How to pronounce Arabic letters correctly
An English speaker learning Arabic hits a wall unrelated to vocabulary: several letters are produced in places English never uses, and the nearest familiar sound is a different word, not an accent.
Arabic letters are produced at specific points in the mouth and throat, and that point of articulation — the makhraj — is what separates one letter from another that sounds close to an English ear. Classical Arabic phonetics groups those points into five regions: the mouth cavity, the throat, the tongue, the two lips, and the nasal cavity, with the tongue carrying most of them. Nine Arabic sounds have no English equivalent at all, and a learner who substitutes the nearest familiar sound is not speaking with an accent but saying a different word. Fixing that is not a reading exercise. It needs someone listening while you speak, because the ear that learned a sound wrongly cannot hear it as wrong — it filed both sounds under one label years ago.
In short
- A makhraj is the point in the mouth or throat where a letter's sound is produced — and it separates two Arabic letters that sound alike to an English ear.
- Classical Arabic phonetics groups the points into five regions: the mouth cavity, the throat, the tongue, the two lips, and the nasal cavity. The tongue carries most of them.
- Nine Arabic sounds have no English equivalent — ʿayn, ḥāʾ, khāʾ, ṣād, ḍād, ṭāʾ, ẓāʾ, qāf and ghayn — and the nearest English sound is a different word, not an accent.
- Seven letters are always heavy (khāʾ, ṣād, ḍād, ghayn, ṭāʾ, qāf, ẓāʾ); three more shift between heavy and light depending on what precedes them.
- An ear that learned a sound wrongly cannot hear it as wrong — recording yourself helps a little, a listener helps a lot.
- The Education Endowment Foundation (EEF) reports best effect from one-to-one sessions of about 30 minutes, three to five times a week.
How do you pronounce Arabic letters correctly?
Correct pronunciation in Arabic comes down to putting the sound in the right place. Each letter has a point of articulation — a spot in the mouth, throat or nasal cavity where the airflow is narrowed or blocked — and moving it by a centimetre turns one letter into another. English uses a narrower range of the vocal tract, so several Arabic letters land where an English speaker has never had to aim deliberately.
Reading about the points is worth roughly one lesson. Producing them is a motor skill that develops the way any motor skill does: hear the target, attempt it, get corrected immediately, repeat. The immediate part matters most, because an unheard error is rehearsed rather than removed. Anyone starting from the alphabet rather than from correction will find the full sequence in online Quran classes for adults.
According to the Education Endowment Foundation (EEF), drawing on 123 studies, one-to-one tuition produces +5 months of additional progress with a “moderate” security rating. The same review records a caveat worth repeating: comparisons between one-to-one and small-group tuition show mixed results, and in some studies groups of two or three did equally well or better. The realistic comparison for pronunciation work is not one-to-one against a group of three but one-to-one against a class of twenty, and EEF's small-group review (62 studies, +4 months) puts the optimum group at two to five with a “noticeable reduction in effectiveness” beyond six or seven.
Which Arabic sounds do not exist in English?
Nine Arabic letters have no English counterpart, and they account for most of what makes a learner hard to understand. English speakers reach for the nearest available sound, which is why ʿayn becomes a glottal catch or disappears, and ḥāʾ becomes an ordinary h.
| Letter | What English speakers usually substitute | What it actually requires |
|---|---|---|
| ع (ʿayn) | A vowel, or nothing at all | Constriction in the middle of the throat, voiced |
| ح (ḥāʾ) | English h | The same middle-throat position, voiceless and with more effort |
| خ (khāʾ) | k, or the ch of “loch” approximated | The upper throat, with audible friction |
| ص ض ط ظ | s, d, t and th | The same tongue positions with the tongue body raised and the sound made heavy |
| ق (qāf) | English k | The very back of the tongue, further back than k reaches |
| غ (ghayn) | A French-style r, or g | The upper throat, voiced, paired with khāʾ |
The four letters in the middle row deserve separate attention, because English already has s, d, t and th, so the learner assumes the job is done. The Arabic versions differ in what the rest of the tongue does, and a learner who never raises the tongue body produces the light letter every time.
Where in the mouth does each Arabic letter come from?
Arabic's 28 letters are distributed across five general regions, and knowing the region is usually enough to name an error before going into finer detail. The tongue accounts for the largest share, and it is also where most confusable pairs sit.
| Region | Where it is | Letters produced there |
|---|---|---|
| Mouth cavity | The open space itself, with no contact point | The long vowels: alif, and silent wāw and yāʾ |
| Throat | Three graded positions, deepest to shallowest | hamza and hāʾ · ʿayn and ḥāʾ · ghayn and khāʾ |
| Tongue | Back, edges and tip, against the palate or teeth | qāf and kāf · ḍād · lām, nūn and rāʾ · ṭāʾ, dāl and tāʾ · ṣād, sīn and zāy · ẓāʾ, dhāl and thāʾ |
| Lips | Inner lip against the teeth, or both lips together | fāʾ · bāʾ, mīm and wāw |
| Nasal cavity | The upper nose — the source of nasalisation only | The ghunna carried by nūn and mīm |
Tajweed texts in common use divide these five regions into seventeen specific points. Other classifications count sixteen or fourteen, differing over whether the mouth cavity is a point of its own and how finely the throat is subdivided. The count is a matter of how the map is drawn, not of how the letters sound, and no learner needs to settle it before fixing their own letters.
A learner who has not yet secured the letters in isolation starts before this page, with the Noorani Qaida. The order of the first months for someone already reading with errors is in online tajweed classes, and the sequence for an adult beginner in online Quran classes for adults. What we teach and how is on the Noorani Qaida course.
What are heavy and light letters in Arabic?
Heaviness is a quality laid on top of position: the tongue body rises towards the roof of the mouth, the sound fills the mouth, and the vowel beside it darkens. Seven letters are heavy in every context — khāʾ, ṣād, ḍād, ghayn, ṭāʾ, qāf and ẓāʾ — and the rest are light, except for three that shift according to what surrounds them: alif, lām and rāʾ.
Heaviness is what most often separates an Arabic letter from the English sound substituted for it. Saying ṣād as a plain s is not a small imperfection: the two are different letters, and swapping them swaps the word. Over-correcting into a guttural sound is wrong in the other direction.
Rāʾ repays attention out of proportion to its size, because it appears on nearly every page of running text and it is one of the three shifting letters. A learner who gets rāʾ and lām right has fixed the most audible thing in their reading, which is why teachers reach for those two early rather than starting at the top of the alphabet.
Which Arabic letters do English speakers confuse most?
Errors cluster in a handful of pairs rather than spreading evenly across the alphabet. Four pairs cover most of what we hear in a first lesson with an English-speaking learner.
| Pair | The usual error | What to watch |
|---|---|---|
| ض and د | ḍād emerges as a slightly heavy d | ḍād uses the edge of the tongue against the upper molars; dāl uses the tip. The difference is placement, not force |
| ذ and ز | dhāl emerges as z | The tongue tip shows between the teeth for dhāl and stays behind them for zāy |
| ح and ه | ḥāʾ thins into an English h | ḥāʾ comes from the middle of the throat and hāʾ from the deepest point; ḥāʾ needs firmer constriction |
| ق and ك | qāf emerges as a heavy k | qāf comes from the very back of the tongue; kāf sits a little forward of it |
Heritage learners show a different pattern. A child who hears Arabic at home usually produces the throat letters accurately and stumbles on the emphatic ones, having absorbed the sound system without ever reading. Pew Research Center, using Census Bureau survey data, counts 1.4 million people aged 5 and over speaking Arabic at home in the United States, about two-thirds of those aged 5 to 17 born there — a large group who hear the sounds daily and were never taught to make them deliberately. Where reading was never established, the starting point is the alphabet itself, in the Noorani Qaida.
Does mispronouncing a letter change the meaning?
Substituting one letter for another usually does change the word, because Arabic builds meaning on consonant roots and swapping a root letter produces a different root. Changing only a letter's quality — making a heavy letter light, or shortening a long vowel — more often degrades the sound without producing a different word.
The split between a changed word and a degraded sound sets the repair order. Errors that change the word are fixed first, and refinements of quality come after, because a learner asked to perfect vowel length before fixing a substituted consonant spends months on detail while the underlying error hardens. Any religious ruling that may attach to either category is a question for qualified scholars, and this article does not enter into it.
The same order applies to how a course is built. A programme that opens with tables of rules before hearing the learner read a line has arranged its syllabus around a book rather than around what the learner gets wrong. The order of the first months, once the letters are secure, is in online tajweed classes.
Can you fix your pronunciation without a teacher?
Some of the work is genuinely solitary. Listening to an accurate reciter, imitating one letter dozens of times, and spreading practice across days are all things a learner does alone and should. A meta-analysis of 839 assessments across 317 experiments drawn from 184 articles (Cepeda and colleagues, 2006) found that repetition distributed across days is retained markedly longer than the same repetition packed into one session — so five minutes on four letters daily beats half an hour weekly on the whole chart.
What cannot be done alone is detection. A learner who says ḍād as d hears themselves saying ḍād, because the sound was filed under that label years ago and does not change by listening harder. Catching a substituted consonant reliably takes an ear from outside.
According to the Education Endowment Foundation (EEF), “short, regular sessions (about 30 minutes, three to five times a week) over a set period of time (up to ten weeks) appear to result in optimum impact”. The source recommends more frequency than most learners buy, so anyone taking one lesson a week should expect to make up the difference with short daily practice between lessons.
We give no fixed timeline for securing the letters before hearing someone read, and publish no average, because we have not started keeping a record of our own students' timelines. What can be said honestly is that the number of letters you get wrong, rather than the number of points on the chart, is what sets the length of the work.
Sources
- Education Endowment Foundation (EEF) — One to one tuition, Teaching and Learning Toolkit, review last updated July 2021: one-to-one tuition shows +5 months of additional progress across 123 studies with a “moderate” security rating, and “short, regular sessions (about 30 minutes, three to five times a week) over a set period of time (up to ten weeks) appear to result in optimum impact”. The same review notes that studies comparing one-to-one with small-group tuition show mixed results, and that in some cases groups of two or three were equally or more effective. (Checked 14 September 2026) — link
- Education Endowment Foundation (EEF) — Small group tuition, review last updated July 2021: +4 months of additional progress across 62 studies, an optimum group of two to five, and a “noticeable reduction in effectiveness” once a group passes six or seven. (Checked 14 September 2026) — link
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132, 354–380 — a meta-analysis of 839 assessments across 317 experiments drawn from 184 articles, finding that repetition spread across days is retained markedly longer than the same repetition massed into one session. — link
- Pew Research Center — Mohamad Moslimani, 5 facts about Arabic speakers in the U.S., 18 May 2023, drawing on U.S. Census Bureau American Community Survey data via IPUMS: 1.4 million people aged 5 and over speak Arabic at home in the United States, and about two-thirds of those aged 5 to 17 were born in the country. (Checked 14 September 2026) — link
- The division into five regions and seventeen points is a descriptive classification standard in tajweed teaching texts, reported here as what is taught rather than as a measured figure. We found no peer-reviewed study comparing methods of teaching the points of articulation, and do not imply one exists.
Last updated:
Related questions
What does makhraj mean?+
How many points of articulation does Arabic have?+
Do I need to learn the whole map to read properly?+
Why can I not hear the difference when I record myself?+
Can Arabic pronunciation apps correct me?+
Is pronunciation different for Quran reading and for conversation?+
Reading is no substitute for a real lesson
15 minutes with a certified teacher, no card and no sign-up — or subscribe from $30 a month.
A free 15-minute intro — no card, no sign-up

