Twenty eight letters, and about seventeen shapes. Most of the alphabet is a small set of skeletons distinguished by dots: one shape with one dot below is b, with two above is t, with three above is th, with one above is n, with two below is y. Learn the skeleton and the dots become a detail rather than five separate letters, and the alphabet shrinks by a third before you have memorised anything.
What actually takes the fortnight is that every letter changes shape according to where it sits in the word, and that six of them refuse to join to the letter after them. Those two facts, not the letter count, are why an Arabic word looks like a continuous line rather than a row of characters.
The six things that make the script readable
- letters change shape by position, having an initial, medial, final and isolated form. There are fewer real shapes than that suggests, because most are the same skeleton with a different join.
- six letters never join to the left, so a word breaks visually in the middle. Learning which six removes most of the confusion about where one word ends.
- the alphabet is 28 letters and it is a fortnight of work, not a term. It is written right to left and there are no capitals at all.
- al- is the definite article, written as one piece with the noun, and it assimilates to the following consonant in speech: al-shams is said ash-shams.
- ta marbuta marks the feminine, a letter that appears only at the end of a word and is usually the difference between maktab and maktaba.
- dots separate b, t and th. Several letters share one skeleton and differ only in the number and position of dots above or below, which is why handwriting is harder to read than print.
Four forms, and six letters that break the chain
A letter has an isolated form, an initial form, a medial form and a final form, and the chart in the front of a textbook shows the isolated one, which is the form you will see least often in running text. That is a genuine design problem in most teaching material: the shapes you are drilled on are the shapes that occur rarest.
The six non-joiners are the other half. Alif, dal, dhal, ra, zay and waw connect to the letter before them and not to the letter after, so a word containing one is visibly broken into pieces even though it is a single word. A beginner reads those gaps as word boundaries and mis-segments the line, which is not a vocabulary failure and not a letter failure. It is a rule nobody stated.
Direction is the smallest of the three surprises. Right to left stops being noticeable within a few days, and the thing that keeps catching people afterwards is that numbers run the other way, left to right, inside a right to left line.
You cannot read your way to vocabulary here
Here is the stall and it is structural rather than personal. In an alphabetic language an unfamiliar word can be sounded out and then asked about, so reading builds vocabulary. Arabic leaves the short vowels out, so an unfamiliar word cannot be pronounced until you know which word it is, and the same three consonants with different vowels are frequently different words.
The advice to read a lot therefore fails in a specific way. A learner opening a newspaper to build vocabulary gets nothing, because every unknown word is unreadable rather than merely unknown, and after three pages they conclude the script defeated them. What works instead is material at a density where almost every word is already known, which makes reading a way of making existing vocabulary automatic rather than a way of acquiring new vocabulary. The general form of that constraint is material just past your level, and why that is the sweet spot, and Arabic moves the threshold much higher than other languages do.
One letter, four jobs
What happens to a letter as it moves through a word
Each frame is the same letter. The chart in a textbook shows the first frame, which is the one that occurs least.
- 01IsolatedStanding alone, and the form every alphabet chart prints
- 02InitialJoined forward only, and usually reduced to its head
- 03MedialJoined on both sides, and often little more than a tooth on the line
- 04FinalJoined backward only, and it gets its tail back
The medial form is where the letters look most alike, since several skeletons reduce to the same tooth and only the dots separate them. That is why reading Arabic quickly is a matter of word shapes rather than of letters, and why it takes a fortnight rather than an afternoon.
Two things to read with
Fully vocalised texts are the training wheels and they are widely available: children's books, religious texts and most teaching editions print every short vowel. It replaced guessing at vowels you do not yet have the vocabulary to infer, which is not a skill that can be practised before the vocabulary exists. Free in any library. Read them until the inference starts happening on its own, then stop, because vocalised text is not what you will meet.
The second is a handwriting sheet showing all four positional forms, worked through by hand. It replaced learning the isolated forms from a chart, which teaches the shape that appears least. Free. Writing rather than reading is the point: the hand notices which letters join forward and which do not, and after a page of copying the six non-joiners stop being a list to remember and become something your pen already knows.
The mistake: learning the letter names
Every chart supplies names, and the names are not the sounds. Knowing that the letter is called jim does not help you read faster, and it adds twenty eight items that will be used only when spelling something out loud.
Learn each letter as a sound inside a short word instead, and learn it in the position it usually occupies. The two exceptions are worth naming because they are not sounds at all: alif is a carrier that changes job depending on what is written above or below it, and the ta marbuta at the end of a word is a t that is often not pronounced and that marks the word as feminine. Both are grammar wearing a letter's clothes, and what they are actually doing belongs to three letter roots, and patterns that generate the vocabulary.
What the script leaves undone
Reading an unvocalised word is an act of inference rather than decoding, and the inference runs on knowing which pattern the word is in, which is the system that supplies the missing vowels. The vocabulary that makes the inference possible is built by root rather than by word, in learning by root instead of by word. Several of the letters stand for sounds English does not have at all, and those have to be built with a recording rather than a description, in the throat sounds, and the emphatic consonants. And nothing on this page applies to the spoken language, which is not usually written down at all, in which to learn first, and what each is used for.