Typing Thai: four levels, and no spaces between words

The space bar does something else here

Thai is written without spaces between words. Spaces exist and they mark the ends of clauses and sentences, so a Thai typist presses the space bar a small fraction as often as an English one does, and never to separate two words.

This has a consequence that reaches well beyond typing: working out where one word ends and the next begins is something the reader does, and something software has to be taught to do. Thai word segmentation is a genuinely hard problem with a research literature behind it, which is why search, spellchecking and line-breaking in Thai are all harder than they look.

For a typing tutor it raises an awkward question, and the honest answer is worth stating. The practice runs here present Thai words separated by spaces, which is not how Thai is written. It is a drill rather than a text: separating the words is what makes it possible to tell you which one you got wrong. The sample sentence on the Thai lessons page is written properly, without them, and reading it next to the drill is a fair picture of the difference.

Why a keyboard chart shows less for Thai

The script has 44 consonants, fifteen vowel symbols that combine into around 28 forms, and four tone marks. A level of a keyboard has about 30 usable keys. The arithmetic does not work, and Kedmanee's answer is to use every level it has: unshifted and shifted, across the letter block and the number row, with Thai characters rather than digits in the unshifted number row.

So the three unshifted rows a keyboard chart draws are a smaller share of Thai than they are of Russian. That is measurable rather than a matter of opinion, and the table below measures it: the share of the distinct characters in each language's own practice words that the rendered chart actually produces.

Thai is near the bottom, and so is every other abugida here for the same structural reason. Russian and Hebrew are at the top because their alphabets fit a level of a keyboard with room to spare — which is exactly the property Thai does not have.

Vowels that are written before the consonant they follow

Five Thai vowels are written to the left of the consonant they are pronounced after. A word beginning with one of them starts, visually and in typing order, with a vowel that is spoken second.

It is not an edge case: 19 of the 51 Thai practice words here begin with one, including some of the most common words in the language — และ, ใน, เป็น, ไม่, ให้.

The good news for a typist is that you type what you see. Thai text is stored and typed in the order it is drawn, so a leading vowel is simply the first key you press, and no reordering is happening in your head or in the software. The reordering is between the writing and the pronunciation, which is a reading problem rather than a typing one.

Other vowels sit above or below the consonant, and tone marks stack above them. Those are typed after the consonant they attach to, so the sequence and the visual arrangement come apart in the other direction — two or three keystrokes produce a cluster that occupies one horizontal position.

Kedmanee and Pattachote

Kedmanee is the standard, and the one these lessons teach. It is what is sold, what is installed, and what a Thai office runs on.

Pattachote is the alternative, designed in the 1960s to distribute the work more evenly between the hands and fingers than Kedmanee does. It is more efficient and it lost, which by now is a familiar sentence: it is the same story as Dvorak against QWERTY, Sebeolsik against Dubeolsik, and Turkey's F-keyboard against its Q-keyboard.

The pattern is consistent enough to be worth naming. A layout designed after a standard is established has to be enough better to overcome every keyboard already in the world, and no layout has ever been that much better. Learn the standard one.