The alphabet problem

Why Thai romanization is a mess, and how to stop it slowing you down

Four spellings of "thank you", none of them wrong, and none of them the way to learn it. Here is what is actually going on and the four rules that make it stop mattering.

Look up how to say thank you in Thai and you will be offered khob khun, kop khun, khàwp khun and khòrp-khun. Look up Thailand's capital city and you will find Krung Thep on the road sign, Krungthep on the airport announcement and grung-thâyp in the app you are learning from.

Nobody is wrong. That is the frustrating part. Thai romanization is not a standard that people keep breaking. It is a genuine absence of a standard, and once you understand why, you can stop trying to find the right one.

There is an official system, and it is not built for you

The Royal Thai General System of Transcription (RTGS) is the official romanization, and it is what you see on road signs, maps and government documents. It is also, for a learner, close to useless, and that is by design rather than by accident.

RTGS exists to write Thai place names in Latin letters so that a person who does not speak Thai can put them on a sign, type them into a form and file them alphabetically. To make that work it deliberately throws away two things: tone, and vowel length.

Here is what that costs. These three words are one of the first things anyone learning Thai gets taught, because they are identical except for pitch:

In RTGS all three are spelled ma. One spelling, three words, and the difference between arriving, a dog and a horse is not recoverable from the page. The system is not hiding the information to annoy you. A road sign has no use for it.

The four khao problem

It gets worse with a syllable this common. All four of these are written khao in RTGS:

Play them. The first two are separated by tone alone. The last two have the same tone and differ only in how long you hold the vowel.

Now look at the middle column, which is our romanization, not RTGS. It separates the first three by tone. It writes the last two identically, and that is a real gap in our own system: it shows vowel length by doubling the letter in most words (มา is maa, สาม is sǎam) but the ao diphthong comes out the same whether it is long or short. So our system is better than the official one and still cannot do the whole job. That is not false modesty, it is the actual state of writing Thai in Latin letters.

The h is not the sound you think it is

If you take one thing from this page, take this one, because it is the single most common mispronunciation by English speakers and it is entirely a notation problem.

In ph, th and kh, the h does not add a sound. It marks aspiration, the small puff of air that comes out with the consonant. Hold your palm in front of your mouth and say "pot": you feel the puff. Say "spot": you do not. English has both sounds and treats them as one letter. Thai treats them as two different letters.

Which raises the obvious question: if ph is the p in "pot", what writes the p in "spot"? Some systems use plain p, which is why you also see "Puket". Others, ours included, use bp, because English speakers reading a lone p will aspirate it out of habit. That is why our romanization of "sour" looks like this:

So which system should you use

Ours keeps tone marks, because the whole reason a learner reads romanization is pronunciation, and pronunciation without tone is not Thai. That is the same choice most teaching materials make, and it is why almost none of them match the road signs.

Both columns describe the same recording. The left one is answering "how do I file this", the right one is answering "how do I say this". Read them as different questions and they stop contradicting each other.

Bangkok is not a romanization of anything Thai people say

Worth knowing, because it explains a lot of what looks like inconsistency on maps.

Thais call the city กรุงเทพgrung-thâyp Bangkok, which is the short form of a ceremonial name so long it holds a world record. Say "Bangkok" to a taxi driver and you will be understood, because everyone knows what foreigners call it, but it is not a transcription of กรุงเทพ and never was. It comes from บางกอก, an older name for a settlement by the river, which stuck in European usage and then stopped changing while the Thai name moved on.

Once you know that, the rest of the map makes more sense. Some spellings are RTGS. Some are older transcriptions frozen in place by international use. Some are a local authority's preference. Some are what a signwriter did in 1974. They are not competing attempts at the same job, so it is not surprising that they disagree.

The two things romanization can never do for you

Four rules that make this stop slowing you down

And if you just want the thing converted

We built a converter that takes Thai text and gives it back word by word with romanization and, where we have a recording, a play button on each word. It matches against the course vocabulary, longest match first, and a word it does not know comes back as Thai with no romanization rather than a guess. Inventing a romanization for an unknown word is worse than admitting we do not have it.

When you are ready to stop needing it: the alphabet, all 68 characters with audio, and the tone rules, which are the part romanization was hiding from you the whole time.

Learn the rest in the app

All 34 chapters, every word with native-quality Thai audio. Free, forever.

Get told when it launches