Thai Letters That Sound the Same: 42 Letters, 21 Sounds

Forty-two letters, twenty-one sounds

The Thai consonant alphabet is listed with 44 signs, two of which, ฃ and ฅ, went out of general use decades ago and survive only on wall charts and old typewriters. That leaves 42 letters a reader actually meets. Between them they spell twenty-one different sounds at the start of a syllable.

Nine letters own a sound outright, with no competition anywhere in the alphabet: ก, ง, จ, บ, ป, ม, ร, ว and อ. The remaining 33 are distributed across twelve sets, and inside a set the letters are pronounced the same. ศ, ษ and ส all open a syllable with s. ท, ธ, ฑ, ฒ, ฐ and ถ all open one with an aspirated t. ข, ค and ฆ all open one with an aspirated k, and the only thing separating them is which class they belong to.

This is not the same problem as Thai letters that look alike. Those are pairs the eye confuses, ก against ถ, บ against ป, and they are fixed by finding the stroke that separates the shapes. The letters on this page are the opposite case: they are easy to tell apart on the page and impossible to tell apart in the air.

The twelve sets

Read the third column first. It is doing more work than the sound is.

Twelve sets holding 33 letters, plus nine letters that stand alone, is twenty-one sounds. The vertical bar in the second column marks a class line, and it is drawn in seven of the twelve rows.

Seven sets where the letter is choosing the tone

Take ฟัน and ฝัน. Same vowel, same final consonant, and the two initial letters are both read f, so a transcription that only records sounds would have to write them the same way. They are not the same word. ฟ is low class, which gives a live syllable with no tone mark the mid tone, so ฟัน is fan and means a tooth. ฝ is high class, which gives the same shape of syllable the rising tone, so ฝัน is fǎn and means to dream.

The same thing happens with ขา and คา, and with ผัด and พัด. In each pair the letter is carrying a piece of information that never reaches the consonant itself: it reaches the tone of the whole syllable. Consonant class is the mechanism, and it is why these seven sets are worth knowing as sets rather than as a list of letters.

There is a pattern in which letter went where. In every one of the seven, the high-class member is the letter that was voiceless in the older language and the low-class members are the ones that were voiced. Thai lost the voicing distinction and gained a tone distinction in the same event, so a difference you would once have heard in the consonant you now hear in the pitch of the syllable. Nothing was thrown away; it moved.

Five sets where the choice is spelling and nothing else

ญ and ย are both low class. ฎ and ด are both mid. ฏ and ต are both mid, ณ and น are both low, ล and ฬ are both low. In these five sets the two letters agree on the sound, agree on the class, and therefore agree on the tone. Reading ญ where ย stands, or the other way round, changes nothing you could hear.

What the letter tells you instead is where the word came from. The second member of each pair turns up mostly in vocabulary borrowed from Pali and Sanskrit and in formal or official words: กฎ, a rule, with ฎ. ปฏิบัติ, to carry out, with ฏ. กรุณา, the polite please, with ณ. กีฬา, sport, with ฬ. ญี่ปุ่น, Japan, with ญ. A Thai reader sees the letter and knows the register of the word before knowing the word.

For reading, that is free information you can ignore at first. For writing it is the whole difficulty, because going from a sound to a spelling means picking one letter out of a set with no acoustic clue at all. This asymmetry is worth naming early: these sets are cheap to read and expensive to write, and most learners only ever need the cheap direction.

Where the extra letters came from

The Thai alphabet was adapted from an Indic writing system built for languages that distinguished sounds Thai never had: a voiced series against a voiceless one, an aspirated series against a plain one, and a retroflex series articulated further back in the mouth. Six letters for one aspirated t is what those distinctions look like once the distinctions themselves are gone.

The letters were kept rather than dropped because the loanwords were kept. Spelling a Sanskrit word with the letters it arrived with preserves the shape of the word across the languages that borrowed it, which is the same reason English holds on to the ph in philosophy. The cost is paid by the learner, and it is paid once: the surplus is a fixed list of 33 letters in twelve groups, not a rule that generates new cases.

You never have to guess which one you are looking at

Here is the part that makes this a smaller problem than it sounds. Reading runs from the page to the sound, and in that direction the ambiguity does not exist. The letter is printed in front of you. ส is not competing with ศ for your attention, because only one of them is on the page, and both would have given you the same s anyway.

Two working habits follow from that:

• Learn each letter with its class attached from the first meeting, because the class is the only part of a homophone set that changes what you say.
• Do not try to memorise the twelve sets as sets. Meet the letters one at a time and let the set assemble itself; the useful moment is recognising that a new letter sounds like one you already have, which takes a second.

Two caveats belong here. ฑ has a second reading, d rather than an aspirated t, in a small group of words taken from Pali, บัณฑิต among them. And ห in front of certain letters is not pronounced at all but changes the tone of what follows, which is a different job entirely and the reason ห appears far more often than the h sound does. Neither caveat affects the count above, since both are about what a letter does in a particular position rather than which sound it owns.

The order these letters arrive in

Meeting a whole homophone set at once is the worst way to take it in, because the letters have nothing to distinguish them except the shape and you end up drilling five near-identical cards. Meeting them spread out is much easier, since each new letter arrives attached to a sound you can already read.

Phayan is an app for learning to read Thai script, and it introduces letters a few at a time from zero rather than handing over a chart. ส arrives early in the high class, alongside ข and ถ, because it is the s you will meet in nearly every s word. ศ and ษ come a couple of rounds later, ซ waits for the low class, and the rare partners, ฆ ธ ณ ภ ฑ ฒ ฌ ฬ, are grouped together much further along, where they can be taken as recognition rather than drilled. Every word has audio, and progress is kept on the device without an account, so you can go and test the sets above straight away.

Frequently asked questions