In South India, there is a language called தமிழ். To pronounce it, த is "tha", மி is "mi" and ழ் is "zh". To pronounce "zh", try to imitate a baby saying the letter "r", but add a tinge of the "z".
Anyway, in English, this language is called Tamil. However, the pronunciation dictates that it be called "Thamizlh", "Thamizh", or at least "Thamil". The sound "Th" had been applied in English, so British traders should have been able to convert த to th. Why did they make a mistake and call it "t"? And, as for the letter "ழ்", it was well used in Thamizh, and I do not understand why the British failed to add the letter "zh" or at least convert ழ் into zlh or rlh or zh instead of the least obvious "l". Many fruits, such as மாம்பழம் (mango) have the "ழ்" letter in it. In fact, all fruits end with the word பழம் which means fruit. I do know about the flourishing Indian Ocean trade. When trading fruits and other commodities, the letter ழ் would have been passed on to Southeast Asia, the Islamic World, the Swahili Coast and China. So, when Europeans entered the trade systems during the age of exploration, they would have been exposed to the unique letter. Once they were exposed to it, the started saying it fluently but did not have a letter for it. When transliterating, they would use "zh", although it sounds more like "rlh" in my opinion. So, they were aware of ழ். Why isn't தமிழ் called "Thamizh" or something better than Tamil? Since Tamil is spelt horribly, many Westerners are pronouncing it wrong.

A side question: In other languages, such as Arabic, Chinese, French and Spanish, is தமிழ் transliterated correctly or badly? I would suspect that in Arabic and Chinese, it is transliterated correctly because of exposure through ancient trade.

  • 2
    I don't speak Tamil, but I thought ழ was sometimes pronounced like the English "L" (IPA /l/), depending on dialect? – Draconis Jun 11 '20 at 2:24
  • 2
    If the Wikipedia page on Tamil phonology is anything to go by, Tamil does not have aspirated stops at all, so th does not even exist in Tamil. According to Wikipedia, the sound /ɻ/ is sometimes transcribed as zh, but given its pronunciation, that seems rather a suboptimal solution. L is not ideal given the presence of both /l/ and /ɭ/ in the language, but zh strikes me as worse. Also according to Wikipedia, the sound colloquially tends to shift towards [l], so l may not be so very bad after all in a broad transcription. – Janus Bahs Jacquet Jun 11 '20 at 6:50
  • 2
    @JanusBahsJacquet - According to Grammar of the Tamil Language, Andronov M, 1987, Tamil really doesn't have aspirated stops phonemes, but the phonemes /p/, /t/, /k/ at the word onset by some speakers are pronounced as aspirated [pʰ], [tʰ], [kʰ]. – Yellow Sky Jun 11 '20 at 10:45
  • 1
    @Manu Aspiration means that there’s a little puff of air after the consonant. If you listen to an American or British person saying the words pin and skin, you may be able to hear that kin sounds a bit like ‘khin’, but skin does not sound like ‘skhin’. If you’re familiar with Hindi, it’s the difference between त /t/ (unaspirated) and थ /tʰ/ (aspirated). Of these two, the Tamil /t/ is described as primarily being pronounced the same as Hindi त and only by some people sometimes the same as Hindi थ. – Janus Bahs Jacquet Jun 11 '20 at 14:11
  • I rolled back this question to its previous version; StackExchange doesn't let users delete questions that have upvoted answers, because upvotes generally mean "this answer is useful or valuable", so we want to keep them around for others who are curious to see. – Draconis Dec 13 '20 at 17:37

It is all because of the fact that the word "Tamil" got into English not from speakers of Tamil. It got into English from the Pali language and it is a cognate of the word Dravida.

The Online Etymology Dictionary says:

1734, from Pali Damila, from Sanskrit Dramila, variant of Dravida

Actually, I cannot understand why you think த is "tha". It is pronounced [t̪ʌ], [t̪] being dental [t], like in Spanish or in Italian. Also, ழ், voiced retroflex approximant [ɻ], is acoustically very close to [l], it is even transliterated as <ḻ>. The traditional way to transliterate the word "தமிழ்" is < tamiḻ >, so I won't be surprised to learn that the word "Tamil" was borrowed into English directly from Tamils, with [ɻ] substituted for [l] as the closest-sounding consonant.

In Spanish it is also Tamil, in Arabic it is تاميلية (tāmīliyya), in Chinese it is 泰米爾 (Mandarin: Tàimǐ'ěr, Cantonese: taai3 mai5 ji5).

UPD.: I've found a book with a detailed description of Tamil phonology and phonetics, Grammar of the Tamil Language, Andronov M, 1987, in Russian. It states Tamil doesn't have aspirated stops phonemes, but the phonemes /p/, /t/, /k/ at the word onset by some speakers are pronounced as aspirated [pʰ], [tʰ], [kʰ] allophones. This means that pronouncing த as "tha" is not predominant. Also, An Intensive Course in Tamil, S. Rajaram, 1979, which has Tamil words transliterated in Devanagari, has त [t̪a] as the first syllable of "தமிழ்", screenshot here.

  • According to Wikipedia, the Cantonese name is 淡米爾 daam6 mai5 ji5. Etymonline is generally hopeless at properly transcribing non-English languages, but I presume the Sanskrit forms should have retroflex and (those should be dot diacritics, not vertical line, but I can’t find the dot on my phone IPA keyboard). – Janus Bahs Jacquet Jun 11 '20 at 6:35
  • @JanusBahsJacquet - In Sanskrit it is द्रमिल dramila (no diacritic dot), but द्रविड draviḍa (with a dot). – Yellow Sky Jun 11 '20 at 11:48
  • @Manu When you say "TH", is it a continuous sound that you can sustain for a long time (like "S"), or a short one-time sound (like "D")? – Draconis Jun 11 '20 at 17:03
  • @Manu huh? ITH + A becomes "Ta". த் + அ = த – a20 Dec 12 '20 at 8:35
  • @Manu As a native Tamil speaker, I don't agree with the "zh" usage. It's quite non-intuitive. The closest would be Tamilh .. There is definitely a L sound in ழ். In fact just like how ந/ன, ண "deepen" in pronounciation .. ல, ள, ழ deepen in pronounciation. Threre is no H .. Pure Tamil doesn't have H, unlike Northern languages or even other South Indian languages. – a20 Dec 12 '20 at 8:42