Tamil phonology ( /ˈtæmɪl/; தமிழ்; Tamiḻ; [t̪ɐmɨɻ]; pronunciation ) is characterised by the presence of retroflex consonants and multiple rhotic consonants. It does not distinguish phonologically between voiced and unvoiced consonants; phonetically, voice is assigned depending on a consonant's position in a word. Tamil phonology permits few consonant clusters, which can never be word initial. Native grammarians classify Tamil phonemes into vowels, consonants, and a "secondary character", the āytam.
The vowels are called uyir ezhuthu ('life letter'). The vowels are classified into short and long (five of each type) and two diphthongs.
The long (nedil) vowels are about twice as long as the short (kuṟil) vowels. The diphthongs are usually pronounced about one and a half times as long as the short vowels, though most grammatical texts place them with the long vowels.
Tamil has two diphthongs /aɪ/ ஐ and /aʊ/ ஔ, the latter of which is restricted to a few lexical items.
The consonants are known as meyyezhuttu ('body letters'). The consonants are classified into three categories with six in each category: vallinam ('hard'), mellinam ('soft' or nasal), and idayinam ('medium'). Tamil has very restricted consonant clusters (for example, there are no word-initial clusters) and does not have aspirated stops. Some scholars have suggested that in Chenthamil (the period of Tamil history before Sanskrit words were borrowed), stops were voiceless when at the start of a word and voiced allophonically otherwise.
A chart of the Tamil consonant phonemes in the International Phonetic Alphabet follows:
- [s] and [ɕ] are allophones of t͡ɕ in some dialects.
- /f/ and /ʂ/ are found only in loanwords and frequently replaced by native sounds.
- [ɦ] and [x] are allophones of /k/ in some dialects.
The voiceless consonants have multiple allophones, depending on position.
Unlike most other Indian languages, Tamil does not have aspirated consonants. The Tamil script does not have distinct letters for voiced and unvoiced stop, although both are present in the spoken language as allophones—i.e., they are in complementary distribution and the places they can occur do not intersect. For example, the voiceless stop [p] occurs at the beginning of the words and the voiced stop [b] cannot. In the middle of words, voiceless stops commonly occur as a geminated pair like -pp-, while voiced stops usually do not. Only the voiced stops occur after a vowel, or after a corresponding nasal. Thus both the voiced and voiceless stops can be represented by the same script in Tamil without ambiguity, the script denoting only the place and broad manner of articulation (stop, nasal, etc.). The Tolkāppiyam cites detailed rules as to when a letter is to be pronounced with voice and when it is to be pronounced unvoiced. The rule is identical for all stops.
With the exception of one rule - the pronunciation of the letter c at the beginning of a word - these rules are largely followed even today in pronouncing centamil. The position is, however, much more complex in relation to spoken koduntamil. The pronunciation of southern dialects and the dialects of Sri Lanka continues to reflect these rules to a large extent, though not completely. In northern dialects, however, sound shifts have changed many words so substantially that these rules no longer describe how words are pronounced. In addition many, but not all, Sanskrit loan words are pronounced in Tamil as they were in Sanskrit, even if this means that consonants which should be unvoiced according to the Tolkāppiyam are voiced.
The failure of written Tamil to provide different characters for voiced and unvoiced stops, unlike Indo-European languages and other Dravidian languages, attests that voice did not originally contrast in stops. Voiced stops may have been the result of rules of allophony, elision, or sandhi.
Elision is the reduction in the duration of sound of a phoneme when preceded by or followed by certain other sounds. There are well-defined rules for elision in Tamil. They are categorised into different classes based on the phoneme which undergoes elision.