How to Build a Language, Part 1

A phoneme is a distinct type of sound that forms a building block of a language. An example in English is the phoneme /k/, which occurs in words such as cow, krill, school, and skid.

Each language has its own phonemes, and a unique system of rules to explain how various phonemes are allowed to be chained together. Human speech is thus built a like chain, with phonemes as the most basic unit: One or more phonemes are chained together to create syllables, syllables are chained together to form words, words form sentences, and so on.

The English language contains 40 phonemes. In comparison, Polynesian contains only 11 phonemes, while Xhosa (a South African language) contains 110. Because our speech muscles are trained from a young age to produce the phonemes of our native language, part of the difficulty in learning a new language is training our speech muscles to produce new phonemes. If your native language is English, give this challenging lesson in Xhosa\’s \”click\” phonemes a try:

Video courtesy of XhosaKhaya.

Unfamiliar phonemes are a challenge for software developers, too. Good speech recognition software has to make clear distinctions between different phonemes, and the more phonemes a language has, the more difficult it is for the software to differentiate them.