Sounds in Systems (13-Nov-2003)

Any spoken language makes use of only a subset of all the possible sounds that can be vocalised. That subset is the phonology of the language. As previously seen, it may be that two or more separate phonetic sounds have equivalent phonemic values in a language. For example, in English, the letter "t" may be pronounced as a glottal-stop in a word without changing the word's meaning. The different phonetic values of a given phoneme are called allophones.

In describing the phonology of a language, one therefore needs to identify the phonemes of that language. This means listening to spoken examples of the language, and identifying which sounds are made by speakers of the language, and then, with the help of a native speaker of the language, determining which of the sounds are linguistically significant. One technique for establishing which sounds are phonemes is using a "minimal pair" test - find words whose meaning can be changed by modifying a single sound, e.g. English "pit" vs. "bit".

In assessing whether different sounds are allophones of the same phoneme, there are certain factors that may come into play:

Invariably, there are patterns to be found in the phonology of a language. For example, with the English plosive consonants

each pair shares the same place of articulation and differs only in the voicing.

Having established a set of phonemes for a language, a further step is to try and determine what rules govern their use. For example, (from Crystal) the "b" sound, which is voiced, loses some of its voicing when it occurs at the end of a word (e.g. "jab").

Useful book readings for this lecture:


Sounds, Grammar and Meaning page