Experimental phonetics

1. Speech can be represented as a series of segments.

2. Each segment has a target, and can be specified uniquely with reference to its target. Each target corresponds to a unique auditory percept in the mind of a trained observer.

3. Any segmental target can be specified using a limited number of dimensions. These dimensions do not need to be defined for each language individually. Instead, there is a universal set of dimensions which are relevant to all human languages. For example, the English word fan might be represented as [fæn].1 This implies that it is made up of a sequence of three segments, as, indeed, is suggested by the spelling. Let us now focus our attention on the first segment, represented by the letter f. It is produced by forcing air through a narrow opening

between the lower lip and the upper teeth. The target may be described as: 'loose contact between lower lip and upper teeth, vocal folds open and not vibrating', and is conventionally specified as a voiceless labiodental fricative. Each term of this three-term label refers to one articulatory dimension. Voiceless refers to the fact that the vocal folds are not vibrating (the voicing dimension), labiodental to contact between the upper teeth and the lower lip (the place of articulation dimension), and fricative to the fact that the constriction between the lips and the teeth is sufficiently narrow for the rasping sound to be produced when air flows between them, but does not cut off the airstream entirely (the manner of articulation dimension). These three dimensions are crucial for the description of consonants in all known.

Acquiring skill in impressionistic phonetics is, in many ways, like studying the pronunciation of a new language. For example, a native speaker of English who sets out to learn French must master the vowel sound which occurs in words