A method of writing down speech sounds in a systematic and consistent way - also known as a ‘notation’ or ‘script’. Two main kinds of transcription are recognized: phonetic and phonemic. Square brackets enclose phonetic transcription; oblique lines enclose phonemic transcription. In the former, sounds are symbolized on the basis of their articulatory / auditory identity, regardless of their function in a language (sometimes called an impressionistic transcription). In the latter, the only units to be symbolized are those which have a linguistic function, i.e. the phonemes. An allophonic transcription adds functional phonetic details. A phonemic transcription looks simplest of all, as in this only the units which account for differences of meaning will be represented. In a phonetic transcription, on the other hand, the aim is not to judge the functional significance of sounds, in the context of some language, but to identify the sounds as such.
Phonetic transcriptions which are relatively detailed are called narrow transcriptions; those which are less detailed are called broad transcriptions. In the broadest possible transcription, only those phonetic segments would be notated which correspond to the functionally important units in the language - in other words, it would be equivalent to a phonemic transcription, and some phoneticians do use ‘broad’ in the sense of ‘phonemic’. But in principle it is important to appreciate that the two transcriptions of [pen] and /pen/ refer to very different entities: the first is a broad phonetic transcription, representing a sequence of concrete, physical articulations; the second is a phonemic transcription, representing a sequence of abstract, functional units, and reflecting a particular theoretical point of view.
In any transcription (whether phonetic or phonemic), each distinguishable sound is given its own ‘symbol’. The whole range of available phonetic symbols is known as a ‘phonetic alphabet’. The most widely used such alphabet is the international phonetic alphabet.