Home > What is Phonology?
The following text was taken from an early version of Paul de Lacy's entry "Phonology" for Oxford Bibliographies Online.
The word "phonology" has several meanings. There are two that are particularly important.
"Phonology" is often used to refer to generalizations about sounds and sound combinations (often called 'sound patterns') in and across languages. While it's not possible to talk about speech sounds without a theory in mind, discussion of sound patterns often avoid committing to a particular theory.
Within Generative Grammar, "phonology" refers to a cognitive module. The module takes inputs consisting of strings of symbols (called 'phonological symbols', 'segments', 'phonemes', 'underlying forms', or 'the input' depending on the theory). The symbols may be accompanied by information about morphology, syntax, and perhaps even some aspects of meaning. The module produces an output representation which serves as the input to the phonetic module(s); these modules ultimately provoke muscle movements that can result in speech sounds. A common point of confusion is the belief that the phonological module manipulates 'speech sounds'; in fact, the phonology manipulates representations which are sent to the phonetic module, which converts them into phonetic representations which are then implemented as muscle movements that can produce audible sound (given the right factors).
The two uses of 'phonology' can be illustrated with the phrase 'the phonology of English'. This phrase can refer to generalizations about sound patterns that are common to humans who are identified as speaking a particular language called 'English' (a socio-political concept). In contrast, the phrase 'the phonological module of English' is meaningless: humans do not share cognitive modules. Even so, and somewhat confusingly, in practice this phrase is often used in the Generative literature as an abbreviation of 'the phonological module of a particular person who identifies as speaking English' or 'the common properties of all phonological modules that belong to speakers of English', or even 'the properties of the phonological module of some idealized speaker who we shall say is speaking 'English'.
The two meanings of 'phonology' are not in opposition. Phonology (sound patterns) is (some of) the data used in theorizing about the phonology (the cognitive module).
There are large variations in sound patterns across languages. For example, Hawai'ian has 9 contrastive consonants, while Ubykh has 86. However, there are commonalities, too, though many are disputed. For example, in at least some environments every language has either an alveolar voiceless stop or a glottal stop (or both). Similarly, no language lacks words that start with a consonant.
There are similarly large variations in phonological modules among humans. Of course, a Hawai'ian speaker has a phonological module that has only 9 distinct consonantal symbols in its outputs, while a Ubykh speaker's phonology has outputs with 86 distinct consonantal symbols. However, a great deal of research has aimed to determine the properties common to all phonological modules. While the outputs of languages are diverse, much work has argued that the representations and processes used to generate phonological outputs are very similar - perhaps identical - in all phonological modules, with only certain aspects of the phonological module (e.g. rules, constraint ranking) differing between modules.
What is the aim of the Phonolab?
The aim of our lab's research is to figure out how the phonological module works. This task is extremely difficult because the phonological module is so deeply embedded in human cognitive systems. The phonological module's computational processes are influenced by many other cognitive modules, too. To make things more difficult, its output is transformed by many other cognitive modules (especially the phonetic module), and distorted by the act of physically producing speech sounds. The main data for determining phonological mechanisms are speech sounds, so determining the properties of the phonological output by inspecting speech sounds involves paring away the influences of other modules - a complicated process.
So, the Phonolab's research does not aim to describe sound patterns; its goal is to determine how human cognition works - and in particular the part of cognition that generates representations that can ultimately be realized as speech sounds.