"Phonology" is the branch of linguistics concerned with the articulatory and auditory domain of grammar—that is, with the theory of what John Langshaw Austin (1962) called phonetic acts. Its subject matter links with but is distinct from that of syntax, semantics, and pragmatics. It covers the forms in which the sounds of words are kept in memory and the manner in which the motions of speech organs are shaped by grammar.
Unlike syntax, semantics, and pragmatics (but like closely related morphology), phonology has been largely ignored by philosophers. On the whole, philosophers consider the fact that natural languages are primarily spoken rather than written as of little interest for what Michael Dummett (1986) calls a "philosophical explanation" of language. This attitude stems largely from the mistaken but widely held view that spoken signs are arbitrary sounds whose individuating traits are those of noises. On that view, utterances contemplated apart from their semantic and syntactic features are merely tokens of acoustical types, bereft of grammatical properties, fully described by the physics of noises, and available for human communication simply because humans can perceive and produce them; there is nothing intrinsically linguistic about them. Nor is this attitude an accident. Historically, philosophers have had little incentive to reflect on the sound of language. Most belong to traditions that admit no crucial differences (except perhaps those that pertain to pragmatics) between natural languages and notational systems developed by scientists, mathematicians, or philosophers for the elaboration of their theories. Such notational systems have a syntax and a semantics of sorts, but they have no phonology. Their constituent elements are typically spatial ideographs that share little with the phonological structures of natural languages. Studying language with such a bias offers few reasons, if any, to focus on what is spoken rather than written. It can, however, entrap one in a false conception of linguistic signs, so false, in fact, as seriously to weaken philosophic doctrines built on it.
Phonology rests on a series of presumptions—each supported by a vast body of observations—that together entail that the sounds of natural languages are not arbitrary human noises, on a par with grunts or snorts, whose individuating attributes lie entirely outside the domain of grammar.
The first such presumption is that when people acquire a word they memorize the underlying phonological representation of that word, a representation that defines—but often only partially—how the word is pronounced. These representations have the structure of linearly arrayed discrete timing positions that are assigned pointers to articulatory organs (lips, blade of tongue, dorsum of tongue, root of tongue, velum, vocal cords) implicated in the pronunciation of the word, and pointers to actions these organs execute during speech. The first timing position for the English pin, for instance, points to the lips, the vocal cords, the velum, full closure of the first, stiffening of the second, and nonlowering of the third.
A second presumption is that these pointers (called phonological features) on timing positions are drawn from a finite repertoire, common to all languages, and that they are combined within and across timing positions in rule-governed ways. Some rules are common to all languages and reflect innate linguistic endowments, others are language specific and reflect the influence of linguistic exposure. No language, for instance, avails itself of nasal snorts. French admits rounding of the lips in combinations of features that English excludes (thus the sound ü in French but not in English). Korean, unlike English (except for h ), admits aspiration in underlying phonological representations. German, unlike English, admits initial sequences corresponding to sounded k followed by sounded n. All languages assemble features in similar (three-dimensional-like) structures.
A third presumption is that underlying phonological representations, in isolation or when compounded in complex words, are subject to rule-governed processes that add, subtract, or modify phonological features, which group them into syllables, feet, and prosodic words, which assign stresses and (in some languages) tones, and which ultimately yield final articulatory instructions, so-called surface phonological representations related to, but often very different from, the underlying representations in memory. Processes of this sort account for the fact that, for example, leaf occurs as leavz (with v instead of f ) in the plural, or that serene is pronounced differently when alone than when a constituent of serenity, or that p gets aspirated in pin though not in spin. The details of these rules, the manner of their application, the universality of their formats, and the options fixed by different languages are all objects of intense research and controversies. But the evidence in behalf of their reality seems irrefutable.
Phonology is of philosophic interest, not only because it brings into question analogies between contrived notational systems and natural languages, but also because it raises conceptual issues of its own. Two can be mentioned here.
First, individual spoken utterances are analyzable in both acoustical and phonological terms. No generalizable exact correspondences between these two analyses are known. None may be forthcoming. For instance, nothing acoustical corresponds to word division. How can this dualism be reconciled? Is there a cogent sense in which the objects of speech production are the same (or belong to the same types) as those of speech perception? Offhand, the problem resembles that raised by other events amenable to multiple descriptions. But in this case solutions must be attuned to much that is already understood about both phonology and acoustics. It is not a simple task.
Second, phonological theory associates multiple representations with each utterance—including an underlying representation and a surface one—and it describes them all in the same notation. Surface representations can be conceptualized as instructions (or intentions) to move articulators in certain ways; their ontological status, though unclear, is at least comparable to that of other familiar cases. Not so the other phonological representations. They do not have familiar analogues. The semantic domain of phonological notation therefore cannot be ontologically homogeneous. Furthermore, part of that domain is deeply perplexing.
Anderson, S. R. Phonology in the Twentieth Century. Chicago: University of Chicago Press, 1985.
Austin, J. L. How to Do Things with Words. Cambridge, MA: Harvard University Press, 1962.
Bromberger, S., and M. Halle. "The Ontology of Phonology." In On What We Know We Don't Know, edited by S. Bromberger. Chicago: University of Chicago Press, 1992.
Dummett, M. In Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, edited by E. LePore. Oxford and New York: Blackwell, 1986.
Kenstowicz, M. Phonology in Generative Grammar. Cambridge, MA: Blackwell, 1994. An introduction to the field and a complete bibliography.
Quine, W. V. O. Word and Object, chap. 3. Cambridge, MA: MIT Press, 1960.
Sylvain Bromberger (1996)
Morris Halle (1996)
pho·nol·o·gy / fəˈnäləjē; fō-/ • n. the branch of linguistics that deals with systems of sounds (including or excluding phonetics), esp. in a particular language. ∎ the system of relationships among the speech sounds that constitute the fundamental components of a language.DERIVATIVES: pho·no·log·i·cal / ˌfōnəˈläjikəl/ adj.pho·no·log·i·cal·ly / ˌfōnəˈläjik(ə)lē/ adv.pho·nol·o·gist / -jist/ n.