The surface phonological representation of a sentence consists of a segmental representation (root nodes and segmental features), a tonal representation, and a prosodic structure representation; the last consists of a hierarchical representation of prosodic constituents and a representation of the prominence (head) of each constituent. According to the Prosodic Structure hypothesis, prosodic structure organizes sentence phonology and phonetics just as it does word phonology and phonetics (Selkirk 1986, 1995, 2000; Nespor and Vogel 1986; Pierrehumbert and Beckman 1988). The constituents of the prosodic hierarchy most directly relevant to sentence phonology are the prosodic word, the minor and major phonological phrases (also called “accentual” and “intermediate” phrases), the intonational phrase, and the utterance, in that order. These constituents and their prominent heads are claimed to be universally present in human languages and constitute the structure with respect to which linguists define phonological and phonetic phenomena.
Tonal phenomena are perhaps most revealing of the organization of the sentence into prosodic constituents. The tones that play a role in sentence phonology have various sources: thus, there are tones contributed by the lexical representation of individual words, tones that are phrase-level morphemes, and tones that are epenthetic. Sentence-final tonal morphemes indicating the role of a sentence in a discourse—e.g., the declarative low (L) boundary tone of English—appear in all types of languages. Other tonal morphemes found in the sentence may be indicators of the focus status of a constituent such as the focus H tone in Swedish, or various pitch accents of English, designated H*,L*,L + H*,L* + H (Pierrehumbert and Hirschberg 1990). Some tones that appear in the sentence are apparently neither lexical nor morphemic but rather epenthetic, present for purely phonological reasons.
Phonological constraints determine the precise location in the sentence of these various sorts of tones and tend to produce representations in which individual tones align either with the prominent head of a prosodic constituent or with the edge of a constituent. In Bengali (Hayes and Lahiri 1991), a default tone is epenthesized on the prominent syllable of a phonological phrase. In English, pitch accent morphemes, plausibly generated as part of the morphosyntax of the sentence, are drawn to phrasally prominent syllables; in Chichewa, a lexical H tone in a word shifts its position in order to coincide with a phrasal prominence. The edges of prosodic phrases are also a standard location for tone. In Bengali and Japanese, epenthetic tones mark the edges of major and minor phonological phrases, respectively. It is also conceivable that the presence of certain tonal morphemes in particular locations in surface morphosyntactic representation—such as the “continuation” H tone of English—may induce the presence of a prosodic phrase edge at that location in surface phonological representation.
In addition to constraints calling for the coincidence of tones and prosodic heads or edges, there are also constraints that call for the spreading of tone, or that rule against certain sequences of tone; these, too, appeal to prosodic constituent structure. For example, for a tone to spread from one syllable to another, the syllables must be adjacent within a particular prosodic constituent. A prosodic constituent may thus constitute the domain within which a phenomenon takes place.
Prosodic structure also organizes sentential segmental phonology and phonetics. Recent phonetic work shows that, in various levels of prosodic constituent structure, the left edge is the locus of articulatory strengthening. This is seen, for example, in the introduction of a prevocalic glottal stop in English, and in the degree of linguopalatal contact of /n/ in French. Moreover, external sandhi phenomena are restricted by prosodic structures, as documented in Nespor and Vogel 1986, among others. In Bengali, for example, word-final r assimilates completely to a following coronal consonant just when this segmental sequence lies within the same phonological phrase (examples from Hayes and Lahiri 1991).
The prosodic constituent relevant for defining the domain of assimilation is exactly the one relevant for determining the presence of the epenthetic H- phrase edge tone and for determining the presence of the phrasal prominence, which receives an epenthetic L* pitch accent. This is an example of the domain convergence that the prosodic structure hypothesis predicts: it is expected that distinct types of phonological phenomena may converge on the same prosodic constituent structure, collectively providing important evidence for that constituency in the first place.
Prosodic structure is itself determined by various kinds of constraints. There are phonological constraints on prosodic structure per se, construable as prosodic markedness constraints. A basic type is constraints on prosodic domination, which require that, in the unmarked case, prosodic structure is strictly layered, in the sense that a constituent of a higher level in the prosodic hierarchy immediately dominates only constituents of the next level down in the hierarchy (Selkirk 1995). One further subtype is alignment constraints, which call for the alignment of edges of prosodic phrases and prosodic phrasal prominences (Truckenbrodt 1995). Another is constraints on the size, and specifically on the binarity, of prosodic constituents (Selkirk 2000). The existence of these properly phonological constraints on the domain structure of phonology provides one of the strongest arguments for the autonomy of this (prosodic) domain structure from the surface morphosyntactic structure of the sentence.
Alongside these phonological constraints on phonologically relevant higher-order structure, there are constraints that call for a certain faithfulness to the morphosyntactic structure of the sentence. Selkirk 1986 proposes a class of constraints requiring that the edges of designated syntactic constituents in morphosyntactic structure (word, maximal projection) align with the edges of corresponding constituents in prosodic structure. Truckenbrodt 1995 proposes that the Focus representation of a sentence is reflected in the prosodic prominence structure of the sentence. An understanding of the precise ways in which the information structure and constituent structure of surface morphosyntactic structure impinge on the surface phonological structure is a matter of continuing research.
Current models of grammar construe the phonological and morphosyntactic output representations as being co-present, and therefore they allow, in principle, that phonological constraints other than the general faithfulness constraints cited above might include direct appeals to morphosyntactic conditions. A strong version of the prosodic structure hypothesis holds that effects of morphosyntactic structure on sentence phonology and phonetics are never direct, but always mediated by prosodic structure (Selkirk 1986, Nespor and Vogel 1986); a weak version of the hypothesis would allow that some constraints on phonological phenomena also have direct access to morphosyntactic representation (Kaisse 1985). The co-presence of phonological and morphosyntactic output representations also raises the question of whether phonological constraints ever determine aspects of surface morphosyntactic form (Inkelas and Zec 1990). Some have proposed that the phonological imperative to align a phrasal prominence toward the right end of a sentence might be responsible for the (syntactic) ordering of non-Focus, non-phonologically prominent constituents away from the right edge. Further research on these questions is needed to determine just what phonology-syntax interactions are possible.
Hayes, Bruce, and Aditi Lahiri. 1991. Bengali intonational phonology. Natural Language and Linguistic Theory 9.47–96.Find this resource:
Inkelas, Sharon, and Dragal Zec. 1990. Prosodically constrained syntax. In The phonology-syntax connection, edited by S. Inkelas and D. Zec, pp. 365–378. Chicago: University of Chicago Press/CSLI.Find this resource:
Kaisse, Ellen M. 1985. Connected speech: The interaction of syntax and phonology. San Diego, Calif.: Academic Press.Find this resource:
Nespor, Marina, and Irene Vogel. 1986. Prosodic phonology. Dordrecht: Foris.Find this resource:
Pierrehumbert, Janet, and Mary Beckman. 1988. Japanese tone structure. Cambridge, Mass.: MIT Press.Find this resource:
Pierrehumbert, Janet, and Julia Hirschberg. 1990. The meaning of intonational contours in the interpretation of discourse. In Intentions in communication, edited by P. R. Cohen et al., pp. 271–311. Cambridge, Mass.: MIT Press.Find this resource:
Selkirk, Elisabeth. 1986. On derived domains in sentence phonology. Phonology 3.371–405.Find this resource:
Selkirk, Elisabeth. 1995. The prosodic structure of function words. In Papers in Optimality Theory, edited by Jill Beckman et al., pp. 439–700. Amherst, Mass.: GLSA Publications.Find this resource:
Selkirk, Elisabeth. 2000. The interaction of constraints on prosodic phrasing. In Prosody: Theory and experiment, edited by M. Horne, pp. 231–262. Dordrecht: Kluwer.Find this resource:
Selkirk, Elisabeth, and Koichi Tateishi. 1988. Constraints on minor phrase formation in Japanese. In Chicago Linguistics Society, 24.316–339.Find this resource:
Truckenbrodt, Herbert. 1995. Phonological phrases: Their relation to syntax, focus and prominence. Cambridge, Mass.: MIT dissertation.Find this resource:
Juncture and Boundary
Boundaries are the edges of syntactic or morphological constituents. It has long been known that these edges are sometimes distinguished phonologically. Thus, in the Sanskrit grammatical tradition, the term sandhi covers phonological phenomena which take place at boundaries; and traditional grammars of Biblical Hebrew recognize pausal forms of words, which occur only at the ends of major phrases.
The term “boundary” was used by Trubetzkoy (1939), the first modern linguist to pay serious attention to such phonological phenomena. American descriptivists used the term “juncture” instead. Both terms are ambiguous—meaning either the edge itself, the symbol that marks the edge, or its phonological signal.
The intensive study of boundary phenomena began in the late 1940s, with detailed studies of English and other languages. Descriptivist techniques permitted the discovery of a host of junctural contrasts, e.g. nitrate vs. night rate vs. Nye trait. Unfortunately, the descriptivists' belief that phonological and grammatical information should be kept strictly separate led to severe theoretical difficulties.
Generative Phonology, by contrast, denies this separation, and thus has been able to treat juncture more directly in terms of syntactic structure. The locus classicus of generative work on boundaries is Chomsky and Halle 1968 (SPE). In SPE, boundaries may be syntactically determined, but they are treated as if they were phonological segments, following descriptivist practice. Three boundary types are posited: of these, the formative boundary (symbolized by +) and the word boundary (symbolized by #) are standardly recognized. A sequence of two word boundaries (##) also plays an important role in SPE, marking off “phonological words.”
SPE made two major contributions to the understanding of juncture. Most important is the formalization of the notion of the phonological word. The procedure starts with a surface syntactic tree (of Standard Theory syntax), and assumes that each formative and constituent carries a particular boundary (either + or #); in this way, any sentence is broken down into a continuous string of phonological words which are not necessarily isomorphic to syntactic structure. Thus the sentence The book was in an unlikely place is divided into three phonological words: the book ## was in an unlikely ## place.
Second, and more influential, is the observation that there are two phonological types of affixes in English, which may be distinguished in terms of boundaries (+ and #). The distinction, though not found in all languages, is widespread; a number of general theories of morphology and phonology have incorporated it in various ways, especially Lexical Phonology.
Since SPE, the notion that boundaries are analogous to segments has been questioned. Most subsequent work has instead referred either directly to morphosyntactic structure—or else to phonological phrases, which are derived from the interaction of syntactic information with factors like rhythm and focus. Thus the importance of the boundaries themselves has diminished, as more attention has been given to the structure marked by them. The study of juncture has become more closely integrated in actual practice with that of intonation and stress, with which it was traditionally classified under the heading of prosody. (See Aronoff and Kean 1980, Selkirk 1984, Nespor and Vogel 1986.)
Aronoff, Mark, and Mary-Louise Kean, eds. 1980. Juncture. Saratoga, Calif.: Anma Libri.Find this resource:
Chomsky, Noam, and Morris Halle. 1968. The sound pattern of English. New York: Harper and Row.Find this resource:
Nespor, Marina, and Irene Vogel. 1986. Prosodic phonology. (Studies in generative grammar, 28.) Dordrecht: Foris.Find this resource:
Selkirk, Elisabeth O. 1984. Phonology and syntax: The relation between sound and structure. (Current studies in linguistics series, 10.) Cambridge, Mass.: MIT Press.Find this resource:
Trubetzkoy, Nikolai S. 1939. Grundzüge der Phonologie. (Travaux du Cercle Linguistique de Prague, 7.) Prague. Translated as Principles of phonology. Berkeley: University of California Press, 1969.Find this resource: