This is an archive page; this conference occurred in May 2001.
The mfm homepage is here.
Phonology and syntax - the same or different? An OT
Ricardo Bermudez-Otero & Kersti Borjars (University of Manchester)
In the generative tradition, specifically linguistic knowledge is regarded as AUTONOMOUS (i.e. as arbitrary and self-contained; see Croft 1995: para 2.2). Within this paradigm, the existence of phonology and syntax as distinct domains of linguistic inquiry raises a fundamental question: do both components of language possess autonomy to the same degree? Inspired by proposals in minimalist syntax, Burton-Roberts (2000) and Carr (2000) answer this question with an emphatic negative. In the Minimalist Program, syntactic knowledge is presented as radically autonomous; in addition, parametrization is shunted elsewhere, and so syntax can be conceived of as a purely biological fact. In contrast, Burton-Roberts and Carr regard phonology as FUNCTIONAL: emerging from the phonetics, relying on general cognitive faculties, and constituted by social convention. Optimality-theoretic approaches to syntax lead to rather different conclusions, particularly if harmony maximization is argued to control all syntactic phenomena and not just deletion (cf. Pesetsky 1997). This research supports a conservative position, closer in fact to GB than to Minimalism, in which both phonology and syntax involve a mixture of the biological and the cultural, for in both domains the ranking of constraints, like the setting of parameters, is ultimately a matter of acquired social convention. Moreover, whilst constraints cannot be directly derived from functional considerations and are therefore partially arbitrary, the observation that they are nonetheless often clearly GROUNDED suggests that the arbitrariness thesis cannot be sustained in a radical form, either in phonology or in syntax.
These conclusions depend on the relative success of OT's phonological and syntactic applications. In this respect, we note that, despite appearances to the contrary, OT phonology and OT syntax enjoy similar empirical advantages and encounter similar problems. The advantages include insightful accounts of conspiracies, hedges to rules and principles, the emergence of the unmarked, and top-down effects. The problems include ineffability, optionality, and opacity, as well as the existence of apparently nonoptimal residues of contingent historical change. Differences between phonology and syntax in these areas turn out to be matters of degree: e.g. whilst Pesetsky (1997) uses ineffability to argue that syntax must incorporate a clash-and-crash component, recent work finds recalcitrant instances of the same problem in phonology (Orgun & Sprouse 1999, Torkenczy 2000).
In relation to grounding, we adduce five empirical arguments against the charge that grammatical statements of markedness superfluously duplicate the impact of performance bias upon acquisition and change (cf. Hale & Reiss 2000, Newmeyer 2000):
(i) performance difficulty landscapes do not precisely match linguistic markedness hierarchies;From a conceptual viewpoint, there remains the task of explaining how grounded constraints come into existence. In this area, both phylogenetic and glossogenetic accounts (Hurford 1990) have been advanced, and it seems likely that both types of explanation will be called for in different instances. In either case, the notion of grounded constraint provides an exciting new way of thinking about the tension between autonomy and functionality in phonological and syntactic knowledge.
(ii) constraints are stated in terms of linguistic categories, which do not precisely match the nonlinguistic entities involved in performance bias;
(iii) apparently grounded constraints are freely interspersed with apparently arbitrary ones;
(iv) there are implicational universals that should be violated by blind performance-driven change but are not;
(v) non-performance-driven change is demonstrably constrained by markedness.
Croft, William (1995). Autonomy and functionalist linguistics. Language 71: 490-532.
Burton-Roberts, Noel (2000). Where and what is phonology? A representational perspective. In Noel Burton-Roberts et al. (eds). 39-66.
Burton-Roberts, Noel, Philip Carr & Gerard Docherty (2000) (eds). Phonological knowledge: conceptual and empirical issues. Oxford: Oxford University Press.
Carr, Philip (2000). Scientific realism, sociophonetic variation, and innate endowments in phonology. In Noel Burton-Roberts et al. (eds). 67-104.
Hale, Mark & Charles Reiss (2000). Phonology as cognition. In Noel Burton-Roberts et al. (eds). 161-184.
Hurford, James R. (1990). Nativist and functional explanations in language acquisition. In Iggy M. Roca (ed.). Logical issues in language acquisition. Dordrecht: Foris Publications. 85-136.
Newmeyer, Frederick J. (2000). Optimality and functionality: some critical remarks on OT syntax. ROA-402-08100. (Rutgers Optimality Archive, http://ruccs.rutgers.edu/roa.html).
Orgun, Cemil Orhan & Ronald L. Sprouse (1999). From MPARSE to CONTROL: deriving ungrammaticality. Phonology 16: 191-224.
Pesetsky, David (1997). Optimality Theory and syntax: movement and pronunciation. In Diana Archangeli & D. Terence Langendoen (eds). Optimality Theory: an overview. Oxford: Blackwell. 134-170.
Torkenczy, Miklos (2000). Absolute phonological ungrammaticality: defective paradigms in Hungarian. Paper presented at the 8th Manchester Phonology Meeting, Manchester, 18-20 May 2000.
A Government Phonology perspective
Monik Charette (SOAS, University of London)
The question as to whether phonology and syntax are similar or different, was raised in the 1980s when Chomsky abandoned the view that the syntactic component of a grammar should consist of ordered transformations that derive surface structures from deep structures through intermediate structures. Transformations and intermediate levels disappeared. Structures established at D-structure were now seen as having to be preserved at S-structure. Language specific transformations were replaced by universal principles and parameters. As Bromberger & Halle (1989:51) put it: "This course of events obviously raises the question whether phonology should not undergo a similar development - that is, whether phonological theory should not be restructured in such a way as to exclude rule ordering and representations that are neither underlying representations nor surface forms."
I intend to present some work that has been done in Government Phonology since the mid 1980s showing that like syntax, phonology could be reformulated without recourse to extrinsic rule order. I will show how phonology can be formulated in terms of principles and parameters and how similar the principles proposed in phonology are to those proposed in syntax.
UG and Syntax-Phonology Parallelisms
Phil Carr (Universite Montpellier III)
Paralellisms between syntactic and phonological knowledge have been postulated in a wide variety of approaches to linguistic knowledge, several of which I discuss here.
One approach is the broadly non-Chomskyan, although partly modular, view of phonological acquisition adopted by Locke (1993), who argues that phonology, morphology and syntax are a function of a postulated Grammatical Analysis Module, unique to our species (but quite distinct from UG, as postulated by Chomskyans). Vihman (1996, 2000), adopting broadly Empiricist assumptions, also postulates parallelisms between phonological and syntactic development in the child, parallelisms resulting from, amomg other things, the child's capacity to form inductive and analogical generalisations.
Among broadly generativist frameworks, Anderson's (1985) work on Dependency Phonology and Dependency Syntax postulates a Principle of Structural Analogy between syntactic and phonological structure, but without postulating the existence of UG. Other generativist theories, notably Kaye, Lowenstamm and Vergnaud's (1985) Government Phonology, accept the Chomskyan notion of UG and argue that the syntactic relation of government holds in phonology as well as in syntax, and is given by UG. A similar idea is proposed by Rice (1992), who adopts a government-binding approach to phonological structure.
Following Burton-Roberts (2000), I argue that the position
Government Phonology (and by Rice, and certain versions of Optimality
is problematic. Assuming that UG exists, I propose (see Carr 2000)
if we are to sustain a coherent conception of UG, we must assume that
is phonology-free. In seeking to substantiate this idea, I present
relating to the following. (a) The distinction between sequencing on
one hand and abstract syntactico-semantic relations on the other (e.g.
scope relations, relations governing anaphor interpretation, logical
and the relation of government). (b) The relationship between
and the signal. (c) The argument for UG from the Poverty of the
in syntax (I will argue that there is no such argument available in the
case of phonological acquisition). (d) The role of general cognitive
in phonological acquisition. (e) The notion 'constituency' in phonology
and syntax as it relates to putative syntax-phonology parallelisms
by Pierrehumbert (1990) and by Carstairs-McCarthy (1999). (f) Headhood,
perceptual salience and the figure-ground relation, and their relevance
for the syntax-phonology parallelisms postulated by van der Hulst
(g) The idea that grammars can't count and its relevance for the nature
of word-stress assignment algorithms. (h) The notion that phonology is
substance-free (Hale and Reiss 2000).
Anderson, J. (1985). 'Structural analogy and dependency phonology'. ALH 26: 5-44
Burton-Roberts, N. (2000). 'Where and what is phonology?' In Burton-Roberts, Carr and Docherty (eds).
Burton-Roberts, N., Carr, P. & Docherty, G. (eds) (2000). Phonological knowledge: conceptual and empirical issues. Oxford: Oxford University Press
Carr, P. (2000). 'Scientific realism, sociophonetic variation and innate endowments in phonology.' In Burton-Roberts, Carr and Docherty (eds).
Carstairs-McCarthy, A. (1999). The origins of complex language. Oxford: Oxford University Press.
Hale, M. & Reiss, R. (2000) 'Phonology as cognition'. In Burton-Roberts, Carr & Docherty (eds).
Hulst, H. van der (2000). 'Modularity and modality in phonology.' In Burton-Roberts, Docherty & Carr (eds).
Kaye, J., Lowenstamm, J. & Vergnaud J-R (1985). 'The internal structure of phonological elements. Phonology 2: 305-28.
Locke, J. L. (1993). The child's path to spoken language. Cambridge, Mass.: Harvard University Press.
Pierrehumbert, J. (1990). 'Phonological and phonetic representation.' Journal of Phonetics 19: 375-94.
Vihman, M-M. (1996). Phonological Development. Oxford: Blackwell.
Vihman, M-M & Velleman, S. (2000). 'Phonetics and the
phonology.' In Burton-Roberts, Carr & Docherty (eds).
Are there really phonological constituents at the
phrasal level ?
Elisabeth Delais-Roussarie (CNRS – UMR 5610 / ERSS)
In Prosodic Structure Theory (see, among others, Selkirk
the prosodic constituents are seen as phonological primitives that are
derived from the surface syntactic structure of the utterance. Once
are derived, the phonological component of the grammar has no access to
the syntactic structure per se, but only to the prosodic constituent
Various phonological phenomena are thus analysed within these
(e.g. metrical patterns, sandhi phenomena, etc.).
In this paper, I will show that these constituents might not be phonological entities. This will lead me to propose a model in which the prosodic structure and accentuation are considered as emerging at the surface level from two types of information: morpho-syntactic and metrical. This work is based on an analysis of various French data.
In a first section, I will present the different phrasal constituents of the prosodic hierarchy. During this presentation, I will try to show :
- how these constituents differ from syntactic one;In a second section, I will show that the existence of these phonological constituents might be questioned. Various French phenomena that are difficult to explain within Prosodic Structure Theory will be used to illustrate this point :
- how the syntax / phonology interface is conceived within Prosodic structure Theory.
(i) the prosodic structure and the accentuation of
In , the sentences have the same surface syntactic structure, but their prosodic structures and stress patterns are different.
(ii) the prosodic behaviour of French clitics.
[1a] (Jean-Francois)? (a apercu) ? (les enfants) ? (de mon voisin) ?
[1b] (Jean-Francois) ? (a vu le fils) ? (de mon voisin) ?
In a third section, I shall explain how accentuation and cliticization phenomena can be analysed in a prosodic model in which the prosodic structure emerge at the surface level from two independently motivated types of constraints : alignment constraints and metrical constraints. The alignment constraints account for the alignment between prominent positions (stressed syllables) and syntactic constituents. The metrical constraints will guarantee the well-formedness of the rhythmic pattern. In this model,
- the rhythmic patterns are generated in a way that is comparable to what is proposed in grid-only frameworks (see, among others, Selkirk (1984), Laks (1997)).
- the interface between the phonological and the syntactic modules is not conceived as serial. The two modules might apply in parallel.
Delais-Roussarie, E., (2000). 'Vers une nouvelle approche de la structure prosodique' in Langue Francaise 126, B. Laks (ed) Ou en est la phonologie du Francais?. Larousse, Paris.
Delais-Roussarie, E. (1999). 'Accentuation et realisation des clitiques en Francais' in Cahiers de Grammaire, 24, J. Durand and C. Lyche (eds), Phonologie: Throrie et Variation, Toulouse.
Laks, B. (1997). Phonologie accentuelle: metrique, autosegmentalite et constituance. Collection Sciences du Langage, Editions du CNRS, Paris
Selkirk, E. O. (1984). Phonology and Syntax : The relation between sound and structure, MIT Press
Selkirk, E. O. (1986). 'On derived Domains in Sentence
Yearbook 3, pp 371-405
Syntactic vs. Phonological Displacement:
Some Empirical Considerations
Geoffrey Poole (University of Newcastle-upon-Tyne)
One seemingly unavoidable property of natural language is "displacement" – the fact that some elements are pronounced in one place but are interpreted in another. Within the model of grammar assumed in Chomskyian syntax, structure is first built up from the Numeration to Spell-Out. Spell-Out engages the phonological computation, ultimately leading to Phonological Form, a representation of pronunciation. The syntactic derivation continues to Logical Form, the interface with the conceptual-intensional systems.
Given this architecture, there are in principle two different places where a given "displacement" could have occurred. The first is between the Numeration and Spell-Out. That is, in the overt syntax. The second is between Spell-Out and PF, as part of the phonological computation.
The question of whether a given displacement is, broadly speaking, syntactic or phonological is an entirely empirical one. To the extent that the phenomenon in question displays characteristics associated with syntactic displacement, and to the extent that the best explanation for the phenomenon invokes syntactic concepts (reference to phrase-structure, syntactic constraints, etc.), then we are justified in claiming that the displacement is syntactic. To the extent that the phenomenon displays an affinity with phonological operations and processes, and to the extent that the best explanation of the phenomenon makes reference to phonological concepts (prosodic boundaries, intonational contours, etc.), then we are justified in claiming that the displacement is phonological. In this paper, I examine Stylistic Fronting (SF) in Icelandic with a view toward determining whether this displacement is syntactic or phonological.
SF is exemplified by word order contrasts such as (1), where (1a) represents the canonical word order, and (1b) the order after SF has applied.
a. Žetta er tilboš sem er ekki haegt aš hafna
This is an offer that is not possible to reject
b. Žetta er tilboš sem ekki er haegt aš hafna
This is an offer that not is possible to reject
There is no meaning difference or focalization associated with SF and its application is entirely optional. (1b) is simply considered a stylistic variant of (1a).
Previous analyses (e.g., Maling 1990 and Holmberg 2000, among
have taken SF to instantiate leftward, syntactic movement of elements
as negation and predicate adjectives. In this paper, however,
that SF is correctly analyzed as rightward, prosodic movement of the
verb. I show that the previous syntactic accounts of SF fail
explanatory accounts for the constraints observed by Maling
I then show that Hale's (1996) "prosodic flip" account of
effects with respect to Vedic Sanskrit clitics, in which a change in
order arises from a phonological restructuring of a prosodically
element, can be naturally extended to SF. I show that
properties, its semantic vacuity, optionality, and Maling's "subject
and locality constraints, follow immediately and in a natural fashion
the proposed analysis.