This thesis develops a new approach to the syntax–prosody interface and establishes the integration of the phonological module into Lexical Functional Grammar (LFG). LFG is a modular grammar theory, which (among other questions) is interested in the relation between form and meaning, i.e., between what is said/perceived and what is intended/understood. An important factor with respect to this question is the distinction between two perspectives that are essential for the communication between speaker and listener: 1) comprehension, which discusses the question as to how information from a concrete speech signal influences syntactic phrasing and with it the fundamental ‘understanding’ of what is being said. And 2) production, which is concerned with the question how the speaker’s intention is transformed into an utterance. The focus in this thesis is on a specific fragment in this larger model of communication: the syntax–prosody interface. Given a concrete speech signal, the prosodic grouping of its elements is, on the one hand, related to the language’s internal syntactic structuring: Prosodic phrasing can influence syntactic phrasing in cases of syntactically ambiguous constructions, and syntax also determines prosodic phrasing in that the grouping of prosodic structure to a certain extent reflects (and is thus in part determined by) syntactic structure. However, on the other hand, the edges of syntactic and prosodic constituents are frequently incongruent. One possible reason for this non-isomorphism between syntactic and prosodic phrasing is the rephrasing of prosodically unstressed material. For example, in some languages, function words can be prosodically phrased with a preceding (stressed) element, but are, at the same time, syntactically phrased with a following syntactic head. Such incongruencies between the syntactic and the phonological module are also frequently found in another group of elements, namely prosodically deficient clitics, whose syntactic and prosodic associations are not necessarily congruent and which can, under some circumstances, even change their position in the clause as a result of specific prosodic requirements. From these observations it can be concluded that prosodic phrasing cannot be solely determined by syntactic phrasing. Instead, processes of prosodic restructuring have to be assumed independently of the syntax–prosody interface. The resulting underlying research question in this thesis is how this tension between intermodular communication and frequent non-isomorphism between syntactic and prosodic struc-
100 Figures and Tables
Figure 2.1: C-structure and f-structure representation of Frida sneezed.
Table 2.1: A short comparison of align and wrap.
Figure 2.11: The determination of prosodic constituency as assumed in this thesis.
Figure 2.12: Violations of the Strict Layer Hypothesis.
Figure 2.13: The end-based approach.
Figure 2.15: P-structure representation in Butt and King (1998, simplified).
Figure 2.16: The prosody–syntax interface in Mycock (2006, 81).
Figure 2.17: Prosodically bracketed input to syntactic structure (Bögel et al. 2009).
Figure 2.18: The string at the heart of the grammar (Dalrymple and Mycock 2011).
Figure 2.19: The prosody–syntax interface (Dalrymple and Mycock 2011, ex.11).
Figure 2.2: Association of the lexical information and c-structure.
Table 2.2: Lexical entry for university (Dalrymple and Mycock 2011).
Figure 2.20: The prosody-syntax interface as proposed in Mycock and Lowe (2013).
Figure 2.3: Relating c-structure and f-structure of Frida sneezed.
Figure 2.4: The parallel projection architecture as represented by Asudeh (2006).
Figure 2.5: The T-model as proposed in Chomsky (1981).
Figure 2.7: The Parallel Architecture as proposed by Jackendoff (2002).
Figure 2.9: The Prosodic Hierarchy according to Selkirk (1978).
Figure 3.1: C- and p-structure representation of sher e panjaab ‘the lion of Punjab’.
Figure 3.10: The p-diagram and the P(ause)_duration dimension.
Table 3.1: Lexical entries for Japanese and Amra.
Figure 3.11: Speech signal for Ravi and Amra or Karla.
Figure 3.12: P-diagram for the speech signal in Figure 3.11.
Figure 3.13: Lexical entry for university (Dalrymple and Mycock 2011).
Figure 3.14: Form stratum as proposed by Levelt et al. (1999)
Figure 3.15: C-structure representations for Ravi and Amra or Karla.
Figure 3.16: Speech signal options for Ravi and Amra or Karla.
Figure 3.17: The prosody–syntax interface in LFG (first version).
Figure 3.18: The prosody–syntax interface in LFG (second version).
Figure 3.19: Fragment of the p-diagram for Ravi ∧ (Amra ∨ Karla).
Figure 3.2: Assignment of tune to ‘i’ means ‘insert’ (O’Connor 2004).
Table 3.2: Abbreviated lexical entry for übersetzen ‘to translate’/‘to cross over’.
Figure 3.20: Ravi ∧ (Amra ∨ Karla): at the prosody–syntax interface.
Figure 3.21: A prototypical dative.
Figure 3.22: A prototypical genitive.
Figure 3.23: P-diagram for the speech signal given in Figure 3.21.
Figure 3.3: Simplified p-structure as proposed by Mycock and Lowe (2013).
Table 3.3: Lexical alignment: the transfer of vocabulary.
Figure 3.4: The Prosodic Hierarchy (Selkirk 1978).
Table 3.4: The German determiner system.
Figure 3.5: Some phrasing possibilities for John went to school.
Table 3.5: Lexical entries for der and Partner.
Figure 3.6: A possible AVM-representation of [æn@] ‘Anna’.
Figure 3.7: Speech signal for Ravi.
Figure 3.8: The p-diagram for the acoustic signal for [ravi] (Figure 3.7).
Figure 3.9: The p-diagram of [ravi]+[pause].
Figure 4.1: The three-way distinction of clitics proposed in Anderson (2005).
Table 4.1: Postlexical phonological rules/constraints and Swabian clitics.
Figure 4.2: Architecture generally assumed in (post-)lexical phonology.
Table 4.2: Part of the lexical entry for the Swabian 1SgNom pronoun.
Figure 4.3: The position of the lexicon as proposed in this thesis.
Table 4.3: P-form of the Swabian 1SgNom pronoun.
Figure 4.4: Syllabification, subject drop, and invalid syllable structures.
Table 4.4: P-form entries corresponding to example (55).
Figure 4.5: Prosodic grouping of 3SgFem pronouns (Hall 1999).
Table 4.5: Phonological rule annotations used in this dissertation
Figure 4.6: Existential constraint and inside-out functional uncertainty.
Table 4.6: An abstract example for Optimality Theory.
Figure 4.7: Transfer of vocabulary: metrical, segmental, and prosodic information.
Figure 4.8: Postlexical phonological rules applied to [vo: =@ =s@].
Figure 4.9: Two possible integrations of OT at the prosody-syntax interface in LFG.
Table 5.1: Paradigm for the Tagalog prefix um-.
Figure 5.1: Two syntactic nodes corresponding to one lexical exponent.
Figure 5.10: Speech signal for e=kótú=n]rel yo " )ι‘... (they) called def’ ((75)).
Figure 5.11: Arrangement of modules as assumed in this thesis.
Figure 5.12: Transfer of vocabulary: ÓmÓ yo OśO↓Ol ‘The child jumped’.
Figure 5.13: Transfer of structure: ÓmÓ yo OśO↓Ol ‘The child jumped’.
Figure 5.14: The micro-components of p-structure.
Figure 5.15: The preliminary (left) and final p-diagram (right) of example (76).
Figure 5.16: The complete syntax-prosody interface in LFG.
Figure 5.2: Lexical Sharing tree (Wescoat 2005, ex.22).
Table 5.2: Lexical entries for ùgó ‘vulture’, úgò ‘butterfly’ and úg↓o ‘a kind of stew’.
Figure 5.3: Lexical Sharing applied to beG-al-le (watch-futII=3Sg).
Table 5.3: Endoclisis in Degema
Table 5.4: Paradigm of the factative clitic in Degema.
Figure 5.5: Partial syntactic tree and architecture for s-string = p-string.
Table 5.5: Structure transfer and prosodic rephrasing in cases of non-isomorphism.
Figure 5.6: Partial syntactic tree and architecture for s-string 6= p-string.
Figure 5.7: Speech signal for úg↓o ‘a kind of stew’ and ùgó ‘vulture’ .
Figure 5.8: p-diagram representations of ugo: ‘stew’ (left) vs. ‘vulture’ (right).
Figure 5.9: Speech signal for (73): ‘The girl jumped and danced.’
Figure 6.1: Postlexical prosodic inversion of a syntactically initial clitic.
Table 6.1: Pashto second position clitics (Tegey 1977, 81).
Figure 6.10: C- and f-structure representation of wα ye xla ‘Buy it’.
Figure 6.11: Transfer of structure and vocabulary: ye w@axle ‘Buy it’.
Figure 6.2: Speech signal for (93a) and (93b): bαylod@+me.
Table 6.2: Tegey’s approach: vowel coalescence before 2P clitic placement.
Table 6.3: The generation of vowel coalescence according to Kaisse (1985, 142).
Figure 6.4: The imperfective and the perfective aspect (Roberts 2000, 51, modified).
Table 6.4: Lexical entries as proposed by Lowe (2016, 37, ex. 35) for t.akwαh@.
Figure 6.5: Syntactic generation of example (98) (Roberts 2000, 58, modified).
Table 6.5: A selection of possible linear orders of elements within the verbal complex.
Figure 6.6: Analysis of (103) as proposed by Lowe (2016, 39, (40) (modified)).
Table 6.6: Postlexical phonological processes applied to example (125).
Figure 6.7: C-structure representations for (107a) and (107b).
Table 6.7: Verification of the predictions made by (128) and (129).
Figure 6.8: Abstract representation of the Pashto clause.
Table 6.8: Lexical entries for example (130).
Figure 6.9: Prosodic domains and vowel harmony in examples (121) and (122a).
Figure 7.1: The prosody–syntax interface in LFG (final version).
Download Full PDF Version (Non-Commercial Use)