Semantics, pragmatics, syntax, cognitive science, linguistic theories and grammatical architecture, language and logic, computational linguistics, psycholinguistics
PerspectivesAsh Asudeh and Gianluca Giorgolo, 2015
Ms., University of Oxford and Carleton University: [47 pages] : [ .pdf ]
An important problem in the philosophy of language and the linguistic study of meaning (semantics and pragmatics) concerns co-referential terms and substitutability in different contexts. This problem is now commonly associated with Frege (1892), and is often called 'Frege’s puzzle'. The puzzle can be presented in various ways, but its essence can be captured as follows: given two co-referential linguistic expressions, why is it that in certain linguistic contexts substitution of one expression for the other is truth-preserving, while in others it is not?
For example, given that (1) is true, since “Hesperus" and “Phosphorus" are different names for the planet Venus, how can it be that (2) can be true while (3) is false?
(1) Hesperus is Phosphorus.
(2) Kim believes that Hesperus is a planet.
(3) Kim believes that Phosphorus is a planet.
Alternatively, we could characterize the puzzle by observing that a sentence like the following can be true without entailing that Kim does not believe a tautology:
(4) Kim doesn’t believe that Hesperus is Phosphorus.
Frege’s own solution was that in addition to a reference, nominals have a sense, or 'mode of presentation', and that in certain contexts, such as those involving propositional attitudes, it is these distinct senses that block substitutability. Frege’s puzzle is thus clearly related to the problem of 'referential opacity' in the study of propositional attitudes (Quine 1953, 1956, 1960). For example, Kripke (1972, 1979, 1980) mounts an influential defence of a Millian theory of names as simply referential.
In this paper, we focus on linguistic aspects of substitutability/opacity. We take it for granted that there is an empirical phenomenon to be explained here — differing truth value judgements despite substitution of co-referential terms — and offer a formal mechanism for capturing and explaining it. We follow Saul (1997, 2007) in observing that problems of substitutability also arise in ‘simple sentences’. Our analysis captures these cases, too. Moreover, we also focus on cases of differential interpretation of the same expression (Kripke 1979, Castañeda 1989, Fiengo & May 1998). Lastly, we briefly indicate how our analysis could give insight into cases other than referential expressions, as discussed by Kripke (1979). Thus, the mechanism we propose is not only formally well-founded on advances in formal logic and theoretical computer science, it also allows us to provide the beginnings of a general formal semantics of what we might informally call perspective.
Monads as a Solution for Generalized OpacityGianluca Giorgolo and Ash Asudeh, 2014
Published version: [9 pages] : [ .pdf ]
Proceedings of the EACL 2014 Workshop on Type Theory and Natural Language Semantics (TTNLS). 19-27.
In this paper we discuss a conservative extension of the simply-typed lambda calculus in order to model a class of expressions that generalize the notion of opaque contexts. Our extension is based on previous work in the semantics of programming languages aimed at providing a mathematical characterization of computations that produce some kind of side effect (Moggi, 1989), and is based on the notion of monads, a construction in category theory that, intuitively, maps a collection of “simple” values and “simple” functions into a more complex value space, in a canonical way. The main advantages of our approach with respect to traditional analyses of opacity are the fact that we are able to explain in a uniform way a set of different but related phenomena, and that we do so in a principled way that has been shown to also explain other linguistic phenomena (Shan, 2001).
One Semiring to Rule Them AllGianluca Giorgolo and Ash Asudeh, 2014
Submitted version: [6 pages] : [ .pdf ]
In this paper we look at the well known phenomenon of conjunction fallacies (Tversky & Kahneman, 1983) from a linguistic perspective. We are particularly interested in understanding the impact of this phenomenon on our understanding of compositionality, a core assumption of contemporary linguistic theories about language meaning. We will argue that contra Hertwig, Benz, and Krauss (2008), conjunction fallacies do not arise because of the ambiguity of coordinating conjunctions such as and, but rather because there are two different strategies available for computing the composite likelihood of a collection of events. Crucially, in our analysis the two strategies are two variants of the same underlying general structure that simultaneously allows subjects 1. to reason in purely logical terms; 2. to follow the rules of probability and 3. to commit fallacies depending on the conditions under which they evaluate linguistic expressions relating uncertain events. We will explain the choice between the two strategies in terms of cognitive/computational economy and the consequences of overestimating the likelihood of an event.
Directionality and the Production of Ungrammatical SentencesAsh Asudeh, 2013
In Cristiano Chesi, ed., Special issue on Directionality of Phrase Structure Building,
Studies in Linguistics. 6: 83–106.
Submitted version: [23 pages] : [ .pdf ]
It is a puzzling fact that English speakers produce resumptive pronouns under certain circumstances, yet reject sentences containing resumptives as ungrammatical. This paper reviews a language production model that attempts to solve this problem through prioritization of local well-formedness over global well- formedness and, more specifically, considers the model’s implications for di- rectionality in grammar. In particular, English resumptive in islands imply that the unbounded dependencies in grammar are filler-driven, featurally defined at the top of the dependency, and that islands are identifiable by the grammar from outside the island, in a top-down fashion. Lexical-Functional Grammar provides the formal framework for incrementality in the production model, which is based on monotonic information growth in LFG grammars. LFG also pro- vides a formal, ‘outside-in’ theory of unbounded dependencies that treats them as filler-driven and allows island barriers to be identified from outside of the island, rather than from the inside.
Constructions with Lexical IntegrityAsh Asudeh, Mary Dalrymple and Ida Toivonen, 2013
Journal of Language Modelling. 1(1): 1–54.
Published version: [54 pages] : [ .pdf ]
Construction Grammar holds that unpredictable form-meaning combinations are not restricted in size. In particular, there may be phrases that have particular meanings that are not predictable from the words that they contain, but which are nonetheless not purely idiosyncratic. In addressing this observation, some construction grammarians have not only weakened the word/phrase distinction, but also denied the lexicon/grammar distinction. In this paper, we consider the word/phrase and lexicon/grammar distinction in light of Lexical-Functional Grammar and its Lexical Integrity Principle. We show that it is not necessary to remove the words/phrases distinction or the lexicon/grammar distinction to capture constructional effects, although we agree that there are important generalisations involving constructions of all sizes that must be captured at both syntactic and semantic levels. We use LFG’s templates, bundles of grammatical descriptions, to factor out grammatical information in such a way that it can be invoked either by words or by construction-specific phrase structure rules. Phrase structure rules that invoke specific templates are thus the equivalent of phrasal constructions in our approach, but Lexical Integrity and the separation of word and phrase are preserved. Constructional effects are captured by systematically allowing words and phrases to contribute comparable information to LFG’s level of functional structure; this is just a generalization of LFG’s usual assumption that “morphology competes with syntax” (Bresnan 2001).
The Logic of Pronominal Resumption2012 ⋅ Oxford University Press
OUP link ⋅ Amazon link
This book is a cross-linguistic investigation of resumptive pronouns and related phenomena. Pronominal resumption is the realization of the base of a syntactic dependency as a bound pronoun. Resumption occurs in unbounded dependencies, such as relative clauses and questions, and in the variety of raising known as copy raising. Processing factors may also give rise to resumption, even in environments where it does not normally occur in a given language.
A new theory of resumption is proposed that is based on two key assumptions, one theoretical and one empirical/typological. The first assumption is that natural language is resource-sensitive (the Resource Sensitivity Hypothesis); this is captured through the use of a resource logic for semantic composition. The second assumption is that resumptive pronouns are ordinary pronouns in their morphological and lexical properties, based on typologically robust observations (McCloskey's Generalization). The theory is formalized in terms of Glue Semantics for semantic composition, with a Lexical-Functional Grammar syntax.
The theory achieves a novel unification of hitherto heterogeneous resumption phenomena. It unifies two kinds of resumptive pronouns that are found in unbounded dependencies — one kind behaves syntactically like a gap, whereas the other kind does not. It also unifies resumptive pronouns in unbounded dependencies with the obligatory pronouns in copy raising. The theory also provides the basis for a new understanding of processing-based resumption, both in production and in parsing and interpretation.
Review: Language (Nicolas Guilliot)
〈M,η,★〉 Monads for conversational implicaturesGianluca Giorgolo and Ash Asudeh, 2012
Submitted version: [30 pages] : [ .pdf ]
In Ana Aguilar Guevara, Anna Chernilovskaya, and Rick Nouwen, eds.,Proceedings of Sinn und Bedeutung 16, Volume 1. MIT Working Papers in Linguistics. 265-278.
We argue for a multidimensional approach to conventional implicatures and present a compositional approach to computing conventional implicatures based on monads. We show how a monadic approach allows us to tightly control the flow of information between dimensions. Our analysis is also based on the idea of two separate stages in the computation of the meaning of a sentence, each one regulating the flow of information in a different way. Our approach improves on previous proposals by better isolating the properties of conventional implicature and by showing precisely how their computation can be related to other semantic phenomena.
Copy raising and perception
Revision of Asudeh and Toivonen (2007), Copy raising and its consequences for perceptual reports, with formal analysis.
We examine copy raising in two closely related Germanic languages, English and Swedish, and offer a formal analysis of its syntax and semantics. We develop a new event semantics analysis of copy raising. In addition to augmenting the body of empirical data on copy raising, we show that copy raising yields novel insights into a number of key theoretical issues, in particular language and perception, the theory of arguments and thematic roles, and the broader semantics of control and raising.
Flexible composition for optional and derived argumentsAsh Asudeh and Gianluca Giorgolo, 2012
[20 pages] : [ .pdf ]
In Miriam Butt and Tracy Holloway King, eds., Proceedings of the LFG12 Conference. Stanford, CA: CSLI Publications. 64–84.
A simple but insightful analysis of optional and derived arguments at the syntax–semantics interface is provided, based on established features of LFG with Glue Semantics (optionality and templates in lexical entries and flexible, resource-sensitive semantic composition).
Missing resources in a resource-sensitive semanticsGianluca Giorgolo and Ash Asudeh, 2012
[20 pages] : [ .pdf ]
In Miriam Butt and Tracy Holloway King, eds., Proceedings of the LFG12 Conference. Stanford, CA: CSLI Publications. 219–239.
In this paper, we present an investigation of the argument/adjunct distinction in the context of LFG. We focus on those cases where certain grammatical functions that qualify as arguments according to all standard tests (Needham and Toivonen, 2011) are only optionally realized. We argue for an analysis first proposed by Blom et al. (2012), and we show how we can make it work within the machinery of LFG. Our second contribution regards how we propose to interpret a specific case of optional arguments, optional objects. In this case we propose to generalize the distinction between transitive and intransitive verbs to a continuum. Purely transitive and intransitive verbs represent the extremes of the continuum. Other verbs, while leaning towards one or the other end of this spectrum, show an alternating behavior between the two extremes. We show how our first contribution is capable of accounting for these cases in terms of exceptional behavior. The key insight we present is that the verbs that exhibit the alternating behavior can best be understood as being capable of dealing with an exceptional context. In other words they display some sort of control on the way they compose with their context. This will prompt us also to rethink the place of the notion of subcategorization in the LFG architecture
Multidimensional semantics with unidimensional glue logicGianluca Giorgolo and Ash Asudeh, 2011
[20 pages] : [ .pdf ]
In Miriam Butt and Tracy Holloway King, eds., Proceedings of the LFG11 Conference. Stanford, CA: CSLI Publications. 236-256.
We present an implementation of multidimensional semantics in Glue Semantics. Our approach differs from the proposal of Arnold and Sadler (2010) by restricting multidimensionality to the meaning language and therefore avoiding the introduction of tensors in the compositional glue logic. We use a construction from category theory — monads — to create structured mappings from the algebra of unidimensional semantics to the multidimensional case. Monads have already been successfully employed in theoretical computer science to provide a denotational semantics of side effects. Here we follow the suggestion of Shan (2001) to use monads to model semantic phenomena and show how monads can be used to capture the analysis of natural language expressions like appositives and expressives. We argue that monads allow us to keep the simplicity of unidimensional composition while also allowing the ability to track multiple meaning dimensions and to control information flow between these different dimensions appropriately.
Multimodal communication in LFG: Gestures and the Correspondence ArchitectureGianluca Giorgolo and Ash Asudeh, 2011
[20 pages] : [ .pdf ]
In Miriam Butt and Tracy Holloway King, eds., Proceedings of the LFG11 Conference. Stanford, CA: CSLI Publications. 257-277.
In this paper we investigate the interaction between verbal language and the non-verbal behaviours that commonly accompany it. We focus on spontaneous hand gestures. We discuss the complex network of interactions between the two modalities and how we can model the interpretation of a multi-channel signal. We embed this model of interpretation in the LFG correspondence architecture and we show how the flow of linguistic information that characterizes the architecture can be used to make the interpretation more precise. The result is an enriched architecture in which a complex signal is first broken up into its component parts. The subcomponents are initially interpreted independently but are then fused at the end into a single meaning object. Our model can capture quite precisely the intuitive meaning associ- ated with multimodal utterances.
Local grammaticality in syntactic production
2011 ⋅ [28 pages] ⋅ © CSLI Publications ⋅ contact me for a copy
In Emily M. Bender and Jennifer E. Arnold, eds., Language from a Cognitive Perspective: Grammar, Usage, and Processing. Studies in Honor of Thomas Wasow. Stanford: CSLI Publications. 51-79
This paper considers the following question: How and why do speakers systematically produce (i.e., utter) sentences that they consider ungrammatical? The question is investigated in light of data from the production and parsing of resumptive pronouns in English, which does not have grammatically licensed resumptives in standard varieties. The paper argues that, at least in the case of English resumption, the sentences are indeed underlyingly ungrammatical, but that the language processing system nevertheless produces such sentences due to a prioritization of local well-formedness over global well-formedness.
Towards a unified theory of resumption
2011 ⋅ [66 pages] ⋅ © John Benjamins ⋅ contact me for a copy
In Alain Rouveret, ed., Resumptive Pronouns at the Interfaces. Amsterdam: John Benjamins. 121-187
This paper presents a unified theory of resumptive pronouns, based on the Resource Management Theory of Resumption (RMTR). In particular, it identifies a common basis for puzzlingly different resumptive pronouns in languages such as Irish, which show evidence of ‘base-generated’ resumptives, versus languages such as Vata, which show evidence of gap-like resumptives.
Adjacency and locality: A constraint-based analysis of complementizer-adjacent extraction
2009 ⋅ [20 pages] : [ .pdf ]
In Miriam Butt and Tracy Holloway King, eds., Proceedings of the LFG09 Conference. Stanford, CA: CSLI Publications. 106-126.
This paper provides a new explanation of phenomena related to extraction following an overt complementizer ('that-t effects'), for which the theory-neutral term 'complement-adjacent extraction' is adopted. The analysis stems from the Correspondence Architecture of Lexical-Functional Grammar, making formally explicit certain implicit, native relations of the architecture. No reference is made to traces. The key insight is that complement-adjacent extraction effects concern linear string adjacency, where the string is understood as part of the syntax-phonology interface. A new metavariable is introduced and formally defined; the metavariable identifies the next word's f-structure. A single constraint is proposed that accounts for a wide range of relevant phenomena.
Expressives and identity conditions
Christopher Potts, Ash Asudeh, Seth Cable, Yurie Hara, Eric McCready, Luis Alonso-Ovalle, Rajesh Bhatt, Christopher Davis, Angelika Kratzer, Tom Roeper and Martin Walkow, 2009
Linguistic Inquiry 40(2): 356-366
Prepublication draft [12 pages] : [ .pdf ]
We present diverse evidence for the claim of Pullum and Rawlins (2007) that expressives behave differently from descriptives in constructions that enforce a particular kind of semantic identity between elements. Our data are drawn from a wide variety of languages and construction types, and they point uniformly to a basic linguistic distinction between descriptive content and expressive content (Kaplan 1999; Potts 2007).
In Bernd Heine and Heiko Narrog, eds., The Oxford Handbook of Linguistic Analysis. Oxford: Oxford University Press. 425-458.
Lexical-Functional Grammar (LFG) is a lexicalist, declarative (non-transformational), constraint-based theory of generative grammar. LFG has a detailed, industrial-strength computational implementation. The theory has also proven useful for descriptive/documentary linguistics. The grammatical architecture of LFG, sometimes called the ‘Correspondence Architecture’, posits that different kinds of linguistic information are modelled by distinct, simultaneously present grammatical structures, each having its own formal representation. These separate structures are formally related by correspondence functions. The different grammatical structures are subject to separate principles and formal descriptions and have distinct primitives. The two core syntactic structures are constituent structure and functional structure, and they are the central focus of this chapter. Other grammatical structures that have been proposed concern argument structure, information structure, semantics and the syntax–semantics interface, prosody and the syntax–phonology interface, and the morphology–syntax interface.
Constructions with Lexical Integrity: Templates as the lexicon-syntax interface
In Miriam Butt and Tracy Holloway King, eds., Proceedings of the LFG08 Conference. Stanford, CA: CSLI Publications. 68-88.
LFG differs from Construction Grammar (CG) in assuming a strict separation between the lexicon and the syntax. The LFG architecture and the principle of Lexical Integrity dictate that fully inflected words are ‘inserted’ one by one into the c-structure, which does not seem to permit the blurring of the boundary between words and larger syntactic units that CG advocates. This paper addresses the question of how the intuitions behind constructions (in the CG sense) can be formalized within LFG, without rejection of the foundational assumptions behind the LFG framework. The key insight in our approach is the use of LFG templates (Dalrymple et al. 2004, Crouch et al. 2008) to factor out grammatical information in such a way that it can be invoked either by lexical items or by specific c-structure rules. C-structure rules that invoke specific templates are thus the equivalent of constructions in our approach, but Lexical Integrity and the separation of lexicon and syntax are preserved.
Automatic extraction of translations from web-based bilingual materials
Qibo Zhu, Diana Inkpen and Ash Asudeh, 2007
Machine Translation 21(3): 139-163.
Published version [25 pages]
MT website (SpringerLink): [ .pdf ]
Password-protected local copy ⋅ contact, © Springer: [ .pdf ]
This paper describes the framework of the StatCan Daily Translation Extraction System (SDTES), a computer system that maps and compares web-based translation texts of Statistics Canada (StatCan) news releases in the StatCan publication The Daily. The goal is to extract translations for translation memory systems, for translation terminology building, for cross-language information retrieval and for corpus-based machine translation systems. Three years of officially published statistical news release texts at http:// www.statcan.ca were collected to compose the StatCan Daily data bank. The English and French texts in this collection were roughly aligned using the Gale-Church statistical algorithm. After this, boundary markers of text segments and paragraphs were adjusted and the Gale-Church algorithm was run a second time for a more fine-grained text segment alignment. To detect misaligned areas of texts and to prevent mismatched translation pairs from being selected, key textual and structural properties of the mapped texts were automatically identified and used as anchoring features for comparison and misalignment detection. The proposed method has been tested with web-based bilingual materials from five other Canadian government websites. Results show that the SDTES model is very efficient in extracting translations from published government texts, and very accurate in identifying mismatched translations. With parameters tuned, the text-mapping part can be used to align corpus data collected from official government websites; and the text-comparing component can be applied in prepublication translation quality control and in evaluating the results of statistical machine translation systems.
Copy raising and its consequences for perceptual reports
In Annie Zaenen, Jane Simpson, Tracy Holloway King, Jane Grimshaw, Joan Maling and Chris Manning, eds., Architectures, Rules, and Preferences: Variations on Themes by Joan W. Bresnan. Stanford, CA: CSLI Publications. 49–67.
Expletives and the syntax and semantics of copy raising
In Miriam Butt and Tracy Holloway King, eds., Proceedings of the LFG06 Conference. Stanford, CA: CSLI Publications. 13-29.
We present an event semantics account of copy raising in Swedish and English. The examination of copy raising gives rise to two puzzles. We demonstrate that our event semantics analysis solves the two puzzles. We examine some challenging copy raising data from expletives, propose a solution for handling the data, and discuss consequences of the solution for the theory of expletives.
Direct compositionality and the architecture of LFG
In Miriam Butt, Mary Dalrymple, and Tracy Holloway King, eds., Intelligent Linguistic Architectures: Variations on themes by Ronald M. Kaplan. Stanford, CA: CSLI Publications. 363-387.
Response to David Adger's ‘Remarks on Minimalist feature theory and Move’
Ash Asudeh and Ida Toivonen,
Journal of Linguistics 42(3): 395-422
Published version [12 pages]
JL website (Cambridge Journals Online): [ .pdf ]
Local copy (provided by permission), © Cambridge University Press: [ .pdf ]
David Adger raises some interesting issues and makes several valuable points in his ‘Remarks on Minimalist feature theory and Move’ (henceforth MFTM), a response to our review article ‘Symptomatic imperfections’ (henceforth SI) in this journal (Asudeh & Toivonen 2006), which was in part a review of his Core syntax (Adger 2003). In this response, we address some of the points in MFTM. We would also like to set the record straight about some points in SI which we feel have been misrepresented. In several instances, MFTM argues against claims that were not made in SI. Whatever the independent merit of these arguments, we do not wish to defend viewpoints we did not propose in the first place.
Note: Remarks on Minimalist feature theory and Move by David Adger can be found here.
Ash Asudeh and Ida Toivonen, 2006
Journal of Linguistics 42(2): 395-422
Published version [28 pages]
JL website (Cambridge Journals Online): [ .pdf ]
Local copy (provided by permission), © Cambridge University Press: [ .pdf ]
Review article on:
David Adger, Core syntax: a Minimalist approach. Oxford: Oxford University Press, 2003.
Andrew Radford, Minimalist syntax: exploring the structure of English. Cambridge: Cambridge University Press, 2004.
Minimalist syntax and Core syntax are reasonably good textbooks. They should be very helpful indeed in teaching a syntax course on current Principles and Parameters Theory (P&P; Chomsky 1981) that focuses on the Minimalist Program (MP; Chomsky 1995, 2000, 2001, 2004, 2005). The books present a range of syntactic phenomena, which are for the most part discussed lucidly and illustrated by considerable relevant data. Nevertheless, the books are not pedagogically faultless and the pedagogical faults are often due to underlying theoretical problems.
Although the Minimalist Program has opened up new research avenues for Principles and Parameters Theory, it has left the analytical part of the theory in poor condition. The theoretical flaws in these textbooks are symptomatic of problems in the Minimalist Program at large and, in the bulk of this review article, we use the books to explore underlying problems with the programmatic Minimalist approach to P&P theory. Consideration of these issues reveals Minimalism to be, at heart, a kind of unification-based, lexicalist framework, but one which eschews formalisation or even just explicitness (contra Chomsky 1957) — with attendant deleterious consequences — and which refuses to give up the unnecessary mechanism of movement, even though it has arguably outgrown it.
Ash Asudeh and Mary Dalrymple,
In Keith Brown, ed., Encyclopedia of language and linguistics (2nd edition). Amsterdam: Elsevier. Vol. 2: 23-31.
Published version [9 pages]
ELL2 website (ScienceDirect): [ .pdf ]
Password-protected local copy ⋅ contact, © Elsevier: [ .pdf ]
Binding theory accounts for whether a nominal expression can, must, or must not refer to the same individual as some other nominal in the sentence. Coreference possibilities for a pronoun like him or herself are determined by binding conditions on the relation between the pronoun and its antecedent, the nominal that a pronoun depends on for its reference. Binding theory is typically formulated on some syntactic representation. Three aspects are important in determining proper and improper binding relations: the class of nominal, the syntactic domain within which binding must or must not hold, and the syntactic relation between a nominal and its potential binder.
Some notes on pseudo noun incorporation in Niuean
Control and semantic resource sensitivity
This paper examines tensions between the syntax of control and semantic resource sensitivity. Structure sharing of controller and control target leads to apparent resource deficit under certain circumstances. An analysis is presented using Glue Semantics for Lexical Functional Grammar. It demonstrates that structure sharing and resource sensitivity can be reconciled without giving up or relaxing either notion. It is shown that the analysis can handle either property or propositional denotations for controlled complements. The analysis is extended to finite controlled complements, which raise the opposing problem of resource surplus. A solution is proposed and its typological implications discussed. The syntax and semantics of control as structure sharing is compared to a recent anaphoric control analysis by Dalrymple (2001). Based on facts of exhaustive and partial control, the present analysis is argued to be superior.
Relational nouns, pronouns, and resumption
This paper presents a variable-free analysis of relational nouns in Glue Semantics, within a Lexical Functional Grammar (LFG) architecture. Relational nouns and resumptive pronouns are bound using the usual binding mechanisms of LFG. Special attention is paid to the bound readings of relational nouns, how these interact with genitives and obliques, and their behaviour with respect to scope, crossover and reconstruction. I consider a puzzle that arises regarding relational nouns and resumptive pronouns, given that relational nouns can have bound readings and resumptive pronouns are just a specific instance of bound pronouns. The puzzle is: why is it impossible for bound implicit arguments of relational nouns to be resumptive? The puzzle is highlighted by a well known variety of variable-free semantics, where pronouns and relational noun phrases are identical both in category and (base) type. I show that the puzzle also arises for an established variable-based theory. I present an analysis of resumptive pronouns that crucially treats resumptives in terms of the resource logic linear logic that underlies Glue Semantics: a resumptive pronoun is a perfectly ordinary pronoun that constitutes a surplus resource; this surplus resource requires the presence of a resumptive-licensing resource consumer, a manager resource. Manager resources properly distinguish between resumptive pronouns and bound relational nouns, based on differences between them at the level of semantic structure. The resumptive puzzle is thus solved. The paper closes by considering the solution in light of the hypothesis of direct compositionality. It is argued that a directly compositional version of the theory is possible, although perhaps not desirable. The implications for direct compositionality are considered.
Avoiding attachment ambiguities: The role of constituent ordering
Jennifer E. Arnold, Thomas Wasow, Ash Asudeh and Peter Alrenga, 2004
Journal of Memory and Language 51(1): 55-70.
Published version [16 pages]
JML website (ScienceDirect): [ .pdf ]
Password-protected local copy ⋅ contact, © Elsevier: [ .pdf ]
Three experiments investigated whether speakers use constituent ordering as a mechanism for avoiding ambiguities. In utterances like "Jane showed the letter to Mary to her mother," alternate orders would avoid the temporary PP- attachment ambiguity ("Jane showed her mother the letter to Mary," or "Jane showed to her mother the letter to Mary"). A preference judgment experiment confirmed that comprehenders prefer the latter orders for dative utterances when the former order would have contained an ambiguity. Nevertheless, speakers in two on-line production experiments showed no evidence of an ambiguity avoidance strategy. In fact, they were slightly more likely to use the former order when it was ambiguous than when it was not. Speakers’ failure to disambiguate with ordering cannot be explained by the use of other ambiguity mechanisms, like prosody. A prosodic analysis of the responses in Experiment 3 showed that while speakers generally produced prosodic patterns that were consistent with the syntactic structure, these patterns would not strongly disambiguate the PP-attachment ambiguity. We suggest that speakers do not consistently disambiguate local PP-attachment ambiguities of this type, and in particular do not use constituent ordering for this purpose. Instead, constituent ordering is driven by factors like syntactic weight and lexical bias, which may be internal to the production system.
Honorific marking: Interpreted and interpretable
A licensing theory for Finnish
A licensing theory specifies a tripartite relation between thematic roles, grammatical functions, and grammatical case. I present a licensing theory for Finnish based on recent Optimality Theory (OT) work by Kiparsky (2001). However, the account presented here uses default unification (Lascarides and Copestake 1999), rather than OT. I argue that this yields a more restrictive licensing theory that nevertheless has several of OT's hallmarks, including violable constraints, typological prediction, and emergence of the unmarked. Thus, this work makes two principal contributions: 1) it provides a licensing theory for Finnish with considerable empirical coverage, predicting the distribution of objects with various grammatical cases, as well as the distribution of VP-external ([Spec, IP]) and VP-internal subjects, and adnominal genitives; 2) it showcases default unification as an alternative to Optimality Theory.
A resource-sensitive semantics for equi and raising
Lexical Functional Grammar (LFG) analyses of equi and raising in- volve functional control equations, which structure-share the values of grammatical functions, making them token-identical at the level of f(unctional)-structure. Glue semantics for LFG works from f-structure and is resource-sensitive. This seems at rst to be incompatible with structure-sharing. It is demonstrated that this is not so. By using LFG's parallel projection architecture, the Glue analysis of equi proposed here treats the clausal equi complement as denoting a property, which has been argued for on independent grounds by Chierchia (1984). It is also shown that the same approach can provide an analysis of raising in which clausal complements of raising verbs denote propositions, as required. The anal- ysis also explains the ability of raising verbs to take expletive subjects and the inability of equi verbs to do so.
In Mary Andronis, Erin Debenport, Anne Pycha and Keiko Yoshimura (eds.), CLS 38: The main session. Chicago, IL: Chicago Linguistic Society. 31-46.
18 pages: [ .pdf | .ps.gz ]
The syntax of preverbal particles and adjunction in Irish
It is shown that five apparently irreconcilable claims about the clausal syntax of Irish can be reconciled in a natural, base-generated LFG analysis that builds on the standard LFG theory of endocentricity and coheads/extended heads, the LFG projection architecture, and Toivonen's (2001) work on non-projecting categories and c-structure adjunction. The analysis also builds on McCloskey's (1996) analysis of Irish adjunction, but does not posit complementizer lowering. The principal theoretical consequences of the analysis are 1) the reconciliation of the five claims, in particular a synthesis of McCloskey's position that the Irish preverbal particles are complementizers and Sells's (1984) position that they are head-adjoined to the verb, 2) the elaboration of Toivonen's (2001) theory of c-structure adjunction, 3) correct predictions about not only adjunction to matrix and subordinate clauses, but also adjunction to appositives.
Coordination and parallelism in Glue Semantics: Integrating discourse cohesion and the Element Constraint
We present initial work on a theory of coordination and parallelism in Glue Semantics (GLUE; Dalrymple 1999, 2001). We will explore points of convergence and divergence between our approach to coordination and similar Categorial Grammar (CG) approaches. We also compare our approach to a previous GLUE approach to coordination (Kehler et al. 1995, 1999) and argue that our approach is superior on the grounds that it preserves a very strong notion of resource-sensitivity (Dalrymple et al. 1993). We conclude by discussing parallelism in connection with the Coordinate Structure Constraint (CSC; Ross 1967). The CSC is a putatively robust condition on extraction which has been argued to be a feature of the CG approach to coordination and of other related approaches. It is standardly assumed to have two parts, the Conjunct Constraint and the Element Constraint (Grosu 1973). The Conjunct Constraint is quite robust, but the Element Constraint has been challenged repeatedly, most recently by Kehler (2002), who argues that the CSC is not a syntactic condition, but rather follows from conditions on discourse coherence and parallelism. We discuss a constraint language on the structure of GLUE derivations, and show how Kehler's theory of discourse cohesion can be related to parallelism in such derivations.
Derivational parallelism and ellipsis parallelism
Glue Semantics for HPSG
Ash Asudeh and Richard Crouch, 2002
In Frank van Eynde, Lars Hellan and Dorothee Beermann (eds.), Proceedings of the 8th. International HPSG Conference. Stanford, CA: CSLI Publications. 1-19.
19 pages: [ .pdf | .ps.gz ]
The glue approach to semantic interpretation has been developed principally for Lexical Functional Grammar. Recent work has shown how glue can be used with a variety of syntactic theories and this paper outlines how it can be applied to HPSG. As well as providing an alternative form of semantics for HPSG, we believe that the benefits of HPSG glue include the following: (1) simplification of the Semantics Principle; (2) a simple and elegant treatment of modifier scope, including empirical phenomena like quantifier scope ambiguity, the interaction of scope with raising, and recursive modification; (3) an analysis of control that handles agreement between controlled subjects and their coarguments while allowing for a property denotation for the controlled clause; (4) re-use of highly efficient techniques for semantic derivation already implemented for LFG, and which target problems of ambiguity management also addressed by Minimal Recursion Semantics.
Shape conditions and phonological context
Ash Asudeh and Ewan Klein, 2002
In Frank van Eynde, Lars Hellan and Dorothee Beermann (eds.), Proceedings of the 8th. International HPSG Conference. Stanford, CA: CSLI Publications. 20-30.
10 pages: [ .pdf | .ps.gz ]
This paper builds on Zwicky's (1986) notion of shape condition, that is, a rule that specifies the phonological shape of inflected forms "by reference to triggers at least some of which lie outside the syntactic word". Zwicky observes that "many rules traditionally classified as external sandhi rules are [shape conditions]". They are not phonological rules in the usual sense, since they only apply to specific lexical items and are active within syntactic rather than phonological domains.
Shape conditions are problematic in many standard grammar architectures. On the one hand, they seem to be constraints on lexical entries, while on the other hand, they make reference to the syntactic context. Hayes (1990) has sketched a theory of "precompiled phrasal phonology" in which allomorph choice is conditioned by syntactic frames stated in lexical entries. However, his approach is not formalized in any detail, and begs the question of how lexicalised syntactic frames fit into the overall grammar. Although they could be identified with subcategorization frames, this would make the implicit claim that the relation between a shape condition target and its triggers can be equated with the syntactic relation between a lexical head and its complement. Although this assumption holds good for the Hausa phenomena he addresses, we do not believe that it holds in general.
HPSG appears to offer a promising framework for formalizing something like Hayes' approach, but the standard machinery also makes it hard to distinguish a shape condition trigger from a complement. In order to overcome this difficulty, we develop the notion of phonological context: a feature of signs which allows us to condition allomorphic alternation in terms of (i) the phonological edges, and (ii) the syntactic properties of an expression's immediate syntactic sisters. We show how our analysis deals with four illustrative cases: the indefinite article alternation in English, syncretic liaison forms for possessive pronouns in French, Hausa verb-final vowel shortening, and soft mutation in Welsh nouns.
Probabilistic Learning Algorithms and Optimality Theory
This article provides a critical assessment of the Gradual Learning Algorithm (GLA) for probabilistic optimality-theoretic (OT) grammars proposed by Boersma and Hayes (2001). We discuss the limitations of a standard algorithm for OT learning and outline how the GLA attempts to overcome these limitations. We point out a number of serious shortcomings with the GLA: (a) A methodological problem is that the GLA has not been tested on unseen data, which is standard practice in computational language learning. (b) We provide counterexamples, that is, attested data sets that the GLA is not able to learn. (c) Essential algorithmic properties of the GLA (correctness and convergence) have not been proven formally. (d) By modeling frequency distributions in the grammar, the GLA conflates the notions of competence and performance. This leads to serious conceptual problems, as OT crucially relies on the competence/performance distinction.
Linking, optionality, and ambiguity in Marathi
In Peter Sells (ed.), Formal and empirical issues in optimality-theoretic syntax.
Stanford, CA: CSLI Publications. 257-312.
|56 pages: [ .pdf | .ps.gz ]|
|Password-protected: Please contact me|
Optimality Theory (OT) has problems with ineffability, optionality, and ambiguity, because OT grammars typically pick exactly one winner for any input. This paper take steps to accommodate ambiguity and optionality in Optimality Theory syntax. I briefly present the basic architecture of standard Optimality Theory and how it relates to ineffability, ambiguity, and optionality, and I argue that optionality and ambiguity in OT are formally the same, the first operating in production and the second in comprehension. After discussing four possible ways of capturing optionality and ambiguity in OT, I argue that Boersma's stochastic OT model provides the most satisfactory way of dealing with these two phenomena in OT. Lexical Functional Grammar (LFG) provides a formalization of the inputs, GEN, and the candidates for OT syntax. My analysis is thus formalized with a stochastic OT-LFG model.
Using this model with Dowty's theory of proto-roles, I give an analysis of optionality in linking arguments to grammatical functions in Marathi. In doing so, I will illustrate four main points. The first is that linking can be achieved with a small set of cross-linguisically plausible, violable constraints. Second, optionality can be captured in OT syntax by modifying the architecture of the theory only slightly. Third, in comprehension-directed optimization, the same OT constraints that are used to capture the linking optionality in production can also capture the resulting ambiguity in the Marathi strings that correspond to the winning candidates in production. Fourth, this OT approach to linking has interesting implications for proto-role theory.
Review of Joost Dekkers, Frank van der Leeuw, and Jeroen van de Weijer (eds.), Optimality Theory: Phonology, Syntax, and Acquisition
Experimental evidence for a predication-based Binding Theory
Ash Asudeh and Frank Keller, 2001
In Mary Andronis, Chris Ball, Heidi Elston, and Sylvain Neuvel (eds.), CLS 37: The main session. Chicago, IL: Chicago Linguistic Society. 1-14.
17 pages: [ .pdf | .ps.gz ]
Constraints on linguistic coreference: Structural vs. pragmatic factors
Binding theory is the component of grammar that regulates the interpretation of noun phrases. Certain syntactic configurations involving picture noun phrases~(PNPs) are problematic for the standard formulation of binding theory, which has prompted competing proposals for revisions of the theory. Some authors have proposed an account based on structural constraints, while others have argued that anaphors in PNPs are exempt from binding theory, but subject to pragmatic restrictions. In this paper, we present an experimental study that aims to resolve this dispute. The results show that structural factors govern the binding possibilities in PNPs, while pragmatic factors play only a limited role. However, the structural factors identified differ from the ones standardly assumed.
Argument structure and animacy restrictions on anaphora
This paper presents a theory of syntactic binding defined on the level of argument structure within the framework of Head-Driven Phrase Structure Grammar (Pollard and Sag, 1994). The particular focus is on three construction types: Super Equi-NP Deletion, Picture NPs, and anaphors in nominal specifier positions.
The case of Picture NPs especially has been vexing for syntactic theories of binding (e.g. Chomsky, 1986, Pollard and Sag, 1994). One previous solution has been to complicate binding theory to account for these cases (Chomsky, 1986). However, Pollard and Sag (1994) have shown that the resulting predictions are inaccurate, and the theory is therefore complicated with no large descriptive or explanatory gain. But, their own solution is also unsatisfying, since it treats Picture NP anaphora (and the other two cases) as being exempt from syntactic binding and subject solely to processing and discourse constraints (which they do not explicitly define).
The theory presented here extends recent work in HPSG that has moved the locus of binding theory to argument structure (ARG-ST) lists (Manning, 1996; Manning and Sag, 1998) by proposing a new syntactic constraint. This constraint crucially uses the representational power of HPSG to include a requirement of animacy on certain potential antecedents of anaphors, the Antecedent Closeness Constraint (ACC). The ACC moves the HPSG binding theory out of the realm of the strictly syntactic, capturing insights from functional accounts (e.g., Kuno, 1987) about the information needed to compute the antecedents of reflexives and reciprocals. However, the ACC is superior to existing functional accounts, because it simultaneously captures syntactic and semantic restrictions on the antecedent-anaphor relationship in an explicit and predictive manner, by using the formalism of HPSG. At the same time, it captures a range of data that have been shown to be problematic for purely syntactic theories of binding.
Functional identity and resource-sensitivity in control
Glue semantics provides a semantics for Lexical Functional Grammar (LFG) that is expressed using linear logic and provides an interpretation for the f(unctional)-structure level of syntactic representation, connecting it to the level of s(emantic)-structure in LFG's parallel projection architecture. Due to its use of linear logic for meaning assembly, Glue is resource-sensitive: semantic resources contributed by lexical entries and resulting f-structures must each be used in a successful proof exactly once. In this paper, I will examine the tension between a resource-sensitive semantics which interprets f-structures and structure-sharing in f-structures as expressed by functional control resulting from lexical functional identity equations. The empirical phenomenon I concentrate on is equi, also known as obligatory control. Although at first blush it seems that structure-sharing poses a serious problem for Glue semantics, I will show that this is not so. In fact, this tension leads to a very restrictive theory, and the analysis I present here solves several long-standing problems in the semantics of equi, by exploiting LFG's grammatical architecture.
Constraints on linguistic coreference: An experimental investigation of exempt anaphors
Previous experimental studies (Gordon and Hendrick, 1997) demonstrated that native speakers' judgments of the coreference possibilities for (non-reflexive) pronouns systematically diverge from the predictions of linguistic binding theory. We extend these results to the coreference possibilities for syntactic anaphors (reflexives). In particular, we deal with the binding of anaphors and pronouns as they occur in picture~NPs. We show, contrary to the predictions of binding theory, that in picture~NPs, anaphors are systematically preferred over pronouns. Furthermore, such anaphors can take remote binders, even when there is a local, intervening nominal. This entails that such anaphors are always exempt from binding theory, contrary to the claims in the literature. We also demonstrate that the same factors that influence the exempt status of anaphors play a role in extraction from picture~NPs. This result indicates that both phenomena should be mediated by the same grammatical representations.
Incorporation in Danish: Implications for interfaces
Anaphora and argument structure: Topics in the syntax and semantics of reflexives and reciprocals
This thesis presents an examination of the syntactic and semantic properties of natural language reflexive and reciprocal expressions, which are commonly referred to as anaphors in syntactic theory. The work is carried out in the broader context of `binding theory' and `control theory'. Binding theory has two main aspects. The first aspect is to explain the distribution of nominal expressions in natural language sentences. The second aspect is to link this syntactic representation to a theory of natural language semantics, such that the semantics is constrained by the syntactic relationships between nominals. Control theory is about the interpretation of understood (phonetically null) arguments of complements to a certain class of predicates that are characterized semantically. The syntactic framework assumed is Head-driven Phrase Structure Grammar (HPSG). In this framework, binding theory and control theory are intimately tied together through the syntax of reflexives.
The binding and control theories developed here are based on a level of representation called `argument structure'. Argument structure is construed as an abstract linking representation between the level of (lexical) semantics and the level of syntactically relevant grammatical relations. It is defined as a representation of the other expressions that a given word or phrase must combine with in order to be semantically and syntactically saturated. A strong interpretation of this definition is assumed here: all syntactically and semantically relevant arguments are present at the level of argument structure.
There are four main goals in this thesis: first, to extend and revise problematic previous versions of the theories of binding and control in HPSG; second, to examine the implications that these revisions and related phenomena have for the representation of argument structure; third, to develop a general program for the interpretation of the index notation used in binding theory; fourth, to test whether reciprocal expressions should be represented in the semantics as quantifiers or plurals.
The outline of the thesis is as follows. Chapter one presents an introduction to the issues involved and a brief discussion of the framework. The HPSG theories of binding and control are presented and formalized in chapter two. There are various problems that ensue which are discussed therein. In chapter three, an extended binding theory is presented. This theory consists of a simplified core, which solves certain of the problems noted in chapter two, and a new constraint on the distribution of anaphora called the Antecedent Closeness Constraint. The coverage of the extended binding theory is demonstrated. In chapter four, the interplay between argument structure and control theory is explored, and several new problems are discussed. A revised control theory is presented which deals with these problems, as well as the ones discussed in chapter two. The coverage of the revised control theory is presented. This includes a demonstration of the application of the Antecedent Closeness Constraint to controlled complements. The fifth chapter presents a program for interpreting indices, with specific reference to Discourse Representation Theory. The chapter ends with a discussion of logophoric reflexives in English. The sixth chapter presents a discussion of reciprocal interpretation, with specific attention paid to whether the reciprocal is a plural or a quantifier.