(Francis book, forthcoming from Oxford University Press)
This book examines a challenging problem at the intersection of theoretical linguistics and the psychology of language: the interpretation of gradient judgments of sentence acceptability in relation to theories of grammatical knowledge. Acceptability judgments constitute the primary source of data on which such theories have been built, despite being susceptible to various extra-grammatical factors. Through a review of experimental and corpus-based research on a variety of syntactic phenomena and an in-depth examination of two case studies, Elaine J. Francis argues for two main positions. The first is that converging evidence from online comprehension tasks, elicited production tasks, and corpora of naturally-occurring discourse can help to determine the sources of variation in acceptability judgments and to narrow down the range of plausible theoretical interpretations. The second is that the interpretation of judgment data depends crucially on the theoretical commitments and assumptions made, especially with respect to the nature of the syntax-semantics interface and the choice of either a categorical or a gradient notion of grammaticality. The theoretical frameworks considered in this book include derivational theories (e.g. Minimalism, Principles and Parameters), constraint-based theories (e.g. Sign-based Construction Grammar, Simpler Syntax), competition-based theories (e.g. Stochastic Optimality Theory, Decathlon Model), and usage-based approaches. The volume shows that while acceptability judgment data are typically compatible with the assumptions of various theoretical frameworks, some gradient phenomena are best captured within frameworks that permit soft constraints-non-categorical grammatical constraints that encode the conventional preferences of language users.
To pre-order from Oxford University Press, or to request an examination copy for teaching, click here.
The interface between syntax and information structure in second language acquisition: Evidence from the English dative alternation
(Weirick 2021 dissertation)
It has been observed that word order variation based on information structure is particularly challenging for second language learners (L2 learners) to acquire. Native-like performance is often not achieved until the L2 learner reaches ‘advanced’ or ‘native-like’ proficiency, if it is achieved at all (Hopp, 2007, 2009; Hopp et al., 2018). Two explanations for this observation are commonly proposed: one explanation claims that L2 learners have fundamentally different representations of the relevant discourse conditions that constrain the variation, due to, for example, first language (L1) transfer. An alternative account claims that advanced/near-native L2 learners do not have fundamentally different linguistic representations, but struggle to integrate multiple types of information (e.g. syntactic and pragmatic/contextual information) in real time during language processing, resulting in non-target-like performance. The current proposal addresses this debate through two experiments that test advanced/near-native L2 learners’ knowledge of the English dative alternation. The dative alternation describes a set of verbs that can occur in two closely related constructions: the double-object construction (DOC) and the prepositional construction (PC). While the two constructions are similar in meaning, they differ subtly in their information structure constraints. A third construction, typically called Heavy NP Shift (HNPS), will also be examined. The HNPS order also demonstrates specific information structure preferences but is infrequent and seems to occur with only a subset of dative verbs. Two experiments, an acceptability judgment task and a reading time task, are proposed to examine L2 learners’ knowledge of these constructions.
The effects of verb type and discourse status on relative clause extraposition in English
(Weirick and Francis, journal article in preparation)
In English, subject-modifying relative clauses are typically adjacent to their antecedent, as in (1). However, another word order in which the relative clause is extraposed to the end of the sentence is also possible, as in (2).
- Some swimmers who had no adult supervision appeared.
- Some swimmers appeared who had no adult supervision.
Using informal judgments, Guéron (1980) and Rochemont & Culicover (1990) observed that relative clause extraposition (RCE) sentences like (2) are more acceptable when they serve a presentational function denoting the appearance of some entity into the world of discourse (Guéron, 1980; Rochemont & Culicover, 1990). Acceptability judgment tasks and corpus studies have confirmed that the acceptability and use of RCE is indeed influenced by factors related to this presentational function, including the lexical semantics (appearance verb vs. other) (Walker, 2013) and the discourse status (given vs. new) (Francis & Michaelis, 2014) of the matrix verb. Rochemont & Culicover (1990) further proposed that some non-appearance verbs can be ‘coerced’ into a presentational reading when the predicate is discourse given. The purpose of this study is to investigate the prediction that RCE sentences containing non-appearance intransitive verbs become more acceptable when the predicate has already been mentioned in the preceding discourse. Using a formal acceptability judgement task, we examine the effects of verb type (appearance, non-appearance intransitive, and transitive reflexive) discourse status of the predicate (given, new) and word order (RCE, canonical) on sentence acceptability.
Effects of frequency and simplicity in L2 English causative motion production
(Zheng, Weirick, and Francis)
How do L2 speakers of English choose a linguistic form when describing a scene involving causative motion? The results of a video elicitation task showed that L2 speakers omitted Path expressions more often than L1 speakers. However, L1 and L2 English speakers did not differ in their choice of structure for expressing Cause. Both groups used the canonical transitive construction most often. Among L2 speakers, level of proficiency showed no significant effect. Our findings support the Complex Adaptive System Principles model (Hawkins & Filipović 2013) which predicts that L2 speakers prefer forms that are frequently-used, structurally simple and resemble those in their L1.
Linguistic flexibility modulates speech planning for causative motion events: A cross-linguistic study of Mandarin and English
(Zheng 2017 dissertation)
This dissertation examined cross-linguistic differences in linguistic options for expressing causative motion (e.g., The man chased the dog into the house) in two typologically distinct languages: Mandarin and English. Furthermore, and more significantly, this research examined the impact of linguistic flexibility on speakers’ utterance planning. A video elicitation task showed that, compared to English speakers, Mandarin speakers displayed greater lexical and structural flexibility in describing the path and cause components of a causative motion event. That is, even though English and Mandarin provide a similar number of options for expressing path and cause information, English speakers select the dominant option most of the time, while Mandarin speakers select the non-dominant options relatively more often. It was predicted that this linguistic flexibility places cognitive demand on speakers, leading them to select the most accessible form during speech planning. The results of two priming experiments supported this prediction. Mandarin speakers were more likely than English speakers to be influenced by the prime sentence when selecting among the different options for expressing path and cause. These results indicate that different languages show different degrees of flexibility in linguistic options for a particular semantic domain, and therefore may differ in the degree to which speakers are susceptible to priming within that domain.
Resumptive pronouns and structural complexity in Cantonese relative clauses
(Francis, Lam, Zheng, Hitz, and Matthews)
Hawkins (2004, Efficiency and Complexity in Grammars, Oxford University Press) predicts that when the grammar of a language permits the option of either a resumptive pronoun or a gap within a relative clause, resumptive pronouns should be used more often as structural complexity increases. Results of two experiments, an elicited production task and an acceptability judgment task, affirm Hawkins’ prediction for Cantonese: resumptive pronouns were used more often and rated as more acceptable as the complexity of the relative clause increased from subject to direct object to coverb object and from non-possessive to possessive. Furthermore, resumptive pronoun use was apparently not governed by categorical grammatical rules. Resumptive pronouns were sometimes used in subject relatives (contrary to the usual restriction on highest subject position) and sometimes omitted in coverb object relatives (contrary to a proposed adjunct island condition). Implications for theories of grammatical competence are considered.
This paper was published as:
Francis, Elaine J., Charles Lam, Carol Chun Zheng, John Hitz, and Stephen Matthews. 2015. Resumptive pronouns, structural complexity, and the elusive distinction between grammar and performance: evidence from Cantonese. Lingua 162: 56-81.
Effects of weight and definiteness on speakers’ choice of clausal ordering in English
(Francis and Michaelis)
Why should English speakers sometimes prefer a discontinuous structure as in (1) when an adjacent ordering can express the same meaning?
(1) Some research was conducted that refutes the existing theories with convincing evidence.
This study examines the effects of definiteness (a correlate of discourse status) and constituent length on speakers’ choice of clausal ordering. The first experiment measured structural preference in reading (Rosenbach 2005), while the second and third experiments measured structural preference in production (Stallings et al 1998). These experiments establish independent effects of length and definiteness on speakers’ ordering preference, with the exact length effect differing depending on the task. The findings here corroborate those of an earlier corpus study which examined several factors contributing the relative clause extraposition in spontaneous discourse (Francis and Michaelis 2014).
This study was published as:
Francis, Elaine J. and Laura A. Michaelis. 2017. When relative clause extraposition is the right choice, it’s easier. Language and Cognition. 9(2): 332-370.
Cartographic approach to syntax and discourse of Shanghainese sentence final particles
This Master’s Thesis project focuses on the syntactic behavior of sentence final particles (SFP’s) in Shanghainese (上海话) including: status in the lexicon,syntactic category, placement/ ordering in the numeration, possible feature inventory, position in the structure and movement, and the existence of discourse related functional heads or Speech Act shells (Haegeman 2014). Particular interest is reserved for so called ‘evidential’ particles, which encode meta-information about a speaker’s subjective attitude or presuppositions about the data contained within either a singular utterance or a section of discourse.
Syntactic representation of motion events in Cantonese, Mandarin, and Chaoshan Hua
Following the framework of Generative–constructivist grammar (Ramchand 2008), I propose a syntactic representation for motion event predicates, as well as their structural relationship with event participants, in three Chinese languages: Chaoshan (潮汕), Cantonese, and Mandarin. Argument structure can be understood via compositional derivation of event structure and hierarchical ordering of functional eventive projections via the syntactic operation “merge and move.” Along with encyclopedic concepts associated with verbal roots, this structure composes the meanings of motion event predicates. In this study, I posit that the predicate and argument structure of motion events can be derived from four layers of functional projections that license eventive interpretation: Cause, Initiate, Process Phrase and Resultaive Phrases. Data on motion event expressions in these languages support this proposal by displaying the process of complex event composition transparently in the syntactic configuration, sharing the same inventory of building blocks and subject to head-selection constraints.
Boundedness of verbal and adjectival predicates in Mandarin
This study reformulates previous boundedness account for ba-construction in Mandarin and extends the account to transitive comparatives and hypothesizes that the selection in ba-construction and transitive comparatives constrained by boundedness, where boundedness can be manifested in terms of telicity of VPs, quantization of internal arguments or measure phrases.
The experimental study aims to test the boundedness hypothesis by collecting data of speakers’ acceptability judgment and interpretation to the sentences (the latter through a picture-selection task). Participants are asked to judge the sentences by acceptability and then indicate their interpretation to the entailment of the sentences. It is predicted the bounded predicates would be rated more acceptable in certain word orders.
This generalized account shows how formal semantic properties affect syntactic selection and explains some sentences that existing accounts do not. Also, the proposal implicates a homomorphic syntax-semantics mapping across V and A categories, which is superior than category-specific theories.
Effects of Complexity and L1 Transfer on the Acceptability of Relative Clauses in English, Mandarin Chinese, and Turkish
(Hitz and Francis)
An English-language acceptability judgment task was conducted to test whether L2 learners would find ungrammatical resumptive pronouns in non-subject relative clauses more acceptable than in subject relative clauses (a complexity effect) and whether L1 Chinese speakers would find resumptive pronouns in English relative clauses more acceptable than L1 Turkish speakers (an L1 transfer effect). The results showed that both learner groups found resumptive pronouns in non-subject positions more acceptable than in subject position (unlike the native speaker control group), but that there were no differences between the two learner groups. There appeared to be some evidence for complexity, and no evidence for L1 transfer. However, two native-language acceptability judgment tasks in Turkish and Mandarin Chinese were also conducted, revealing that contrary to previous linguistic analyses, Turkish and Mandarin differed little from each other. Thus, the results of the English-language task turned out to be ambiguous with respect to complexity and transfer accounts.
This study was published as:
Hitz, John and Elaine J. Francis. 2016. On the usefulness of formal judgment tasks in syntax and in second-language research: The case of resumptive pronouns in English, Turkish, and Mandarin Chinese. Linguistics 54: 1241-1280.