About GLEAMS Welcome to GLEAMS 2022! GLEAMS (the Graduate Linguistics Expo at Michigan State) is the graduate student colloquium of the Linguistics program in the Department of Linguistics, Languages, and Cultures. It is organized by the MSU Linguistics Student Organization. This workshop will showcase graduate work in phonology, phonetics, syntax, semantics, psycholinguistics, sociolinguistics, and child language acquisition, giving graduate students the opportunity to present their work to colleagues within and outside the department. This year's invited speakers are Dr. Nick Danis (Washington University in St. Louis) and Dr. Daniel Goodhue (Leibniz-Zentrum Allgemeine Sprachwissenschaft). GLEAMS 2022 takes place Oct 28th and 29th, 2022.
|
|
Event Schedule
Room: Wells Hall B342
Zoom link: https://msu.zoom.us/j/95197043427
Meeting ID: 95197043427
Passcode: LSO
Zoom link: https://msu.zoom.us/j/95197043427
Meeting ID: 95197043427
Passcode: LSO
Event |
Time |
Introduction and Welcome |
Friday 10/28 12:30pm-1:00pm |
An analysis of Khalkh Mongolian possessive markers (Yaxuan Wang) |
1:00pm-1:30pm |
Telicity in English: Behavioral Data of Event Interpretations (John Ryan) |
1:35pm-2:05pm |
Telicity in Mandarin Preschoolers (Jingying Xu) |
2:10pm-2:40pm |
Extraction is Relative: Relative Clauses in Mende (Jason Smith) |
2:45pm-3:15pm |
Coffee Break |
3:15pm-3:30pm |
Keynote: Biased questions and speech acts: Grammar and pragmatics (Dr. Daniel Goodhue) |
3:30pm-5:00pm |
Breakfast |
Saturday 10/29 10:00am-10:30am |
The Semantics of Mandarin Hai (Yunting Gu) |
10:35am-11:05am |
An introduction to the syntax and semantics of Contrarian hardly (Philips Pellino) |
11:05am-11:35am |
Lunch |
11:45am-1:00pm |
Keynote: Comparing phonological representations (Dr. Nick Danis) |
1:00pm-2:30pm |
Coffee Break |
2:30pm-2:45pm |
Sociophonetic variation and imitation in nonbinary speakers (Jack Rechsteiner) |
2:45pm-3:15pm |
Aspectual Verbs (Shannon Cousins) |
3:20pm-3:50pm |
Cue-based Memory Retrieval in Sentence Processing: Individual Differences in Cue Weighting (Jiasheng Guo) |
3:55pm-4:25pm |
The early-adolescent indexical system amid rapid sound change: How are Northern Cities Shift and Low-Back-Merger Shift vowel realizations employed in acts of stance-taking? (Adam Barnhardt) |
4:30pm-5:00pm |
Talk Summaries
Keynote : Biased questions and speech acts: Grammar and pragmatics
Dr. Daniel Goodhue
Some questions convey a bias for a particular answer. High negation questions like Isn't it Halloween? require the speaker to have a prior expectation that the positive answer is true. Rising declaratives like It's Halloween? require contextual evidence in favor of the proposition denoted by the declarative. Complicating the picture is the fact that rising declaratives can also be used as assertions, as in the classic example of Mark Liberman walking into a dentist's office and saying to the receptionist, My name is Mark Liberman/. In this talk, I will explore how the global interpretations of these kinds of utterances come about, addressing the following questions: What are the roles of the grammatical elements involved (e.g. clause type, rising intonation, preposed negation) and the meanings attached to them? What is the role of pragmatics?
Keynote : Comparing phonological representations
Dr. Nick Danis
Within and across frameworks, many different theories of phonological representations exist. The exact differences and predicted behavior between them, as in feature matrices versus autosegments (Goldsmith 1976; Hayes 1986), one geometry versus another (e.g. Halle, Vaux & Wolfe 2000), binary versus unary features (Sagey 1986), features versus gestures (Clements 1992), and so forth, are often myriad and obscure, beyond the (seemingly) obvious. Additionally, these different representations are couched in different methods of computation, from ordered rules to parallel constraints to dynamic systems of gesture. With this variety of computation, the difference between two representations in one framework may be neutralized, obscured, or amplified, in a different framework. This makes precise, formal claims about how theories differ markedly difficult.
In recent years, however, insights from logic and model theory have been used in the formal definitions and comparisons of phonological representations (Strother-Garcia 2019; Danis & Jardine 2019; Oakden 2020). Thus, a formal definition of notational equivalence among phonological representations is given: those representations whose models are bi-interpretable under some restricted type of logic (usually first- order or quantifier-free). This work argues that this bi-equivalence is not a sufficient condition for notational equivalence, and further restricts this definition in phonology to include those transductions that are also natural class preserving.
Essentially, two theories of representation are natural class preserving if each predicts the same set of extensions of segments to be natural classes under the structural and featural assumptions of each. What is surprising is not that theories predict different natural classes—this is often by design in the definition of those theories—but that two theories can be quantifier-free bi-interpretable under logical transduction, yet not be natural class preserving. From a logical point of view, the models might be equivalent, but from a linguistic point of view, the two models still make different predictions based on how phonological processes are often assumed to behave.
Two assumptions in phonology lay the groundwork for comparing theories in this way: that a representational theory of phonology organizes segments in to natural classes, and that the computational theory of phonology targets natural classes for phonological processes. Thus, even if two theories of representation are shown to be logically equivalent, different extensions of natural classes predicts differing phonological behavior. By focusing on these assumptions, and using natural classes in the representation as a proxy for how a theory of computation would operate over such representations, a notion of strong generative capacity for phonology emerges (cf. Dolatian, Rawski & Heinz 2021).
In recent years, however, insights from logic and model theory have been used in the formal definitions and comparisons of phonological representations (Strother-Garcia 2019; Danis & Jardine 2019; Oakden 2020). Thus, a formal definition of notational equivalence among phonological representations is given: those representations whose models are bi-interpretable under some restricted type of logic (usually first- order or quantifier-free). This work argues that this bi-equivalence is not a sufficient condition for notational equivalence, and further restricts this definition in phonology to include those transductions that are also natural class preserving.
Essentially, two theories of representation are natural class preserving if each predicts the same set of extensions of segments to be natural classes under the structural and featural assumptions of each. What is surprising is not that theories predict different natural classes—this is often by design in the definition of those theories—but that two theories can be quantifier-free bi-interpretable under logical transduction, yet not be natural class preserving. From a logical point of view, the models might be equivalent, but from a linguistic point of view, the two models still make different predictions based on how phonological processes are often assumed to behave.
Two assumptions in phonology lay the groundwork for comparing theories in this way: that a representational theory of phonology organizes segments in to natural classes, and that the computational theory of phonology targets natural classes for phonological processes. Thus, even if two theories of representation are shown to be logically equivalent, different extensions of natural classes predicts differing phonological behavior. By focusing on these assumptions, and using natural classes in the representation as a proxy for how a theory of computation would operate over such representations, a notion of strong generative capacity for phonology emerges (cf. Dolatian, Rawski & Heinz 2021).
Complement Control in Early Child Mandarin: Evidence from a Preferential Looking Experiment
Jingying Xu
The present study investigates Mandarin-speaking children’s early comprehension of complement control. We tested 32 Mandarin-speaking 2-year-olds in a preferential looking experiment and assessed their ability to choose the right controller of PRO in subject control and object control structures. The findings indicate that children as young as 2 years of age are sensitive to the anaphoric relation between the controller and PRO and to the structural distinctions between different types of complement control.
Sociophonetic variation and imitation in nonbinary speakers
Jack Rechsteiner
The research I am currently working on aims to examine the construction of nonbinary gender identity through language and whether language usage is influenced by the presence of shared social identities. The two specific experiments I am employing are (1) sociolinguistic interviews of nonbinary speakers as a means of observing the presence or absence of topic-based shifting and (2) a shadowing task that will allow an investigation into the effects of differing social contexts on the degree to which nonbinary speaker’s imitate a model speaker.
The interview study provides an opportunity to increase the sociolinguistic understanding of how social identity informs linguistic production by analyzing the ways that nonbinary speakers employ linguistic resources to construct gender identity through stylistic choices. The phonetic imitation study allows for an investigation of the degree to which speakers are informed and influenced by socially salient identities, even in minimally interactive conditions, and then has the potential to raise implications for what factors motivate linguistic accommodation in a general sense. Taken together, these two experiments can provide a new and interesting avenue for examining how social identity and the perception of social identity play a role in a speaker’s linguistic production.
The interview study provides an opportunity to increase the sociolinguistic understanding of how social identity informs linguistic production by analyzing the ways that nonbinary speakers employ linguistic resources to construct gender identity through stylistic choices. The phonetic imitation study allows for an investigation of the degree to which speakers are informed and influenced by socially salient identities, even in minimally interactive conditions, and then has the potential to raise implications for what factors motivate linguistic accommodation in a general sense. Taken together, these two experiments can provide a new and interesting avenue for examining how social identity and the perception of social identity play a role in a speaker’s linguistic production.
The early-adolescent indexical system amid rapid sound change: How are Northern Cities Shift and Low-Back-Merger Shift vowel realizations employed in acts of stance-taking?
Adam Barnhardt
Pre/early-adolescence constitutes the liminal space between childhood and early adulthood – between adherence to the linguistic norms of childhood, acquired primarily from caregivers, and more intense adoption of innovative forms, mirroring (older) peers (Holmes-Elliott, 2021; Labov, 1989; Sankoff, 2019). Research suggests pre/early-adolescents make indexical associations between social information and linguistic forms salient to these two life-stages – evidence for this often manifests in the deployment of linguistic style. For example, in Eckert (2011: pp. 93 - 95), a Northern Californian pre-adolescent in a popular school clique is shown to front GOAT (the innovative form) when adopting a “cool teenage persona,” and back GOAT (the conservative form) when adopting a child-like “poor little me persona.” On the other hand, recent work suggests that pre/early-adolescents often pattern similarly to adults in the associations they make with marked and unmarked variants, although the evaluative difference between the two is less for pre/early-adolescents than for adults (Vaughn & Becker, 2021). The question can be posed, then, as to what associations early-adolescents make with conservative variants they presumably index to ideas of childhood but which are also indexed to other social information by adults, and how these associations compare to those early-adolescents make with innovative variants presumably indexed to ideas of teenager-hood. This is further complicated when adult evaluations of the conservative variants partially overlap in meaning with presumed early-adolescent associations with the innovative variants. Which variants are employed when early-adolescents are stylistically performing this meaning?
Michigan provides an ideal milieu in which to pursue this question. Recent research shows that the Northern Cities Shift is receding rapidly across apparent time and being supplanted by the Low-Back- Merger Shift (Nesbitt, 2021b; Nesbitt, Wagner, & Mason, 2019). Nesbitt (2021a) also demonstrates that NCS pre-oral raised TRAP is indexed by young adults to associations of “hard working-ness” and “uneducatedness.” Given this, I aim to investigate whether/how elements of these chain-shifts are employed by early-adolescents in acts of stance-taking in line with Eckert (2011), and what the implications are for our understanding of their indexical systems. Specifically, I will evaluate whether LOT and pre-oral TRAP, two phonemes maximally differentiated between the chain-shifts, are realized systematically as Northern Cities-shifted and Low-Back-Merger-shifted in the carrying out of “childlike” (innocent, helpless, uncool) and “teenager” (mature, competent/hardworking, cool) stances, respectively, and how this is navigated alongside partially-overlapping meanings associated with pre-oral raised TRAP by adults (crucially, “hardworking”). I will evaluate speech from 1 - 3 early-adolescents collected via asynchronous self-recordings of participants responding to prompt questions. The transcripts of their speech will be coded for instances of potential relevant stance-taking, and LOT and TRAP tokens extracted, measured, and statistically analyzed to determine if/which stance-types are significant predictors of variation. I hypothesize that NCS variants will be employed in childlike stances (due to the conservative nature of participants’ childhood input), as well as in hardworking/competent stances, and that LBMS variants will be employed in teenager stances (due to participants’ contact with older peers), except those implicating hardworking-ness/competence.
Michigan provides an ideal milieu in which to pursue this question. Recent research shows that the Northern Cities Shift is receding rapidly across apparent time and being supplanted by the Low-Back- Merger Shift (Nesbitt, 2021b; Nesbitt, Wagner, & Mason, 2019). Nesbitt (2021a) also demonstrates that NCS pre-oral raised TRAP is indexed by young adults to associations of “hard working-ness” and “uneducatedness.” Given this, I aim to investigate whether/how elements of these chain-shifts are employed by early-adolescents in acts of stance-taking in line with Eckert (2011), and what the implications are for our understanding of their indexical systems. Specifically, I will evaluate whether LOT and pre-oral TRAP, two phonemes maximally differentiated between the chain-shifts, are realized systematically as Northern Cities-shifted and Low-Back-Merger-shifted in the carrying out of “childlike” (innocent, helpless, uncool) and “teenager” (mature, competent/hardworking, cool) stances, respectively, and how this is navigated alongside partially-overlapping meanings associated with pre-oral raised TRAP by adults (crucially, “hardworking”). I will evaluate speech from 1 - 3 early-adolescents collected via asynchronous self-recordings of participants responding to prompt questions. The transcripts of their speech will be coded for instances of potential relevant stance-taking, and LOT and TRAP tokens extracted, measured, and statistically analyzed to determine if/which stance-types are significant predictors of variation. I hypothesize that NCS variants will be employed in childlike stances (due to the conservative nature of participants’ childhood input), as well as in hardworking/competent stances, and that LBMS variants will be employed in teenager stances (due to participants’ contact with older peers), except those implicating hardworking-ness/competence.
Cue-based Memory Retrieval in Sentence Processing: Individual Differences in Cue Weighting
Jiasheng Guo
The cue-based retrieval theories assume that readers/listeners try to match searched nouns against several features (syntactic vs. non-syntactic features) when searching for the correct noun in their memory (e.g., *The key to the cells unsurprisingly were rusty from many years of disuse.). It is possible that retrieval cues are weighted differently in certain dependency types (e.g. reflexive-antecedent dependency). Yadav et al. (2022) propose that systematic individual-level variation in cue weighting is also noteworthy in addition to the observed average behavior. They argue that reading proficiency (approximated by reading speed) might be associated with cue weighting. Another eye-tracking study (Frazier et al., 2015) suggests that wh-filler-gap dependency might facilitate the reflexive antecedent search during sentence processing, which coincides with the notion that syntactic cues might have a higher weighting than others. This project aims to explore the individual differences by analyzing the existing empirical data with ACT-R cue-based memory retrieval model and test our hypothesis that fast and accurate readers have a higher weighting for syntactic cues during sentence comprehension. Moreover, we also plan to conduct a new study in order to validate our hypothesis, preferably with eye-tracking, self-paced reading, and incorporate comprehension questions to ensure that faster readers demonstrate correct understanding despite the grammatical illusion.
The Semantics of Mandarin Hai
Yunting Gu
The sentences with Hai can almost always be interpreted with aspectual, additive, or scalar meaning depending on intonational pattern and shared background information of the interlocutors. I argue that the Mandarin focus-sensitive operator Hai has three components: the scalar component, the additive component, and the aspectual component. The scalar component is the basis of Hai, and factors such as intonation, adjacent utterances, or assumptions of interlocutors could trigger one or more components of Hai and thus derive the meaning of Hai in its each existence.
Extraction is Relative: Relative Clauses in Mende
Jason Smith
In this talk I present the first analysis of relative clauses in Mende, an understudied Mande language spoken in Sierra Leone and Liberia. Mende is a ‘non-rigid’ SOV language, in that CP complements obligatorily follow the verb, according to Cinque’s (2005) classification. Two distinct features of relative clauses in the language are noteworthy. First, the head of the relative clause can strand the remainder of the clause in a post-verbal position, which I argue is a result of movement of the object to the left of the verb (Kayne 1994), perhaps for case. This lines up with Koopman's argument for case-driven movement in the Mande languages Vata (1984) and Bambara (1992). I briefly show further evidence for movement in Mende including quantifier stranding in both A- and A’- movement. Second, intriguingly, extraction is possible out of subject-modifying relative clauses but not out of object-modifying relative clauses. In my presentation I propose a structure for DPs and relative clauses, and further set out some initial thoughts on how relative clause structures are derived, pointing to a possible explanation for the subject non-subject asymmetry in permitting movement out of relative clause islands.
Telicity in Mandarin Preschoolers
Jingying Xu
The present study investigates the acquisition of telicity in Mandarin preschoolers. Previous literature has documented both non-adult-like but apparently contradictory interpretations of telic structures in English- and Mandarin-speaking children: while English-speaking children allow telic expressions to describe incomplete events more liberally than adults (van Hout, 1998; Ogiela, 2007), Mandarin-speaking children are overly restrictive, rejecting incomplete readings for telic expressions (Chen, 2005; Liu, 2018; Li, 2019). However, the methodological variations in previous studies have made cross-linguistic comparisons hard to make. The present study adopts the same method as used in Ogiela’s (2007) English study, i.e., a modified truth-value judgment task, to test Mandarin-speaking 3- to 6-year-olds’ interpretation of telic structures paired with complete and incomplete events. We manipulate the verb type, the determiner type, and the type of the particle -le: the verb-final -le, which is arguably analyzed in the literature as a partitive perfective operator, allowing the partial attainment of a telic event (Smith, 1999; Chief, 2007), and the sentence-final -le, which does not convey any aspectual meaning. Our results showed that, like English-speaking children, Mandarin-speaking children across all age groups accepted telic expressions for incomplete events to some extent, but unlike English-speaking children, no effect of verb type and determiner type was found in Mandarin-speaking children. There was a significant effect of the interaction between the type of -le and age. 3- to 4-year-old children allowed telic expressions for incomplete events more often in the verb-final -le condition than in the sentence-final -le condition. This suggests that the verb-final -le is a strong cue to the incomplete reading for the younger children but not for the older children.
An analysis of Khalkha Mongolian possessive markers
Yaxuan Wang
One mysterious syntactic phenomenon frequently asked in Khalkha Mongolian is why the possessive suffix, which is constrained by Condition A of binding theory, can probe a potential antecedent outside of its binding domain. I argue that binding theory alone is not sufficient to explain the linguistic facts and propose an analysis adopting the Agree operation. My analysis correctly predicts all the possible and impossible structures, with an additional hypothesis that Mongolian possessive suffixes serve as an antecedent for PROs in adjunct. If the current analysis is on the right track, one stipulation seeking an explanation is why Mongolian possessive particles are optional.
Telicity in English: Behavioral Data of Event Interpretations
John Ryan
Telicity is integral to the interpretation of event descriptions, providing events with a logical endpoint (telic) or the lack of a logical endpoint (atelic). In this study, we examine the factors which are relevant in shaping event interpretation for English speaking adults (n=118). Linguistic models of aspectual composition predict that telic interpretations arise compositionally by combining a quantity sensitive verb with a quantized object (Verkuyl, 1972; Borer, 2005). Despite this description, behavioral data from previous experiments suggest that speakers are not always categorical in their judgements and allow telic constructs to describe incomplete events (Ogiela et al. 2007, 2014; Ryan et al, 2021). That is, preliminarily it seems that more fine-grained properties of verbs and determiners contribute to the variability in judgement. This raises questions about exactly which features affect people’s judgements.
This study assesses adults’ interpretation of telic and atelic events by varying the verb type, the direct object, and the determiner of the verb object. This operates on the framework provided by Verkuyl (1972, 1989, 1993) and Krifka (1989) and takes the experimental design of Ogiela et al (2007, 2014) as a base. The experiment was carried out by means of modified truth-value judgement tasks to examine telicity interpretation. In addition to providing support to the current theoretical models of telicity, the results seem to indicate that not only do verb subtypes and direct object determiner type contribute to telic/atelic interpretation, but it seems that there are properties of the direct object itself (besides determiner type) that may contribute to telicity interpretations. This finding opens up many more questions about the acquisition of aspect and telicity among children as well as crosslinguistically.
This study assesses adults’ interpretation of telic and atelic events by varying the verb type, the direct object, and the determiner of the verb object. This operates on the framework provided by Verkuyl (1972, 1989, 1993) and Krifka (1989) and takes the experimental design of Ogiela et al (2007, 2014) as a base. The experiment was carried out by means of modified truth-value judgement tasks to examine telicity interpretation. In addition to providing support to the current theoretical models of telicity, the results seem to indicate that not only do verb subtypes and direct object determiner type contribute to telic/atelic interpretation, but it seems that there are properties of the direct object itself (besides determiner type) that may contribute to telicity interpretations. This finding opens up many more questions about the acquisition of aspect and telicity among children as well as crosslinguistically.
Aspectual verbs
Shannon Cousins
English aspectual verbs (such as begin, finish, continue, etc.) have been analyzed under multiple theoretical and experimental frameworks in recent literature. Such verbs have been shown to incur excess processing costs during online comprehension; the crux of such cost varies across accounts. This presentation highlights the differences between two main theoretical accounts, the Structured Individual Hypothesis by Piñango & Deo (2016) and the Underlying (Anti) Causation account by Morounas & Williamson (2019) and briefly discuss the theoretical and experimental predictions for each.
An introduction to the syntax and semantics of Contrarian hardly
Philips Pellino
Within the class of adverbials referred to as approximatives, there is a member, hardly, that in addition to functioning as an approximative modifier akin to barely or rarely, may also be used to express Contrariety. When employed in this specialized manner, hardly functions like a negative operator, negating the entire proposition and implicating that some Contrary proposition actually holds. Central to this Contrarian usage of hardly are several unique inferences as well as a pattern of behavior with other sentential elements that seems counterintuitive considering the adverb’s strong negative flavor. One notable example is the tendency for Contrarian hardly to anti-license Strong NPIs. This talk will introduce some aspects of the syntactic and semantic issues involved and provide a few remarks about how an explanation is being pursued.
Organizing Committee Members
Yongqing Ye
Mitch Klein
Shannon Cousins
John Ryan
Nghi Lam
Jingying Xu
Suzanne Wagner (Director of Graduate Studies)
Yongqing Ye
Mitch Klein
Shannon Cousins
John Ryan
Nghi Lam
Jingying Xu
Suzanne Wagner (Director of Graduate Studies)
