Current Research Grants
Year Awarded 2019 - 2021
Year Awarded | 2018
- Title: Modality across Categories: Modal Indefinites and the Projection of Possibilities.
- Investigators: Luis Alonso-Ovalle (PI), Junko Shimoyama, Bernhard Schwarz,
- Duration of Grant: 2018-2022
- Granting Agency: SSHRC Insight Grant
Project Description: This project seeks to contribute to the development of a cross-categorial theory of modality by bringing together two domains of inquiry with limited overlap (modal verbs and modal indefinites). A basic question about modal items is how they determine the possible worlds that they range over (their 'modal domains.') Recent research on verbs maintains that grammar mediates the determination of their modal domains: According to the Modal Anchor Hypothesis (MAH) (Hacquard 2006) auxiliary verbs project their domains from a covert anaphoric element (a 'modal anchor'), whose interpretation can be grammatically restricted. We seek to explore whether the MAH extends to modal indefinites.
- Title: Testing Theories of Linguistic Productivity
- Investigator: Timothy O'Donnell.
- Duration of Grant: 2018-2020
- Granting Agency: FRQSC, Research support for new academics
Project Description: The most celebrated property of natural language is its creative potential. Language allows us to express an unbounded number of thoughts, emotions, and experiences. One foundation of this creative potential is the way in which language users can create new words by combining existing stems and affixes (e.g., "truthniness"). The phenomenon whereby speakers are able to form new expressions from an inventory of reusable, stored pieces is known in linguistics as "productivity." For any given language, however, there are many more potential productive stems, prefixes, and suffixes than are actually used in practice---each giving rise to many logically possible, but spurious ways of forming novel expressions. For example, English contains suffixes which are highly productive and generalizable (e.g., -ness; "Lady- Gagaesqueness," "pine-scentedness") and suffixes which can only be reused in specific words, and cannot be generalized (e.g., -th; "truth," "width," "warmth"). How can language users determine that "truthiness" is a possible word but "coolth" is not? What units are generalizable? What are the basic, stored building blocks?
In recent years, a number of formally explicit theories of morphological productivity have emerged in the literature (e.g., Baayen, 2007; O'Donnell, 2015; Yang, 2016). This project is aimed at developing empirical tests of theories of morphological theory by developing predictions of theoretical models and then testing these predictions with novel datasets. The project focuses on three specific classes of empirical phenomena for which various theories make differing predictions. These are: (i) "multiple generalizations" the phenomenon whereby speakers sometimes produce several forms with the same meaning; (ii) "minority generalizations" the way in which sometimes infrequent forms become the most productive; (iii) "missing generalization” the way in which certain forms which are predicted to exist in a language simply seem to be impossible.
- Title: Towards Robust Unsupervised Language Learning
- Investigator: Timothy O'Donnell
- Duration of Grant: 2018-2022
- Agency: NSERC Discovery program
Project Description: You know many things about new words, even before you have ever heard them used. For instance, you know that a new messaging app is more likely to be called a nonsense word like “jimble” than one like “jinble”. You know that if it is called “jimble,” then someone might say “jimble me a message,” but probably wouldn’t say “jimble me with a message.” You know that a person who uses the app could be called a “jimbler,” and many other facts about how this word can and can’t be used.
Understanding the computational processes that underlie this kind of generalization is a central problem for language science and is crucial to engineering computer systems capable of interacting with people. Despite tremendous progress in recent years, there remains a large gap between human-level natural language learning and the state-of-the-art in natural language processing. In particular, state-of-the-art systems tend to require supervision on specific tasks and large quantities of training data. This contrasts with human learners who acquire flexible knowledge like that above, in a largely unsupervised fashion. This project focuses on building systems that generalize in more human-like ways by incorporating theoretical insights from linguistics into probabilistic models of unsupervised language learning.
This project has two objectives. First, it will use techniques from probabilistic programming to develop systems for specifying models of unsupervised language learning and for automatically deriving inference algorithms from such specifications. This work will build on insights from mathematical linguistics as well as recent advances in probabilistic inference. In particular, it will make use of results about the formal complexity of linguistic theories as well as tools from “variational Bayesian” inference. Second, it will make use of these frameworks to explore a variety of models that encode substantive theoretical assumptions from the linguistics literature about the sound structure of words and the structure of sentences.
- Title: Prosodie, morphosyntaxe, et gestuelle des questions produites en face à face
- Investigators: Francisco Torreira
- Duration of grant: 2018-2021.
- Granting agency: FRQSC, Research support for new academics
Project Description: One of the main characteristics of verbal utterances spoken in face-to-face situations is that they necessarily consist of several parallel communicative levels: prosody (e.g. the choice of a specific intonation contour type), morphosyntax (e.g. the choice of a particular sequence of words), and gesture (e.g. raising one’s eyebrows or shrugging one’s shoulders). Prosody consists of the articulatory and acoustic aspects of speech that are not governed by the choice of words: a) the melodic modulation of the voice, b) the duration of subparts of an utterance, such as syllables, and c) the modulation of acoustic intensity, as well as d) other aspects related to phonation and articulation conditioned by the prosodic structure of the utterance. One of the most frequent roles of prosody in verbal interaction is that of signaling the communicative attitude of the speaker. More precisely, the expression of the questioning character of an utterance through prosodic means is particularly recurrent across languages. In French, for instance, final rising intonation in an utterance such as “Il pleut” ‘it is raining’ can signal a question as opposed to an assertion in specific contexts. However, prosody is not the only means of distinguishing questions from assertions. In French, absolute questions can be signaled by morphosyntactic formats (e.g. “Pleut-il ?”), or through gesture (e.g. raising one’s eyebrows).
The goal of this project is to gain a better understanding of how prosody, morphosyntax, and gesture are integrated during speech production and comprehension in the case of French questions. First, we will carry out a quantitative and qualitative analysis of observational data. In order to do this, we will collect a new corpus of Montreal French comparable to the Nijmegen Corpus of Casual French, a corpus of conversational Parisian French that we have collected in the past. Second, based on our initial observations, we will formulate a series of hypotheses related to the form and function of the questioning utterances present in the two corpora. Third, we will carry out production and perception psycholinguistic experiments in order to test relevant hypotheses. We will pay particular attention to the existence of prosodic, morphosyntactic, and gestural patterns associated to different types of questions, as is the case for several languages such as English, Spanish, or Italian. The use of conversations held by speakers of Montreal French will allow us to determine the role of prosody and gesture in two varieties of French differing in their use of morphosyntax, since, in conversational language, Quebec French appears to use morphosyntactic formats (“Pleut-il ?”) more often than Parisian French. More generally, the results of this project will allow us to better understand how speech acts are encoded and interpreted in human verbal interaction.
- Title: Three Dimensions of Sentence Prosody
- Investigator: Michael Wagner
- Duration of Grant: 2018-2023
- Granting Agency: NSERC Discovery Grant
Project Description: Speech synthesis has come a long way, today's computers can produce words and sentences highly accurately, but if a computer reads out a novel text, the result is still often hard to listen to and at times comically incoherent. A major reason is our lack of understanding of sentence prosody. Sentence prosody can be characterized as all those acoustic properties of an utterance that are not a function of the words it contains, but rather are due to other factors: Intonation (e.g., falling intonation on assertions vs. rising intonation on questions), Phrasing (the way the words in a sentence group together, for example due to their syntactic constituent structure), and Prominence (for example highly predictable or discourse given words are often reduced, while contrastive information is boosted). A first step to a better understanding of sentence prosody is to disentangle the three mentioned dimensions in the signal. The syntactic and semantic functions (type of speech act, constituency, contrast) are orthogonal to each other, but whether their prosodic correlates (tune, phrasing, prominence) are remains controversial. This project consists of a series of production and perception experiments designed to establish the true interactions, and develops a more appropriate representational model. We make use of novel tools (e.g, The Montreal Forced Aligner, developed in my lab, Mcauliffe et al. 2017, Interspeech) and statistical modelling, to develop models that can disentangle the dimensions, just as humans do when processing speech.
Year Awarded | 2017
- Title: Agreement and anti-agreement across languages
- Investigator: Jessica Coon
- Duration of grant: 2017–2022
- Granting agency: SSHRC Insight Grant
Project description: This project investigates agreement as well as the absence of otherwise-expected agreement in two unrelated languages: Chuj (a Mayan language of Guatemala) and Kabyle (an Amazigh or Berber language of Algeria). This project will contribute to theories of grammatical agreement through a careful examination of when agreement fails. The project has three major objectives: (1) theoretical research and the advancement of linguistic theory; (2) documentation of under-studied languages through original fieldwork; and (3) training students and native-speaker linguists in linguistic theory and documentation.
- Title: SPeech Across Dialects of English (SPADE): large-scale digital analysis of a spoken language across space and time
- Investigator: Morgan Sonderegger (Canada PI), Jane Stuart-Smith (UK PI), Jeff Mielke (US PI)
- Duration of grant: 2017-2020
- Funding agency and program: Trans-Atlantic Platform Digging Into Data (Canada funding from SSHRC & NSERC)
Project description: Historically, linguistic research has tended to carry out fine-grained analysis of a few aspects of speech from one or a few languages or dialects. The current scale of speech research studies has shaped our understanding of spoken language and the kinds of questions that we ask. Today, speech research is entering its own ‘big data’ revolution – massive digital collections of transcribed speech are available from many different languages, gathered for many different purposes. This project aims to develop and apply user-friendly software for large-scale speech analysis of existing public and private English speech datasets, and to understand how English speech has changed over time and space. The datasets are comprised of both Old World (British Isles) and New World (North American) English across an effective time span of over 100 years. See the project's web site for more information.
- Title: Uncovering the structure and sources of speech variability through large-scale studies
- Investigator: Morgan Sonderegger (PI)
- Duration of grant: 2017-2022
- Funding agency and program: SSHRC Insight Grant
Project description: Anyone who has tried to learn a foreign language or understand an unfamiliar accent knows that speech sounds are subject to massive variability, due to many factors. This project investigates the structure of speech variability (along what dimensions it occurs) and its sources (e.g. in physiology, cognitive biases), within and across languages. Understanding the structure and sources of variability is important for understanding fundamental aspects of human communication, such as speech perception and language change, as well as for practical applications, such as developing speech technology systems. This project’s goal is to investigate the structure and sources of variability within and across languages in realization of stop consonants (e.g. /p/, /b/), through large-scale cross-linguistic studies, enabled by computational and quantitative methods.
Year Awarded | 2016
- Title: What makes a good listener? Individual differences in speech perception and their implications
- Investigator: Meghan Clayards
- Duration of grant: 2016-2020
- Granting agency: SSHRC Insight Grant
Project Description: One of the most important things we do every day is understand spoken language. We effortlessly handle variability in different talkers and contexts with more flexibility than any automatic speech recognition system. However, listeners are themselves variable. In the past 5-10 years there has been an explosion in interest in individual differences in speech perception. However, as this field is still in its infancy, research is fragmented. What is currently lacking is a theory of how individuals differ across contexts and tasks and in the skills that underlie success in challenging situations. The central goal of this project is to bring the study of individual differences to a new level: Rather than observing differences between individuals in specific cases, we will identify whether certain general perceptual strategies are systematic and reflect differences in flexibility.
- Title: Antitonicity: a study in semantic universals
- Investigator: Bernhard Schwarz (PI)
- Duration of grant: 2016-2019
- Granting agency: SSHRC Insight Grant
Project Description: A central theme in linguistic research is the investigation of language universals, properties that hold across all natural languages and may illuminate the cognitive foundations of human language. This research program pursues a cross-linguistic investigation of a notion here named antitonicity. Antitonicity mirrors monotonicity, a central notion in semantic research for decades. Yet, surprisingly, antitonicity has not previously been investigated in linguistics. Monotone operators always preserve or always reverse entailments among their arguments. Antitonicity is radical non-monotonicity: an antitone operator does not merely fail to always preserve or reverse entailments, but it never preserves or reverses entailments. Antitonicity is thereby a central element in the logical space spanned by non-montonicity. The proposed research program is a first attempt to investigate this space, starting from the working hypothesis that antitonicity is a semantic universal.
Year Awarded | 2015
- Title: Perspectives neurocognitives sur l'acquisition, la perte et le traitement du langage (Neurocognitive perspectives on the acquisition, loss and processing of language)
- Investigators: Lydia White (PI, Linguistics), Fred Genesee (Psychology), Heather Goad (Linguistics), Yuriko Oshima-Takane (Psychology), Phaedra Royle (Université de Montréal), Karsten Steinhauer (SCSD), Elin Thordardottir (SCSD)
- Duration of grant: 2015-2019
- Granting agency: FRQSC Team Grant
Project Description: The overall objective of this research program is to investigate neurocognitive underpinnings of language acquisition and use amongst learners who are bilinguals, early or late L2 learners, or learners with language impairment. The approach is interdisciplinary, embracing different theoretical and methodological perspectives, both linguistic and psycholinguistic. We measure linguistic behaviour, using off-line and on-line measures. We also use neuro-imagining methods in order to examine more directly the neural substrates implicated in–or affected by–language learning, language loss and language processing. A number of projects are planned investigating a variety of linguistic phenomena and involving comparisons between monolinguals and bilinguals, impaired and unimpaired language learners, early and late acquirers of second languages, and learners experiencing language loss at different ages.
- Title: Phonological effects on grammatical representation and processing
- Investigators: Heather Goad & Lydia White
- Duration of grant: 2015-2020
- Granting agency: SSHRC Insight Grant
Project description: Recent literature in linguistics has witnessed a growing interest in how different components of the grammar formally relate to each other. This research program explores the relationship between phonology and other domains of the grammars of second language learners and bilingual speakers. We are investigating how this relationship accounts for knowledge and use of a second language from a variety of perspectives including: (i) extending our Prosodic Transfer Hypothesis beyond production to include processing as well as various types of functional morphology in different combinations of languages; (ii) extending our work on parsing of ambiguous sentences, focusing on prosodic cues to syntactic constituency in a different range of constructions and a wider variety of languages; (iii) considering the role of prosody in determining how pronouns are (mis)interpreted; (iv) investigating prosodic realization of information structure, especially where the L1 and L2 differ as to whether certain aspects of information structure are realized prosodically or syntactically; and (v) identifying situations where the input available to learners, especially in classroom contexts, is potentially or actually misleading such that prosodic evidence might misguide the learner as to the appropriate representation for the target grammar.
- Title: Non-Canonical Relative Clauses: Universals and Variation in Compositionality
- Investigators: Junko Shimoyama (PI) and Keir Moulton (Simon Fraser University)
- Duration of grant: 2015-2019
- Granting agency: SSHRC Insight Grant
Project description: The general goal of the proposed research program is to provide a detailed and systematic cross-linguistic description and analysis of non-canonical types of clausal subordination: (i) internally headed relative clauses and (ii) pseudo-relatives. As these constructions do not have direct counterparts in English, English-centered basic compositional mechanisms simply do not seem to work for them. There have been only limited attempts in the literature to better understand their syntactic structure, semantic interpretation, and how to connect the two at the interface level. The proposed research program will pursue two specific goals: (i) we will extend the descriptive coverage of the two non-canonical clausal subordination constructions under investigation by focusing on their scopal properties and temporal/aspectual properties; and (ii) we will interpret and analyze collected data, guided by the working hypothesis that the theoretical notion of a situation plays a key role as compositional ‘glue’ in these constructions.
Previous Research Grant
Summary: s
Awarded from 2015-2018
Year Awarded | 2018
- Title: Documenting word order variation in Mayan languages: A collection of Ch'ol narratives
- Investigator: Jessica Coon
- Duration of grant: 2018
- Granting agency: National Geographic Explorers Grant
Project description: The present project will contribute to linguistic research and language conservation through the creation of an annotated corpus of Ch’ol oral tradition. This corpus will be made available online to researchers interested in Mayan languages and culture, as well as in a site designed to engage the public in issues surrounding language conservation. The creation of this corpus will further foster capacity building and collaborative research with native-speaker linguists and trainees in Mexico, the US, and Canada.
Year Awarded | 2017
- Title: Vers un outil pour l'étude de la parole humaine à grand échelle
- Investigator: Morgan Sonderegger (PI)
- Duration of grant: 2017-2018
- Funding agency and program: Ministère des Relations internationales et de la Francophonie Coopération Québec-Israël
Project Description: La capacité d’analyser facilement de vastes quantités de textes écrits à l’aide de technologies numériques est maintenant tenue pour acquis par les chercheurs en sciences humaines et sociales, ainsi que par le grand public. Par contre, la recherche sur la parole humaine a tendance à impliquer des analyses détaillées de seulement quelques-uns des aspects de la parole humaine dans peu de langues (et surtout en anglais) et son impact est limité en raison des demandes élevées de temps et de fonds monétaires nécessaires pour amasser et annoter les données langagières. Nous initions une collaboration à long terme entre les chercheurs du Québec et de l’Israël afin de développer des technologies numériques avancées (« technologies de l’information ») permettant l’étude de la parole humaine à grande échelle.
- Title: Games and probability: a new approach to antipresuppositions
- Investigator: Bernhard Schwarz (PI), Timothy O’Donnell (co-PI)
- Duration of grant: 2017-2018
- Funding agency and program: Internal Social Sciences and Humanities Development Grant
Game theoretic and probabilistic approaches have led to new insights regarding the dividing line between conventional and conversational meaning, and promise to deliver quantitative predictions about speakers' utterance choices and listeners' interpretation that can fruitfully be related to the statistically interpreted results of controlled experiments. Yet the field has only begun to fully evaluate the utility of game theoretic and probabilistic approaches in the understanding of conversational meaning. This project explores the application of game theoretic and probabilistic approaches to presuppositions and so-called antipresuppositions, that is, inference listener’s draw based on speakers’ refraining from encoding certain presuppositional content in their utterances.
Year Awarded | 2015
- Title: North American English in Film and Television: Variation and Change
- Investigators: Charles Boberg (PI)
- Duration of grant: 2015-2017
- Granting agency: SSHRC Insight Development Grants
Project description: The objective of this proposal is to create and analyze new set of impressionistic and acoustic phonetic data on variation and change in the language of North American English film and television. Both variation and change in North American English, among the general population, and film and television language have been extensively studied in the past, the former by sociolinguists and dialectologists and the latter by film critics and media studies scholars outside of Linguistics. By contrast, the interdisciplinary research proposed here will produce the first comprehensive linguistic analysis, using state-of-the-art techniques of acoustic phonetic analysis, of dialect and social variation and change over time in the language of North American film and television. The resulting data set will be parallel and complementary to the best-known data set on variation and change in North American English among the general population, presented in the Atlas of North American English(Labov, Ash and Boberg 2006). This will allow for direct comparison between how phonetic variation and change are represented in film and television and how they operate in the ‘real world’ that film and television seek to portray. Such a comparison will enable the P.I., as well as other researchers, to explore the relationship between media, or ‘public’, language, and the ‘private’ language of individuals and groups in the general population, thereby addressing questions about the potential linguistic influence of the mass media on non-media speech; the ideology of regional and social linguistic variation in North America and in the film industry in particular, including the emergence over the 20th century of ‘General American English’; and the role of language in the expression of regional and national identity in American and Canadian culture.
- Title: Chuj Electronic Database Creation: Documentation and Revitalization of a Mayan Language
- Investigator: Jessica Coon
- Duration of grant: 2015–2016
- Granting agency: SSHRC Connection Grant
Project description: This project will create an online database of the Chuj language, an endangered and under-studied Mayan language spoken in the Guatemalan highlands. This project involves the unique opportunity for a group of academics to partner with a community organization––the Comunidad Lingüística Chuj––to create an accessible database which will aid both linguistic research on the language, as well as contribute to ongoing language revitalization efforts.
Awarded from 2012-2014
Year Awarded | 2014
- Title: Variability in the voicing contrast across languages and over time: corpus and simulation studiesInvestigator
- Investigator: Morgan Sonderegger
- Duration of grant: 2014-2017
- Granting agency: SSHRC Insight Development Grant
Project description: This project addresses two longstanding questions about human language: how speech sounds are realized across the world’s languages, and why languages change over time. Both are related to a fundamental aspect of language, variability. How the "same sound" (a phoneme such as /p/ or /b/) actually sounds, in terms of the speech signal, varies greatly across languages. This project will make progress on the questions of how and why sounds vary across languages and over time, focusing on the case of voiced and voiceless consonants, by scaling up relative to previous work: mapping variability in how these sounds are produced across several languages, using automatic measurement algorithms and datasets adapted from speech technology; and using large-scale computational simulations to explain why the pronunciation of these sounds can change over time. Contributions will be made in three key areas: description of how speech sounds vary across languages, theories of why speech sounds vary over time, and training of students.
- Title: Variabilité interlinguistique et temporelle dans les contrastes sonores voisés et non-voisés: Une application à grande échelle
- Investigator: Morgan Sonderegger
- Duration of grant: 2014-2017
- Granting agency: FRQSC Établissement de Nouveaux professeurs-chercheurs
Project description: La variabilité est un aspect fondamental de la langue. Un même "son" (un phonème tel /p/), varie grandement en termes de signal sonore d'une lange à l'autre. Dans une langue, la production des sons relève d'un système complexe de plusieurs indices acoustiques. Cependant, comme la collecte et l'annotation de données sont coûteuses et prenantes, nous sommes loin d'être en mesure d'expliquer comment la prononciation d’un phonème varie d'une langue à l'autre. Même à l'intérieur d'une communauté, la prononciation des sons varie aussi avec le temps. Beaucoup d’études décrivent la trajectoire de changements de certains sons et tentent d'expliquer les raisons de ces changements. Depuis peu, il est maintenant possible de tester les explications proposées en utilisant des simulations informatiques de l'évolution de la langue à l'intérieur d'une communauté. Le projet proposé vise à mieux comprendre pourquoi les sons varient d'une langue à l'autre et avec le temps, portant une attention particulière aux consonnes voisées et non-voisées. Le projet ajoutera donc aux études antérieures: décrivant la variabilité de la production de ces sons dans plusieurs langues, utilisant des algorithmes de mesure automatiques et des bases de données adaptées à la reconnaissance automatique de la parole ainsi que des simulations informatiques à grande échelle afin d'expliquer la raison pour laquelle la prononciation de ces sons changent avec le temps. Le projet contribuera particulièrement à la description des variations en production des sons d'une langue à l'autre, et aux théories ayant pour but d'expliquer pourquoi et comment les sons varient avec le temps.
- Title: Breaking into the Acoustic Stream: The role of allophonic patterns in processing language
- Investigators: Michael Wagner, Meghan Clayards
- Duration of Grant: 2014-2017
- Granting Agency: SSHRC Insight Grant
Project Description: Understanding speech requires decoding multiple dimensions of information that are encoded in the incoming speech stream. Prosodic cues to word and constituent boundaries provide an important way for a listener to parse the signal into meaningful units, and these cues involve both segmental and super segmental changes to phonetic structure. There is, however, substantial variability in how prosody is realized, and concomitantly there is substantial variability in the realization of segmental information crucial for lexical access. Our research program aims at developing a model of how prosodic and segmental variability interact in production and how they are processed during on-line language understanding.
Year Awarded | 2013
- Title: Language Modelling Laboratory
- Investigator: Morgan Sonderegger
- Duration of grant: 2013-2018
- Granting agency: CFI John R. Evans Leaders Fund
Project description: Funding to create the Montreal Language Modeling Laboratory in the Department of Linguistics, for computational and empirical investigation of speech sounds. Research in the lab will address how and why the sounds of a language vary over time and across languages, by “scaling up” relative to previous work: using large datasets, large-scale computational simulations, and complex statistical models, to paint a fine-grained picture of phonetic and phonological variability, and ask new questions about its sources.
- Title: Modality in the Nominal Domain.
- Investigator: Luis Alonso-Ovalle (PI), Junko Shimoyama & Bernhard Schwarz.
- Duration of grant: 2013-2017
- Granting agency: SSHRC Insight Grant
Project description: Determiners ('the', 'every,' ...) and the phrases that they form with nominal expressions ('the book', 'every book,'...) refer to and quantify over entities. Since antiquity, the study of determiner phrases has centered on these two basic semantic operations. The project seeks to advance the study of determiner phrases by exploring an understudied but cross-linguistically widespread family of instances that complement reference or quantification with modal meanings of the type that auxiliary verbs like 'must', 'can' or 'should' express. Our general objective is to elucidate linguistic principles in DP interpretation and the expression of modality through a systematic cross-linguistic description and analysis of modal DPs, bringing together two classic domains of inquiry (DPs and modality) with very limited overlap up to now.
- Title: Variability in Speech: New Tools for a New Era of Language Research (Speech Learning Lab)
- Investigator: Meghan Clayards (PI), Morgan Sonderegger, Michael Wagner, Kris Onishi (Psychology) & Aparna Nadig (SCSD)
- Duration of Grant: 2013-2015
- Granting Agency: McGill Collaborative Research Development Grant
Project Description: A central puzzle for understanding natural language is how, despite vast variability in how we speak, we communicate rapidly and efficiently, adapting to each situation. This skill is one that automatic speech recognition tools so far fail to emulate. To understand how humans (and machines) cope with speech variability, we need better information about the extent and limits of variation. Significant progress has been made, however, collecting and annotating speech data have been done primarily by hand, a lengthy and expensive process that limited the questions which could be asked. Our primary objective is to develop and validate new quantitative tools to remedy the data bottleneck taking advantage of recent improvements in speech technology. The challenge is that these tools (unlike humans) do not yet generalize well to data from different contexts (e.g., speech of children, to children, or in conversational settings) which generally contain more variability and noise. Our approach is to apply specific computational tools we have developed to different data sets we have collected (including conversational speech and speech involving children), yielding proof-of-concept results.
- Title: Personne et nombre dans les langues Mi'gmaq et Kaqchikel: Conséquences pour la concordance
- Investigator: Jessica Coon
- Duration of grant: 2013–2016
- Granting agency: FQRSC Nouveaux Chercheurs
Project description: This research project seeks to document and analyze a cross-linguistically rare pattern of verbal agreement in Kaqchikel (Mayan) and Mi'gmaq (Algonquian). In certain constructions in both languages, a particular morphological slot is filled by features of either the subject or the object, depending on their relative ranking along a "prominence hierarchy." In both languages, both person and number features must be taken into account. Through investigating the interaction of person and number features in these languages, the project seeks to contribute to our understanding of agreement more generally.
- Title: Developing mobile learning applications for the Mi'gmaq language: New opportunities for language research and revitalization (Project website)
- Investigator: Jessica Coon (PI), Alan Bale (Concordia), Gail Metallic (Listuguj Education Directorate), and Michael Wagner
- Duration of grant: 2013–2016
- Granting agency: SSHRC Partnership Development Grant
Project description: The proposed project aims to contribute to existing community efforts to reverse the trend of language loss in the Listuguj Mi'gmaq First Nation and to simultaneously revitalize, document, and research the Mi'gmaq language through a unique opportunity to partner four core entities: 1) Linguists at McGill and 2) Concordia Universities; 3) the Listuguj Education Directorate (LED) and 4) a Montreal-based software company, iLanguage Lab. These four entities will work together to achieve three inter-related and mutually re-enforcing goals: 1) Teaching: increase Mi'gmaq language teaching and learning capacities in Listuguj; 2) Research: document, describe, and analyze Mi'gmaq; and 3) Technology development: develop language-learning and language database software tools for use by linguists and community members. The strength of this partnership rests on the fact that these objectives feed into each other in novel ways and will integrate Aboriginal students as researchers and contributors to knowledge on their heritage language.
- Title: Language Acquisition and Breakdown: Experiments with Quantifiers and Quantities at the Syntax-Semantics Interfac
- Investigator: Yosef Grodzinsky
- Duration of grant: 2013-2015
- Granting agency: SSHRC Insight Grant
Project description: The proposed research has three objectives: A. to provide neurolinguistic and psycholinguistic evidence that would help to adjudicate between competing linguistic analyses of certain complex natural language quantifiers. B. to better understand, and more precisely characterize, the language deficit subsequent to focal damage to (some of) the language regions in the human brain, through theory-driven experimental investigations of the various aphasic syndromes. C. to help elucidate the relation between the language deficits observed in these syndromes and the temporarily incomplete linguistic abilities observed in the normally developing child. This might bear on the well-known Regression Hypothesis, concerning the relation between language acquisition and breakdown (Jakobson, 1941; Caramazza & Zurif, 1978, passim).
Year Awarded | 2012
- Title: Variations entre langues dans la sémantique des groupes nominaux indéfinis: l’expression de l’ignorance et de l’indifference
- Investigator: Luis Alonso-Ovalle
- Duration of grant: 2012–2015
- Granting agency: FQRSC Nouveaux Chercheurs
Project description: Determining a typology of natural language quantification remains a central goal in the search of cross linguistic stable properties that can shed light on the universal properties of natural language. This project contributes to this goal by examining the interpretation of a number of modal indefinite noun phrases that express meal meanings. The project seeks to determine the source of these modal meanings, the dimensions along which they can vary from language to language, and what these two issues reveal about natural language quantification. On the theoretical side, results will contribute to our understanding of the source of implicit modal meanings, and to the computation of non-literal meanings.
- Title: A community-linguistics collaboration for revitalizing Mi'gmaq in Listuguj (Project website)
- Investigators: Jessica Coon (PI) and Michael Wagner
- Duration of grant: 2012-2013
- Granting agency: SSHRC #611-2012-0001
Project description:This project will aid efforts to revitalize the Mi'gmaq language in the community of Listuguj (Gaspé region, Québec) through a collaboration between McGill and the community. Mi'gmaq, a member of the Eastern Algonquian branch of the Algonquian language family, is in acute danger of falling out of use in Listuguj. If nothing is done, it will likely be completely gone within decades. The situation in this community is similar to that in other Mi'gmaq communities in the Northeast, as well as to other First Nations communities across Canada. Young members of the community are no longer acquiring the language at home, and the majority of fluent Mi'gmaq speakers are over the age of 50. Despite this state of affairs, there is room for optimism: there has been a recent surge of efforts to develop new strategies to teach and learn the language, and bring it back into the daily lives of the community. The present proposal aims to augment these efforts through the co-creation of knowledge between linguists and native speakers. Given the rapidly aging base of speakers, the language revitalization efforts will focus on second language learning and documentation, as well as developing teaching materials for younger speakers. A critical theme underlying these efforts will be creating a culture in which the spoken language regains a central role in the community.
- Title: The mental representation of language variation: macro- and micro-parameters (Ergativity Lab website)
- Investigator: Lisa Travis (PI) and Jessica Coon
- Duration of grant: 2012–2015
- Granting agency: SSHRC #435-2012-0882
Project description: This research project seeks to understand how language variation is encoded by examining what has been described as ergativity in the typological and syntactic literature. Within the scope of this project, we have two sets of goals. One set of goals involves extending the relevant data set to under-documented ergative languages as well as languages that have some ergative characteristics but have not traditionally been considered to be ergative. We plan to add to the existing descriptive generalizations of how ergative properties pattern cross-linguistically. We will primarily investigate closely related languages/dialects within the Austronesian and the Mayan language families. Since both families show variation in the presentation of ergative characteristics, they are appropriate laboratories for investigating the ergativity continuum. The other set of goals involves investigating the theoretical tools of the language faculty in order to shed light on the nature of parameters focusing specifically on: (i) the viability of macroparameters (ii) the restrictions on parameters, and (iii) the particular characteristics of the ergativity continuum.
Awarded from 2009-2011
Year Awarded | 2011
- Title: Foreign-a words and the third low vowel of Canadian English
- Investigator: Charles Boberg
- Duration of grant: 2011-2014
- Granting agency: SSHRC #410-2011-1409
Project description: This project examines variation in the way words borrowed from other languages are pronounced in Canadian English. In particular, it examines pronunciation of vowels spelled with the letter <a>, which has three regular phonemic values in English: /ey/ (fate); /æ/ (fat); and /ah/ (father). Previous research has shown that Canadian English favors /æ/ in these words, where other dialects favor /ah/. Whereas American English uses /ah/ in words like pasta and words like drama, British English uses /æ/ in pasta and Canadian English uses /æ/ in both. However, my most recent research on this pattern found that some Canadian productions cannot be classified as either /æ/ or /ah/, but fall somewhere in the intermediate phonetic space between these vowels. For instance, many speakers pronounce façade in a way that rhymes neither with sad nor with sod, but is halfway between those vowels. Are such productions atypical instances of either /æ/ or /ah/, or simply imitations of foreign pronunciations without any formal status in the grammar, or do they constitute a new vowel phoneme that is being created by the foreign element in the lexicon of Canadian English? I examine these questions by eliciting productions of a wide range of foreign words spelled with <a> (pasta, lava, mantra, nachos, saga, Bali, etc.) from native speakers of Canadian and American English, along with tokens of neighboring vowels, in order to assess the pattern of vowel assignments across a large data set, and the exact phonemic relations that characterize the low vowel space in Canadian and neighboring American varieties of English. The analysis involves both auditory-impressionistic judgments and acoustic measurement.
- Title: An acoustic-phonetic analysis of conversational speech: the acoustic richness of communicatively-rich versus communicatively-poor words
- Investigator: Meghan Clayards
- Duration of grant: 2011-2014
- Granting agency: FQRSC Nouveaux Chercheurs
Project description: This research investigates the acoustic characteristics of conversational style speech. Much is known about the acoustic characteristics of careful laboratory style speech. Conversational style speech is less well studied but understanding it is important for our understanding of normal spoken interaction, building and using human-machine interfaces and extracting information from spoken digital medial. One important difference between careful and conversational speech is that in careful speech all acoustic information is available and in conversational speech some words are highly compressed. The research proposed here investigates how acoustic information is altered in cases of compression. It will use scripted conversations to elicit instances of words in situations where they provide new and important information and in situations where they are informationaly redundant. The amount and type of acoustic information will be compared in these two cases. Experiments will also test whether the amount of acoustic information correlates with ease of recognition in perception.
- Title: Second language acquisition at the phonology/syntax interface
- Investigators: Lydia White and Heather Goad
- Duration of grant: 2011-2015
- Granting agency: SSHRC #410-2011-0809
Project description: This research program extends our work on the Prosodic Transfer Hypothesis (PTH). The PTH argues for a prosodic account of L2 learners' omission or mispronunciation of inflectional morphology and function words, such as tense, agreement, determiners, etc. In particular, L2 learners are claimed to have difficulties constructing prosodic representations which are disallowed in the L1. In this program, we investigate additional L1/L2 combinations and new morphological domains. A series of experiments will be conducted, investigating the performance of child and adult L2 learners, comparing spoken production of functional material with performance on a variety of other tasks, both online and offline.
Year Awarded | 2010
- Title: Degree constructions cross-linguistically: the case of Japanese
- Investigators: Junko Shimoyama (PI) and Bernhard Schwarz
- Duration of grant: 2010-2014
- Granting agency: SSHRC #410-2010-1264
Project description: This research program aims to contribute to the theory of language variation and universals in the domain of syntax-semantics through a detailed and comprehensive study of so-called degree (or comparison) constructions in Japanese. A central question we address is whether observable differences between corresponding degree constructions in Japanese and English are attributable to independent grammatical variation between the two languages, or such differences are indicative of fundamental differences between the two degree systems. The research will form a basis for further projects on broader cross-linguistic studies.
- Title: Language and Context
- Investigators: Brendan S. Gillon (PI), Alan Bale, Sabine Bergler, Michael Blome-Tillmann, Iwao Hirose, Andrew Reisner
- Duration of grant: 2010-2013
- Granting agency: SSHRC #410-2010-1254
Project description: How one formulates utterances and how one understands utterances can, and usually does, depend on context, taken in a very, very broad sense. How one understands an utterance depends not only on the words used, their literal meanings, their grammatical properties, the grammar of the utterance (morphology and syntax), how it is pronounced (phonology), but also on the situation in which it is uttered, and even on the beliefs of the utterer and the listeners. Context dependence is central to major issues hotly debated in a variety of areas of contemporary philosophy and to problems pivotal to theoretical and computational linguistics. The overall objective of the project is to bring together philosophy, on the one hand, with theoretical and computational linguistics, on the other, to obtain a rigorous understanding of linguistic context dependence. We shall then use this rigorous understanding to re-examine the philosophical debates and to improve methods in computational linguistics. More specifically, the project has two sub-objectives: to distinguish, in a principled way, between grammatically and non-grammatically based linguistic context dependence and to apply the results to address issues raised in computational linguistics and in philosophy, concentrating in the latter case on issues in decision theory, moral theory and epistemology.
Year Awarded | 2009
- Title: Effets de maturation sur l’acquisition et le traitement langagiers (Effects of maturation on the acquisition and processing of language)
- Investigators: Lydia White (PI), Fred Genesee, Heather Goad, Yuriko Oshima-Takane, Karsten Steinhauer, and Elin Thordardottir
- Duration of grant: 2009-2013
- Granting agency: FQRSC #2010-SE-130727
Project description: This research program adopts a multi-disciplinary approach, investigating the nature and extent of age effects in language acquisition. The focus is on how age differences affect linguistic representations, language development, processing and use. A number of inter-related projects are included, involving comparisons between monolingual and bilingual language learners of different ages, impaired and unimpaired language learners, early and late acquirers of second languages, and learners experiencing language loss at different ages. Parallel and complementary experiments on a variety of linguistic phenomena are conducted.
- Title: Cues to Segmenting Complex Syllables from the Speech Stream in Infancy
- Investigators: Kris Onishi & Heather Goad.
- Duration of grant: 2009-2011
- Granting agency: CRLMB New Initiative Funds
Project description: This research program experimentally examines the linguistic and statistical cues that infants use to segment the speech stream into syllables. The focus is on the syllabification of word-internal stop+liquid clusters, as these clusters are phonotactically ambiguous when viewed from a cross-linguistic perspective: languages like French syllabify them as tautosyllabic while languages like Arabic syllabify them as heterosyllabic.
- Title: McGill Syntactic Interfaces Research Group (McSIRG)
- Investigators: Lisa Travis (PI), Glyne Piggott, Bernard Schwarz, Junko Shimoyama.
- Duration of grant: 2009-2011
- Granting agency: FQRSC #2010-SE-130906
Project description: The goal of this cross-modular research group is to closely examine technical issues that arise at the two syntactic interfaces – PF and LF – when syntactic representations are interpreted by either the phonological component or the semantic component. The investigation pays particular attention to the implications to both the phonology and the semantics of expressions of the theory of cyclic Spell Out of phases (Chomsky 2001).
- Title: Interfaces in second language acquisition: accounting for the difficulties of second language learners.
- Principal investigator: Lydia White
- Duration of grant: 2008-2011
- Granting agency: SSHRC #410-2008-1001
Project description: This project investigates possible causes of non-native performance in L2 acquisition. It has recently been suggested that L2 learners have problems in integrating different kinds of grammatical knowledge, for example, syntax with discourse or pragmatic requirements, syntax with morphology, or morphology with phonology. These are areas where different components of the grammar must 'interface' with each other. Projects conducted within the program will address different interfaces, in order to establish which interfaces (and which properties within particular interfaces) are problematic for second language learners and which are not.