Lee-Ann Vidal Covas, PhD

My Portfolio

Tue, 21 Apr 2026 00:00:00 +0000

Covariation & salience in linguistic contact: a sociophonetic study of liquid variation in Boston Spanish

Sun, 21 Sep 2025 00:00:00 +0000

Click to view more.

This dissertation examines how Spanish-speaking Bostonians of Puerto Rican and Dominican origin navigate linguistic variation and lectal coherence, focusing on the role of salience in shaping the outcomes of linguistic contact. Speakers do not adopt, resist, or modify linguistic features at random—social and cognitive factors shape how variation unfolds in contact settings. While prior research has shown that some features vary in their susceptibility to change in bilingual environments, the extent to which salience influences this process remains underexplored. By analyzing both low-salience features (e.g., subject pronoun expression, subject placement, and filled pauses) and high-salience features (e.g., coda /s/ and liquid variation), this study explores how speakers engage with variation in response to social and linguistic pressures in a dynamic contact setting.

Using sociolinguistic interviews from 22 Spanish speakers in Boston, the study employs quantitative variationist methods to analyze the distribution and interaction of multiple variables. The results show that low-salience features exhibit modest, yet, measurable and systematic structural convergence with English grammatical norms, reflecting bilingual optimization strategies. In contrast, high-salience features provide speakers with greater social and stylistic flexibility, allowing them to deploy these features in ways that align with their identity, ideology, and interactional goals.

In some cases, such as with rhotics, speakers maintain a broader sense of Caribbean linguistic identity—even after extended residence in the mainland U.S.—using these features as shibboleths of national identity. For example, the lateralization of the tap continues to differentiate Puerto Rican and Dominican speech styles, serving as a salient marker of regional affiliation. However, other high-salience features, such as laterals, appear to carry less social weight, making them more susceptible to contact-induced restructuring or stylistic reinterpretation.

To examine how these patterns of variation intersect, the study incorporates a covariation analysis that investigates whether and how the five variables pattern together—particularly in relation to speakers’ exposure to English and broader contact-induced pressures. This analysis reveals a fundamental asymmetry: lower-salience features tend to shift in tandem, reflecting a broader structural realignment shaped by bilingual optimization. Higher-salience features, by contrast, do not show the same kind of coordinated movement. Instead, they remain sites of sociolinguistic agency, affording speakers the flexibility to reinforce, neutralize, or reinterpret these forms in ways that serve social and identity-related goals. These findings contribute to sociolinguistic theory by demonstrating that salience is not merely an inherent property of linguistic variables but a dynamic, ideologically mediated phenomenon—one that shapes how and why speakers retain, modify, or reconfigure particular linguistic features in language- and dialect-contact settings.

Moreover, the current research advances understanding of linguistic agency, identity, and change in multilingual communities by conceptualizing salience as a continuum rather than a dichotomy. By framing linguistic features along a salience continuum, the study reveals how different variables interact with social forces in contact settings, offering insights into the ways speakers navigate variation to manage group affiliations and respond to sociolinguistic pressures. Additionally, by incorporating liquid variation into the study of salience and covariation, this dissertation broadens the empirical and theoretical scope of the study of Spanish in the United States, shedding new light on the ways bilingual speakers actively engage with variation.

Covariation & salience in linguistic contact: a sociophonetic study of liquid variation in Boston Spanish

Wed, 07 May 2025 00:00:00 +0000

Spanish in Boston Project

Thu, 01 May 2025 00:00:00 +0000

I have previously served as a Grad Assistant and Lab Manager for Dr. Daniel Erker in the Spanish in Boston Project, where I focused on transcribing and annotating sociolinguistic interviews related to various social and linguistic phenomena.

Currently, I continue to contribute to the project through data creation, extraction, and cleaning, as well as running statistical analyses and creating visualizations.

Salient Features, Strong Identities: Social Meaning and Stylistic Choices in Boston Spanish

Tue, 03 Dec 2024 00:00:00 +0000

How Salience Influences Dialectal Persistence and Covariation: Insights from Boston's Spanish-Speaking Community

Thu, 07 Nov 2024 00:00:00 +0000

Investigating salience and ideologies in Greater Boston: A sociophonetic analysis of liquid variation

Thu, 17 Oct 2024 00:00:00 +0000

Variation, Contact, and Change in Boston Spanish: How social meaning shapes stylistic practice and bilingual optimization

Fri, 06 Sep 2024 00:00:00 +0000

Riding the Waves of Salience: Exploring Spanish Liquid Variation among Boston Spanish speaker

Thu, 04 Jan 2024 00:00:00 +0000

Making a Splash: The Salience of Liquid Variation among Boston Spanish speakers

Fri, 13 Oct 2023 00:00:00 +0000

¿Va primero el verbo? OR ¿El sujeto va primero?: Subject-verb order in Latin American Spanish

Sat, 01 Jul 2023 00:00:00 +0000

Optional Extended Content

Making a Splash: The Salience of Liquid Variation among Boston Spanish speakers

Thu, 23 Mar 2023 00:00:00 +0000

Making a Splash: The Salience of Liquid Variation among Boston Spanish speakers

Mon, 20 Mar 2023 00:00:00 +0000

¿Va primero el verbo OR El sujeto va primero?: Subject-Verb order in Latin American Spanish

Thu, 05 Jan 2023 00:00:00 +0000

Qué decimos cuando no decimos nada: Claves del cambio lingüístico inducido por contacto en las pausas llenas del español conversacional

Wed, 21 Sep 2022 00:00:00 +0000

Note

Esta es una traducción, ofrecida por el Observatorio, del original inglés remitido por los autores. This is a translation, provided by the Observatorio, of the English original submitted by the authors.

What we say when we say nothing at all: Clues to contact-induced language change in Spanish conversational pause-fillers

Tue, 20 Sep 2022 00:00:00 +0000

The Mirror Principle Violations Project

Sun, 01 Aug 2021 00:00:00 +0000

I worked as a graduate assistant on Mirror Principle Violations Project, which is one component of the SULa Lab’s research on argument structure and morphology.

My role was to survey descriptive materials for a set of under-researched languages, including Indonesian, Wichita, West Greenlandic, Yagua, and Zulu, and document how causative and applicative morphology behaves in each case.

For each language I summarized typological morphology/syntax facts, recorded which markers were used for causatives versus applicatives, checked whether the same marker could do both jobs, and noted whether the sources discussed when both markings appeared on the same verb.

For more about the lab, see the .

(Eds.) Proceedings of the 45th Annual Boston University Conference on Language Development

Sat, 01 May 2021 00:00:00 +0000

Note

Danielle Dionne and Lee-Ann Vidal-Covas served as editors of this two-volume set.

Filled pauses as a site of language contact among Spanish-speaking Bostonians

Wed, 06 Jan 2021 00:00:00 +0000

Filled pauses as a site of language contact among Spanish-speaking Bostonians

Wed, 02 Dec 2020 00:00:00 +0000

English/Spanish Contrastive Phonology & Filled Pauses

Thu, 15 Oct 2020 00:00:00 +0000

Teaching Foreign Language in the 21st Century Classroom

Wed, 01 Mar 2017 00:00:00 +0000

The Verb as a Predictor of Variable Pronominal Use in Spanish

Thu, 23 Oct 2014 00:00:00 +0000

Los efectos condicionantes del verbo en el uso variable de los pronombres personales de sujeto

Mon, 14 Jul 2014 00:00:00 +0000

Los efectos condicionantes del verbo en el uso variable de los pronombres personales de sujeto

Mon, 14 Jul 2014 00:00:00 +0000

Verb-related Constraints on Variable Pronominal Use: Do we have all the Answers?

Thu, 03 Apr 2014 00:00:00 +0000

El uso variable de los pronombres sujetos en el castellano puertorriqueño hablado en Luisiana y Puerto Rico

Thu, 27 Feb 2014 00:00:00 +0000

The Effects of Verb Semantics on Pronominal Expression: A Preliminary Study

Thu, 27 Feb 2014 00:00:00 +0000

El uso variable de los pronombres sujetos en el castellano puertorriqueño hablado en Luisiana y Puerto Rico

Fri, 20 Dec 2013 00:00:00 +0000

El uso variable de los pronombres sujetos en el castellano puertorriqueño hablado en Luisiana y Puerto Rico

Tue, 01 Oct 2013 00:00:00 +0000

Variable use of subject pronouns in Spanish

Thu, 04 Apr 2013 00:00:00 +0000

Variable use of subject pronouns in the Spanish in Louisiana

Mon, 01 Apr 2013 00:00:00 +0000

Cogito: Speech Data Annotation for Machine Learning

Mon, 01 Jan 0001 00:00:00 +0000

Overview

Worked on annotation and quality assurance for speech and language datasets used in machine learning models, focusing on improving model performance and annotation consistency.

My Role

Annotated speech data for emotion, engagement, and speech patterns
Designed and refined annotation approaches across projects
Conducted prompt engineering to improve model outputs
Tested pre-trained language models and suggested calibration improvements

Data & Workflows

Processed audio and text data for ML training pipelines
Built and validated annotated datasets for internal and external clients
Handled dynamic annotation requests across teams
Contributed to workflow improvements and QA processes

Impact

Improved annotation consistency across datasets
Contributed to higher-quality training data for ML models
Provided insights that informed model behavior and UX decisions

Covariation & Salience in Linguistic Contact

Mon, 01 Jan 0001 00:00:00 +0000

I investigate how the noticeability—or salience—of different linguistic features shapes how people use and adapt language in contexts of contact with other dialects and languages. Instead of analyzing features one by one, I look at how lower- and higher-salience features pattern together to understand when speakers converge toward another language, maintain their own patterns, or repurpose features as identity markers. This approach bridges morphosyntax, phonetics, and discourse, integrating reproducible annotation and quantitative modeling across multiple feature types.

Future work expands this framework by adding more higher-salience variables (e.g., intervocalic /d/ lenition, vowel lowering, prosodic “cantadito”) and by including new communities beyond Boston as well as additional national identity groups. By comparing both speech production and listener perception, I explore when bilingual speakers adapt for efficiency and when they preserve differences as a form of cultural expression.

Dissertation Research

Conducted as part of the Spanish in Boston Project (NSF BCS-1423840), my dissertation examines how lower- and higher-salience features co-vary in the speech of Puerto Rican and Dominican Bostonians. Using mixed-methods analysis, I show that lower-salience features—like subject-pronoun expression, subject placement, and filled pauses—tend to align with English-compatible patterns, while higher-salience features—such as coda /s/ and liquids (/r, ɾ, l/)—remain socially meaningful resources. The results demonstrate that salience—social as much as perceptual—explains which variables shift together and which persist as identity markers.

At a glance


Scope	Covariation across morphosyntax (pronouns, word order), discourse (filled pauses), and phonology (coda /s/, liquids)
Design	Sociolinguistic interviews from the Spanish in Boston Corpus plus metalinguistic commentary
Speakers	Puerto Rican and Dominican speakers (N=22, ≈24,197 tokens)
Methods	Acoustic analysis in Praat Mixed-effects modeling in R	Praat RStudio

Key findings

Lower-salience features show modest, systematic convergence with English norms (bilingual optimization).
Higher-salience features remain stable & socially indexed, marking national identity distinctions.
Salience is continuous and ideologically mediated, structuring which variables shift together and which become identity markers.

Filled Pauses & Language Contact in Boston Spanish

Mon, 01 Jan 0001 00:00:00 +0000

When people speak, they often pause with short sounds like uh, ah, or eh while planning what to say next. These filled pauses keep the conversational floor and help listeners follow along—yet they vary across languages. English speakers tend to say uh/um, while Spanish speakers more often use eh/em. This project investigates how bilingual Spanish speakers in Boston use these tiny hesitation sounds, revealing how everyday speech encodes patterns of language contact and sound change.

At a glance


Design	Sociolinguistic interviews from the Spanish in Boston Corpus
Speakers	Spanish-speaking Bostonians (N=80, ≈6,364 tokens, 13 national origins)
Methods	Acoustic analysis in Praat Mixed-effects modeling in R	Praat RStudio
Predictors	Age of arrival, language use with interlocutors, and phonetic context
Variable	Vowel quality in filled pauses — fronted [e] (eh) vs. centralized [a]/[ə] (ah, uh)

Key findings

Speakers with greater English exposure (earlier arrival, more bilingual networks) favor centralized pauses ([a], [ə]), while Spanish-dominant speakers prefer [e].
Phonetic context plays little role — the pattern is socially, not structurally, driven.
The shift is conservative: bilinguals reorder pause-filling norms rather than adopting English-like forms wholesale.

Spanish in Boston: Sociolinguistic Dataset & Analysis

Mon, 01 Jan 0001 00:00:00 +0000

Overview

This project involved designing and analyzing sociolinguistic datasets to investigate variation in Spanish speech. It demonstrates end-to-end experience in data creation, annotation design, quality assurance, and statistical analysis.

My Role

Designed for novel linguistic variables
Managed dataset collection, curation, and QA workflows
Supervised and trained student annotators
Led full research lifecycle from data design to statistical modeling

Data & Methods

Built and analyzed datasets of 70k+ tokens
Conducted coding, , and in R
Applied to investigate linguistic variation
Developed workflows for annotation consistency and data quality

Outcome

Produced structured datasets for analyzing Spanish variation
Generated findings contributing to
Demonstrated scalable approaches to linguistic data annotation and QA

Subject Pronoun Expression in Puerto Rican Spanish

Mon, 01 Jan 0001 00:00:00 +0000

In Spanish, finite verbs can occur with or without an expressed subject pronoun — as in yo hablo ‘I speak’ vs. hablo ‘(I) speak’ — without altering the meaning. Drawing on sociolinguistic interviews from Puerto Rican speakers in Louisiana and Puerto Rico, I examine how grammatical and discourse constraints pattern this variation. Results reveal Caribbean-typical pronoun rates (~37%) and show that internal morphosyntactic factors, not English contact, drive pronoun use in this sample.

At a glance


Design	Sociolinguistic interviews with 20 speakers (≈2,300 tokens); analyzed with Goldvarb
Predictors	Person/number, switch-reference, tense–mood–aspect, prior-subject realization, clause type, and verb semantics
Timeline	Conducted at Louisiana State University; later informed dissertation research expanding the same variable to new samples and contexts

Key findings

Pronoun rates ≈ PR 37.8% vs. LA 37.0%.
1SG/3SG and switch-reference favor pronoun usage; plurals disfavor it.
Only location shows a clear social effect — routine English contact did not increase pronoun usage.

Word Order & Information Structure in Latin American Spanish

Mon, 01 Jan 0001 00:00:00 +0000

How do Spanish speakers judge Subject-Verb (SV) vs. Verb-Subject (VS) word orders—and how do verb type (unergative vs. unaccusative) and pragmatic condition (wide focus, subject-narrow focus, subject-given) shape those judgments? Using a controlled, online experiment, I test links between argument structure and information structure in four Latin American dialects.

At a glance


Design	Online acceptability-judgment task (Likert 1–5), 2×3 (Verb Type × Pragmatic Condition)
Survey Hosting Platform	Qualtrics	Qualtrics
Recruitment	Prolific	Prolific
Participants	N = 69 (Chile, Cuba, Mexico, Puerto Rico) → 1,656 ratings (SV/VS)
Methods	Mixed-effects modeling in R	RStudio

Key findings

SV preferred overall; VS favored with unaccusatives.
Verb type and pragmatic condition jointly predict acceptability.
Region not predictive in these data.
Supports a link between argument structure & information structure (Unaccusative Hypothesis).