About

Sarah Murray is an Associate Professor in the Department of Linguistics at Cornell University. Her primary research focuses on the semantics and pragmatics of natural language, specifically analyzing the formal representations needed to understand a variety of linguistic structures across grammatically diverse languages. Her work encompasses topics such as evidentiality, modality, plurality, connectives, sentential mood, and speech acts. Murray combines formal theories of meaning and discourse with documentation and analysis of understudied languages, demonstrating a strong interest in language documentation, methodology for elicitation, community-based language work, and collaboration for language revitalization and reclamation. She has worked with the Cheyenne community in Southeastern Montana since 2006 on various language projects and is the director of the Cornell Language Documentation Lab. Her research interests include crosslinguistic semantics and pragmatics, dynamic semantics, philosophy of language, cognitive science, and language documentation. Murray is affiliated with the American Indian and Indigenous Studies Program and is actively involved in projects related to Algonquian languages, including Cheyenne. Her work emphasizes the importance of language documentation and revitalization, contributing to the understanding of how meaning is coded in language and supporting indigenous language preservation efforts.

Research topics

• Artificial Intelligence
• Sociology
• Computer Science
• Natural Language Processing
• Social Science
• Political Science
• Psychology
• Linguistics
• Philosophy
• Gender studies
• Epistemology
• Criminology
• Social psychology
• Law
• Medicine

Selected publications

• Variation in Prosodic Structure across Algonquian

  Michigan State University Press eBooks · 2026-02-01

  book-chapter
  PublisherDOI

• Effective Early Mother Tongue Literacy Instruction in African Languages

  2024-11-29

  book-chapter

  The work of Funda Wande, an NGO located in Cape Town, South Africa, focuses on prioritizing, thinking through, and testing interventions that will lead to all children learning to read for meaning and calculate with confidence by the age of 10 by 2030. The Funda Wande instructional design and assessment teams have developed research-based early literacy curriculum (Foundation Phase through to Grade 3) in the mother tongues of isiXhosa, Sepedi, and Afrikaans. This curriculum has been implemented in schools across the Eastern Cape, Western Cape, and Limpopo and has been evaluated using RCT design. This chapter addresses the challenges faced in designing this early literacy curriculum, which honor the mother tongues and cultures of South African learners and promote antiracist pedagogy across ethnic groups. In addition, implementing “best practice” from the USA context required thoughtful examination for its cultural relevance in the South African context.

  PublisherDOI

• Cheyenne Demonstratives:

  Michigan State University Press eBooks · 2023-03-01

  book-chapter1st authorCorresponding
  PublisherDOI

• Prosodically Conditioned Phonology in Cheyenne

  Michigan State University Press eBooks · 2022-04-01

  book-chapterSenior author
  PublisherDOI

• Chapter 5: Sensing and Perceiving

  2021-01-01

  article1st authorCorresponding
  Publisher

• Chapter 10: Intelligence and Language

  2021-01-01

  article1st authorCorresponding
  Publisher

• Evidentials

  2020-11-04 · 2 citations

  other1st authorCorresponding

  Evidentials indicate source of information. This chapter reviews the key concepts and distinctions from the semantic literature on evidentials, including essential data and theoretical approaches. Several semantic diagnostics establish various empirical properties of evidential systems, including what propositions are conveyed by sentences with evidentials, what status they have, how evidentials embed, and how they behave in questions. While there are important empirical differences across languages, there are also many similarities. This chapter also reviews three types of theoretical approaches to evidentials – as modals, as speech act operators, and as not‐at‐issue restrictions – and how these approaches apply crosslinguistically.

  PublisherDOI

• Evidentiality, Modality, and Speech Acts

  Annual Review of Linguistics · 2020 · 18 citations

  1st authorCorresponding
  • Computer Science
  • Linguistics
  • Natural Language Processing

  Evidential constructions have two main semantic effects: They contribute information about an individual's source of evidence, and they potentially modify the force of a sentence. In this article, I review the at-issue status of the evidential information, the indexical and anaphoric properties of evidentials, their force-modifying effect, and the connection throughout to epistemic modality. In some languages, evidentials occur as part of the grammatical morphology, but evidential information can be expressed through a variety of constructions across languages. As such, the study of evidentiality highlights the important role of cross-linguistic semantics and the collaboration between language typology and linguistic semantics.

  PublisherOA PDFDOI

• The structure of communicative acts

  Linguistics and Philosophy · 2020 · 76 citations

  1st authorCorresponding
  • Computer Science
  • Computer Science
  • Linguistics
  PublisherDOI

• A Hamblin Semantics for Evidentials and Evidential Questions

  2020-09-30

  book-chapter1st authorCorresponding

  This paper proposes a non-modal analysis of evidentials that can account for their interpretation in declaratives and interrogatives as well as for cross-linguistic variation in evidential behavior. On the proposed view, evidentials contribute not-at-issue, truth conditional content which is new, not presupposed, and they are not analyzed via a separate level of illocutionary meaning (cf., Faller 2002) or as a modal (cf., Izvorski 1997, McCready & Ogata 2007, Matthewson et al. 2007). Unlike other analyses, this proposal can account for the fact that when a sentence with an evidential is rejected, only the main content is rejected—the evidential content is not.

  PublisherDOI

Frequent coauthors

• Miloje Despić

  Cornell University

  3 shared

• William B. Starr

  Cornell University

  2 shared

• Michael David Hamilton

  2 shared

• Jessica Rett

  University of California, Los Angeles

  2 shared

• Rachel I. Vogel

  University of Minnesota

  1 shared

• Henk Zeevat

  University of Amsterdam

  1 shared

• Marianne Huijsmans

  University of British Columbia

  1 shared

• Mehrnoosh Sadrzadeh

  1 shared

Awards & honors

• Klarman Fellow