About

Kathryn Davidson is a Professor of Linguistics at Harvard University whose research explores the human capacity to understand an infinite number of novel sentences through the study of semantics, pragmatics, and language development. Her work investigates how contextual information is incorporated into meaning and how humans learn to interpret language. She addresses fundamental questions about the types of symbolic and other representations that capture meaning in human language, the impact of language modality—including speech, sign, gesture, and writing—on meaning creation and sharing, and how meaning in language differs from meaning outside of language. Additionally, she explores the use of new tools for collecting and analyzing linguistic data to develop mathematically precise models of meaning, as well as the role of language input in language development, such as the benefits of sign language exposure for deaf children's overall language abilities. Davidson directs the Meaning & Modality Laboratory, where traditional linguistic theory is combined with psycholinguistic experimental methods to gather and analyze behavioral data from a variety of spoken and signed languages. Her lab also collaborates with computational linguists to compare human linguistic behavior with that of advanced artificial models, including Large Language Models, across similar tasks. Among her recent scholarly outputs are an Element on Semantics and Depiction and a book on Formal Semantics and Pragmatics in Sign Languages, both in production with Cambridge University Press. She is based at Harvard's Department of Linguistics in Boylston Hall and the Meaning & Modality Lab at 1100 Mass Ave.

Research topics

• Immunology
• Biochemistry
• Internal medicine
• Algorithm
• Cell biology
• Medicine
• Surgery
• Gastroenterology
• Biology

Selected publications

• Tumour specific HORMAD1 expression perturbs mitotic arrest and drives sensitivity to mitotic kinase inhibitors

  Nature Communications · 2026-03-11

  articleOpen access

  HORMAD1 expression is usually restricted to germ-line cells but is also aberrantly expressed in 60% of triple-negative breast cancers (TNBCs), where it is bi-modally expressed and associated with genomic instability. Here, we show that out-of-context HORMAD1 expression in mitotic cells perturbs mitotic arrest and generates aneuploidy. These phenotypes are caused by out-of-context HORMAD1 expression driving a weakening of the spindle assembly checkpoint (SAC) and/or in kinetochore-microtubule error correction. These mitotic effects of HORMAD1 are MAD2L1-independent, but instead caused by a HORMAD1/Aurora B interaction and defective Aurora B/INCENP signalling. Consistent with this mechanism, aberrant HORMAD1 expression causes sensitivity to MPS1, Aurora B or BUB1 inhibitors currently being investigated as cancer treatments. Our data suggests how out-of-context expression of a meiotic gene imparts genomic instability upon tumour cells and also identifies several associated dependencies as mechanism-based therapeutic targets for a large, biomarker-defined, subset of cancers. HORMAD1 expression is typically restricted to germline cells where it has an important role in meiotic recombination but has been shown to be upregulated in triple negative breast cancer (TNBC). Here, the authors report that aberrant HORMAD1 expression weakens the spindle assembly checkpoint, driving sensitivity to AURORA kinase inhibition.

  PublisherOA PDFDOI

• Anaphoric demonstratives in Mandarin

  Proceedings from Semantics and Linguistic Theory · 2025-02-12

  articleOpen accessSenior author

  The goal of this study is to experimentally evaluate contrasting claims in the theoretical literature on the acceptability of Mandarin demonstratives and definite bare nouns in anaphoric contexts. Jenks (2018) argues that Mandarin differentiates between uniqueness-based (weak) and anaphoric (strong) definites through bare nouns and demonstratives, respectively. In contrast, Dayal & Jiang (2022), Bremmers, Liu, van der Klis & Le Bruyn (2022), and Simpson & Wu (2022) claim that both bare nouns and demonstratives can be used in anaphoric contexts in Mandarin, proposing slightly differing explanations with regards to their felicity, tied to factors such as discourse coherence between context and follow-up sentences. Our findings illustrate that Mandarin demonstratives are strongly preferred across the board in anaphoric contexts, patterning with anaphoric definites (rather than demonstratives) in languages such as English, Turkish (Saha, Sa˘g & Davidson 2023), and Bangla (Saha 2023). Additionally, we observe that definite bare nouns are also felicitous in anaphoric contexts, albeit as a less preferred option. We argue that this preference for demonstratives arises because Mandarin bare nouns can have (i) generic interpretations due to the absence of tense and aspectual marking, and (ii) indefinite interpretations in post-verbal positions (Cheng & Sybesma 1999; Simpson & Wu 2022). Demonstratives, by contrast, are unambiguously anaphoric, driving their overall preference.

  PublisherOA PDFDOI

• Type 1 laryngeal clefts – Which patients can be managed medically? A retrospective cohort study

  International Journal of Pediatric Otorhinolaryngology · 2025-06-20

  article
  PublisherDOI

• Investigating the interaction of linguistic and mathematical reasoning in language models using multilingual number puzzles

  2025-01-01

  articleOpen access

  Across languages, numeral systems vary widely in how they construct and combine numbers.While humans consistently learn to navigate this diversity, large language models (LLMs) struggle with linguistic-mathematical puzzles involving cross-linguistic numeral systems, which humans can learn to solve successfully.We investigate why this task is difficult for LLMs through a series of experiments that untangle the linguistic and mathematical aspects of numbers in language.Our experiments establish that models cannot consistently solve such problems unless the mathematical operations in the problems are explicitly marked using known symbols (+, , etc, as in "twenty + three").In further ablation studies, we probe how individual parameters of numeral construction and combination affect performance.While humans use their linguistic understanding of numbers to make inferences about the implicit compositional structure of numerals, LLMs seem to lack this notion of implicit numeral structure.We conclude that the ability to flexibly infer compositional rules from implicit patterns in human-scale data remains an open challenge for current reasoning models.

  PublisherOA PDFDOI

• Crisis as catalyst? Exploring cities’ climate emergency declarations for transformative urban governance

  Urban Transformations · 2025-08-22 · 1 citations

  articleOpen access

  Deep into the ‘Climate Decade’, radical and swift action to avert and prepare for climate disaster remains absent in cities, hindered by pervasive institutional barriers. In this perspective, we propose capacities for transformative urban governance as a lens to study the diffuse, institutional impacts of local governments’ declarations of ‘climate emergency’. We aim to illustrate an alternative approach to evaluating trans-municipal policy phenomena such as Climate Emergency Declarations – one that moves beyond linear assessments of policy progress and instead focuses on changes in urban governance arrangements. Drawing on existing scholarly reviews, we explore whether and how Climate Emergency Declarations reshape underlying governance conditions to support the pursuit of transformative change. In doing so, we foster a dialogue between reviews of Climate Emergency Declarations and Urban Transitions and Transformations research. This allows us to derive strategic directions for advancing transformative urban governance through Climate Emergency Declarations. Furthermore, Climate Emergency Declarations open new research avenues within Urban Transitions and Transformations scholarship to engage with the fear, grief, and conflicts arising from the urgency and threats associated with the climate crisis. Transformative governance capacities to assess cities’ Climate Emergency Declarations (CEDs). CEDs offer opportunities for institutional learning beyond linear policy progress. Transformative urban governance capacities support critical reflection on changes in urban governance. CEDs open research directions on fear, grief, and conflict in transformative urban governance.

  PublisherOA PDFDOI

• Investigating the interaction of linguistic and mathematical reasoning in language models using multilingual number puzzles

  ArXiv.org · 2025-06-16

  preprintOpen access

  Across languages, numeral systems vary widely in how they construct and combine numbers. While humans consistently learn to navigate this diversity, large language models (LLMs) struggle with linguistic-mathematical puzzles involving cross-linguistic numeral systems, which humans can learn to solve successfully. We investigate why this task is difficult for LLMs through a series of experiments that untangle the linguistic and mathematical aspects of numbers in language. Our experiments establish that models cannot consistently solve such problems unless the mathematical operations in the problems are explicitly marked using known symbols ($+$, $\times$, etc., as in "twenty + three"). In further ablation studies, we probe how individual parameters of numeral construction and combination affect performance. While humans use their linguistic understanding of numbers to make inferences about the implicit compositional structure of numerals, LLMs seem to lack this notion of implicit numeral structure. We conclude that the ability to flexibly infer compositional rules from implicit patterns in human-scale data remains an open challenge for current reasoning models.

  PublisherOA PDFDOI

• Title Pending

  Glossa a journal of general linguistics · 2025-01-01

  articleOpen access

  This is an accepted article with a DOI pre-assigned that is not yet published.Children can use distributional information about where words occur to figure out their meanings. But what happens when two very different words not only have most of their distribution in common, but also compose to form indistinguishable sentential meanings in those common cases? As a negative polarity item (NPI), any is selectively licensed by certain linguistic environments, the most common of which is negation. This is the context in which children hear any in around 80% of their input. However, under negation, the meaning of any looks just like a negative quantifier in concord with the higher negation (a negative concord item; NCI). While studies of children’s production indicate that they hardly ever produce any without a licensing negation, suggesting competence with its distribution, we hypothesize that some children may have misanalysed its meaning. To investigate what children think any means, we tested 106 monolingual English-speaking children between 2 and 6 years of age, as well as 20 adults, in two picture-choice comprehension tasks. These tasks assessed their interpretation of any without a preceding negation, both in a licensed (free choice) and an unlicensed context. While most children interpreted any the same way adults did, we also found a group of children who systematically responded to any as if it meant no, consistent with a negative concord (mis)analysis. In addition to illustrating how much children rely on distributional information to learn such abstract words, this finding bears on several debates. It raises the question of whether it is possible to represent the licensing conditions of NPIs prior to knowing their meanings. And it suggests that children may be biased to assume that their language uses negative concord constructions even when it does not.

  PublisherDOI

• Fake reefs are sometimes reefs and sometimes not, but are always compositional

  Experiments in Linguistic Meaning · 2025-01-24

  articleOpen accessSenior author

  The semantics of adjective modification often begins with set intersection,such that [[yellow flower]] = [[yellow]] ∩ [[flower]]. Thus a yellow flower is a flower. Such an account, however, runs into problems for adjectives like fake or counterfeit, which display a privative inference: a fake gun is not a gun and a counterfeit dollar is not a dollar. Moreover, recent work shows privativity cannot easily be encoded as a property of specific adjectives like counterfeit, since e.g. counterfeit watch robustly licenses the subsective inference of being a watch (Martin 2022). We gather judgments on nearly 800 adjective-noun bigrams (of which 180 are novel, i.e. zero corpus frequency), andshow that privativity depends on the adjective, noun and context, and can be manipulated for the very same adjective-noun bigram by presenting it in different contexts. This poses a challenge for theories which fix privativity as a property of the adjective and always use the same method of composition (Partee 2010, del Pinal 2015). Moreover, we find no difference in participant behavior between novel adjective-noun bigrams and high frequency ones, suggesting that the process is nonetheless compositional and not the result of convention or memorized idiosyncrasy. Our results support compositional accounts like Martin (2022) (which modifies del Pinal 2015) and Guerrini (2024), which treat privativity as context-dependent.

  PublisherOA PDFDOI

• Non-apoptotic caspase-8 is critical for orchestrating exaggerated inflammation during severe SARS-CoV-2 infection

  Nature Communications · 2025-11-13 · 3 citations

  articleOpen access

  Inflammation and excess cytokine release are hallmarks of severe COVID-19. While programmed cell death is known to drive inflammation, its role in SARS-CoV-2 pathogenesis remains unclear. Using gene-targeted murine COVID-19 models, we here find that caspase-8 is critical for cytokine release and inflammation. Loss of caspase-8 reduces disease severity and viral load in mice, and this occurs independently of its apoptotic function. Instead, reduction in SARS-CoV-2 pathology is linked to decreased IL-1β levels and inflammation. Loss of pyroptosis and necroptosis mediators in gene-targeted animals provides no additional benefits in mitigating disease outcomes beyond that conferred by loss of caspase-8. Spatial transcriptomic and proteomic analyses of caspase-8-deficient mice confirm that improved outcomes are due to reduced pro-inflammatory responses, rather than changes in cell death signalling. Elevated expression of caspase-8 and cFLIP in infected lungs, alongside caspase-8-mediated cleavage of N4BP1, a suppressor of NF-kB signalling, indicates a role of this signalling axis in pathological inflammation. Collectively, these findings highlight non-apoptotic functions of caspase-8 as a driver of severe COVID-19 through modulation of inflammation, not through the induction of apoptosis.

  PublisherOA PDFDOI

• Preclinical evaluation of a cross-protective ß-SARS-CoV-2 virus-like particle vaccine adjuvanted with MF59

  Research Square · 2024-08-22 · 2 citations

  preprintOpen access
  PublisherOA PDFDOI

Frequent coauthors

• Ivano Caponigro

  University of California, San Diego

  69 shared

• Carlo Geraci

  École Normale Supérieure - PSL

  69 shared

• Onno Crasborn

  Radboud University Nijmegen

  65 shared

• Labex Efl

  Radboud University Nijmegen

  64 shared

• Caterina Donati

  Laboratoire de Linguistique Formelle

  64 shared

• Carlo Cecchetto

  64 shared

• Valentina Aristodemo

  University of Trento

  64 shared

• David Sugden

  King's College London

  26 shared

Labs

• Reading GroupsPI

Education

• Ph.D., Linguistics

  Harvard University

  2005

• B.A., Linguistics

  University of California, Berkeley

  1999