About

Robert Henderson is a Professor and the Director of Graduate Studies in the Department of Linguistics at the University of Arizona. His research focuses on the semantics and pragmatics of human languages, involving complex representations and operations over those representations. His goal is to design logics that define these representations and operations, closely mimicking what is observed in natural languages, thereby modeling how people interpret the meanings of arbitrary expressions in the languages they speak. He emphasizes the importance of empirical evidence from a wide variety of languages to understand how meaning in human languages works in general. To this end, he has conducted primary fieldwork on several Mayan languages spoken in Guatemala, working to describe, document, and analyze these languages both in their own right and to gain insights into the general workings of human language.

Research topics

• Computer Science
• Artificial Intelligence
• Linguistics
• Sociology
• Philosophy
• Psychology
• Political Science
• Natural Language Processing
• History
• Epistemology
• Human–computer interaction
• Social psychology

Selected publications

• Acute postoperative complications after scleral fixation of Akreos AO60 intraocular lens

  Journal of Surgical Case Reports · 2026-02-20

  articleOpen accessSenior author

  Abstract We report three cases of postoperative exogenous endophthalmitis and toxic anterior segment syndrome (TASS) following vitrectomy for dislocated intraocular or crystalline lenses and scleral fixation of an Akreos intraocular lens (IOL) with Gore-Tex sutures. To our knowledge, postoperative exogenous endophthalmitis and TASS have not been previously described with this technique. A 67-, 56-, and 36-year-old underwent pars plana vitrectomy, lens removal or vitreolensectomy, and Akreos IOL scleral fixation. Three days later, cases 1 and 3 required vitreous biopsy and intravitreal Vancomycin/Amikacin for culture-positive endophthalmitis (Staphylococcus epidermidis and Streptococcus pneumoniae). Case 2 underwent removal of a dislocated fluocinolone implant, biopsy, and intravitreal antibiotics; negative cultures supported a diagnosis of TASS. Scleral-fixated IOLs carry risks of endophthalmitis and TASS. Large corneal wounds increase risk; meticulous wound construction is essential. Akreos IOLs do not prevent anterior migration of steroid implants. Immediate management is crucial.

  PublisherDOI

• Virtual vitreoretinal clinics: a service delivery pathway of the future

  International Journal of Retina and Vitreous · 2025-06-01 · 1 citations

  articleOpen access

  BACKGROUND: Vitreo-macular interface (VMI) disorders, including epiretinal membrane (ERM) diagnosed on optical coherence tomography (OCT), form a significant proportion of elective referrals to vitreoretinal (VR) surgeons. An in-person visit to a clinician involves travelling, waiting, investigations then an interaction with the surgeon, which entails many inefficiencies in a large institution. We report the pilot studies of a VR virtual service where these patients can be more efficiently reviewed, investigated, listed for surgery or discharged. METHODS: This was a prospective observational study comparing the outcomes of a virtual assessment to standard face-to-face clinics. All patients included were referred from optometry practices for ERM diagnosed on macula OCT. A first pilot study comprised 79 patients, who attended a diagnostics centre staffed with ophthalmic-trained technicians. A short history, visual acuity and ocular pressures were recorded. Widefield colour photographs and macular OCT images were acquired. Cases were asynchronously reviewed by trained ophthalmologists and senior nurses within the week, and following a telephone consultation with the patient, a virtual management plan was documented. All patients attended 1 week later for a face-to-face appointment, following which, virtual and face-to-face management plans were compared. A second pilot comprised 65 patients, through the same pathway, to examine consistency. A post-hoc analysis was carried out to identify the cohort of patients who would be suitable for a virtual management decision without a telephone consultation. RESULTS: ERMs comprised 35% of overall elective referrals in this study. In Pilot 1, 42% were virtually assessed for discharge, with high concordance with face-to-face outcomes (positive predictive value = 89%). There were 3 cases of missed retinal tears, and 1 OCT misdiagnosis. In the second pilot, 43% were discharged virtually, with higher concordant discharge rates (positive predictive value = 93%). There were no missed peripheral pathology and no misdiagnoses in this pilot. CONCLUSIONS: Our virtual model demonstrates a safe and effective way of managing and discharging patients without a face-to-face clinic. This is especially suitable for low-risk conditions such as ERMs, which comprise a large proportion of referrals.

  PublisherOA PDFDOI

• Oxford Studies in Semantics and Pragmatics

  2024-06-05 · 14 citations

  other1st authorCorresponding
  PublisherDOI

• Identifying dogwhistles

  2024-06-05

  book-chapter1st authorCorresponding

  Abstract The goal of this chapter is to understand what contextual conditions make so-called ‘identifying’ dogwhistles possible, to understand when it is rational for speakers to use identifying dogwhistles, and to understand what kind of language makes for a good 56identifying dogwhistle. To reach these goals we must go beyond standard sociolinguistic theory in terms of how to think about variants, and also beyond the formal model of variation developed in Burnett 2017, 2019. Interactions involving identifying dogwhistles do not just seem more complicated, but prove to require richer formals systems to model them. By building such a model in this chapter, we will begin to see more clearly the shape of the phenomenon. In particular, we extend Burnett 2017, 2019 and show how dogwhistles arise under various parameter settings of the model.

  PublisherDOI

• Dogwhistles and trust

  2024-06-05

  book-chapter1st authorCorresponding

  Abstract Why should listeners ever trust a speaker whose utterances are not truth-tracking? It may seem to be irrational, yet there are many examples, especially in the political sphere. This chapter uses the tools presented in this book so far, augmented with existing work on reliability and trust, to address this puzzle in three parts. First we spell out and contextualize the starting puzzle in a theory of source evaluation based on interactional histories and heuristics for judgments of reliability (McCready 2015). We then build a theory of how ideological considerations are valued alongside truth-conditional content. This is then used together with temperature parameters on truth-conditional content and social meaning to model the degree to which each is valued in particular contexts.

  PublisherDOI

• General preface

  2024-06-05

  other1st authorCorresponding
  PublisherDOI

• A probabilitic pragmatics for dogwhistles

  2024-06-05

  book-chapter1st authorCorresponding

  Abstract While we propose that dogwhistles involve a kind of meaning mostly novel to formal semantics, namely social meaning, in an ideal situation the pragmatic mechanisms that operate over these meanings will be those same mechanisms involved in reasoning about standard truth-conditional content. This chapter lays the foundation for this kind of unified account of the pragmatics of truth-conditional and social meanings. We do so by introducing the Bayesian Rational Speech Act (RSA) framework (Franke and Degen 2016; Franke and Jager 2016; Goodman and Frank 2016, among others). After reviewing the recent pragmatics literature making use of RSA, we then introduce the reader to so-called Social Meaning Games (SMGs), which are a variety of RSA signaling games introduced by Burnett (2017, 2019) to unify variationalist sociolinguistics and modern game-theoretic pragmatic theory. This game-theoretic approach to sociolinguistic interaction will provide firm ground for building an account of dogwhistles.

  PublisherDOI

• The Archetype of the Thug

  Psychological Perspectives · 2024-04-02

  article1st authorCorresponding
  PublisherDOI

• Vigilance and hypervigilance

  2024-06-05

  book-chapter1st authorCorresponding

  Abstract In this chapter we will continue to explore how listeners do/should react to dogwhistles, given that persona construction is interactional. In particular, we return to the core RSA framework developed in Chapter 4 and consider the behavior of L1, the sociolinguistically aware listener. How should a listener optimally reason about a speaker’s persona, assuming a speaker that is attempting to maximize their social utility relative to an audience? This chapter begins to look at this question and identifies two strategies that we see in actual listeners, which we dub vigilance and hypervigilance. The first follows 92from the RSA theory we have developed in previous chapters, and, in fact, we will see that it involves a standard kind of implicature familiar from the RSA literature on implicatures in the truth-conditional domain. The second we relate to the credibility result for cheap talk games of Farrell 1993, where a neologism is credible to the degree that the interests of the sender align with those of the receiver.

  PublisherDOI

• Enriching dogwhistles

  2024-06-05

  book-chapter1st authorCorresponding

  Abstract In this chapter, we turn to another type of dogwhistle, the enriching dogwhistle, where the recognition of the dogwhistle alters the semantic content of the utterance that hosts it. Our aim is to propose an account of dogwhistles of this type. We propose an account on which recognition of speaker persona invites certain kinds of inferences, which result in alterations of the meaning recovered by ‘savvy’ interpreters. Doing so requires some explication of the nature of personas and ideologies, and of the ways in which the recognition of someone’s persona influences our views of the attitudes and beliefs they have. We claim that certain kinds of personas, mainly those 76associated with ideologies and political stances, ‘project’ sets of beliefs and values. Such projections enable certain kinds of invited inferences which, we claim, ground the phenomenon of enriching dogwhistles.

  PublisherDOI

Recent grants

• Collaborative Research: Investigations into tone and stress in a complex prosodic system

  NSF · $167k · 2016–2022

Frequent coauthors

• Ryan Bennett

  26 shared

• Meg Harvey

  University of Arizona

  19 shared

• Elin McCready

  13 shared

• Alan Merry

  Auckland City Hospital

  11 shared

• Craig S. Webster

  University of Auckland

  10 shared

• Chris Frampton

  University of Otago

  10 shared

• Boaz Shulruf

  UNSW Sydney

  9 shared

• Kaylene Henderson

  Auckland UniServices (New Zealand)

  9 shared