Zara Harmon

Zara Harmon

Post-doctoral Researcher

Max Planck Institute for Psycholinguistics

Research

Accessibility-driven Language Production

Published in Cognitive Psychology, 2017

Our work on accessibility (Harmon & Kapatsinski, 2017) and my dissertation work (Harmon, 2019), demonstrates that high accessibility of a form results in the extension of that form to novel related contexts, an extension that is deemed less appropriate by the speakers when the production pressures are eliminated.

Highlights
  - Semantic change results from extension of frequent forms to new uses.
  - Extension of frequent forms to new uses occurs via accessibility in production.
  - In contrast frequent forms are restricted to their original meanings in comprehension.
  - Entrenchment in comprehension can curb (over-)extension in production.

Harmon, Z., & Kapatsinski, V. (2017). Putting old tools to novel uses: The role of form accessibility in semantic extension. Cognitive Psychology, 98, 22–44. [Paper]

Learning Mechanisms in Cue Weighting

Published in Cognition, 2019

Feedback has been shown to be effective in shifting attention across perceptual cues to a phonological contrast in speech perception. However, the learning mechanisms behind this process remain obscure. In this work, we compare the predictions of supervised error-driven learning (Rescorla & Wagner, 1972) and reinforcement learning (Sutton & Barto, 1998) about the effect of feedback on cue reweighting using computational simulations.

Highlights
  - Experimental data show that when given explicit error feedback, learners downweight a primary cue to a phonetic contrast, but only when they can use another cue to reduce prediction error.
  - Simulations of experimental data implicate reinforcement learning as the mechanism behind the effect of feedback on cue weighting in speech perception.

Harmon, Z., Idemaru, K., & Kapatsinski, V. (2019). Learning mechanisms in cue reweighting. Cognition, 189, 76–88. [Paper]

Repetition and Retrieval in Langauge Production

Published in Psychological Review, 2021

Repetitions are ubiquitous in all domains that involve producing an action sequence, and occur when the upcoming plan is not activated enough to be executed. Repetitions are helpful in this situation because the repeated action sequence activates the likely continuation. In this work, we propose a theory of repetition and retreival in the form of three hypotheses: The faciliation hypothesis, the reactivation hypothesis, and the initiation hypothesis.

Highlights
  - The Facilitation Hypothesis maintains that repetitions facilitate accessing the upcoming item.
  - The Reactivation Hypothesis claims that previously produced words must be reactivated to be re-produced, and that this reactivation process uses the words that follow as cues.
  - The Initiation Hypothesis claims that, after an interruption, speech is restarted from words that have occurred relatively unexpectedly in the speaker’s prior experi- ence. The Initiation Hypothesis attributes this effect to Cue Competition between preceding context and top-down cues: initiation from words that tend to occur in predictive preceding-word contexts is relatively unlikely because such words have a weaker association with top-down cues.

Harmon, Z., & Kapatsinski, V. (2021). A theory of repetition and retrieval in language production. Psychological Review, 128(6), 1112–1144. [Paper]

The Competition–Compensation Account of Developmental Language Disorder

Published in CSS proceedings, 2021

Children with developmental language disorder (DLD) regularly use the base form of verbs (e.g.,dance) instead of inflected forms (e.g., danced). Our work suggests that inconsistent use of inflectional morphemes by children with DLD could stem from a combination of processing difficulties that affect novelty detection combined with a resulting learned bias to rely on unanalyzed chunks.

Highlights
  - The competition–compensation account proposes that processing difficulties of children with DLD disproportionally affect processing novel inflected verbs in their input.
  - For children with DLD, the inflected form faces stronger competition from alternatives because these children infer a lower probability for the inflection in unfamiliar and novel contexts.
  - Competition is resolved through a compensatory behavior which involves producing a more accessible alternative with high phonological and semantic overlap with the inflected form: in English, the bare form.

Harmon, Z., Barak, L., Shafto, P., Edwards, J., & Feldman, N. (2021). Making Heads or Tails of it: A Competition–Compensation Account of Morphological Deficits in Language Impairment. Proceedings of the Annual Conference of the Cognitive Science Society, 43, 1872–1878. [Paper]