Broadly, I’m interested in syntax, language acquisition, learnability, and semantics.
More specifically, I’m interested in studying grammatical competence (i.e., what it is that we know when we know a language). However, we only have access to this knowledge via its use in individual humans. One major strand of my research involves careful investigation into how we can parcel out performance factors to ensure we’re identifying aspects of grammatical competence. I do this type of psycholinguistic research both with children and with adults.
The other major strand of my research involves investigations into learnability. Grammatical theories and their concomitant representations are often evaluated on the basis of data fit. I’m interested in furthermore using learnability considerations to help adjudicate between theories and representations.
Below you’ll find a selected list of specific research projects. Things that are not listed here but that are listed on my CV are available upon request.
Non-actional passives can be comprehended by 4-year olds
- Collaborator: Jeff Lidz
- Keywords: acquisition syntax pragmatics
- Output: This research is currently in poster format and is available here.
- Overview: English-acquiring children have difficulty comprehending passives of non-actional verbs, such as “forget”, until 5 or later (e.g., Maratsos et al. 1985). Many have hypothesized that English-acquiring children lack grammatical knowledge of the passive (e.g., Borer & Wexler 1987, 1992; Fox & Grodzinsky 1998; Wexler 2004; Snyder & Hyams 2015). We report experimental evidence that 4-year olds can comprehend passives of non-actional verbs, if their use is licensed by the context. The “by”-phrase of an English passive normally carries narrow focus; it is most naturally used when what is at issue is the content of the “by”-phrase. We conducted a TVJT with stories where there were always three potential forgetters. 4-year olds performed significantly above chance on the passive trials. 4-year-old children do have grammatical knowledge of passives; their failure to comprehend them in previous studies is due to discourse-inappropriate use.
Modeling the learning of the Person Case Constraint
- Collaborator: Naomi H. Feldman
- Keywords: morphosyntax learnability features representations PCC
- Output: This project has resulted in 1 publication.
- Overview: We demonstrate the usefulness of using learnability considerations to evaluate different theories and their concommitant feature representations, above and beyond parsimony considerations, in the domain of the Person Case Constraint (PCC). We find that a theory of the PCC that relies on atomic person feature representations may systematically learn an unattested variant of the PCC when learning from data drawn from the grammar of an attested variant of the PCC. On the other hand, a theory of the PCC that relies on person feature representations consisting of bundles of binary-valued Author and Participant features quickly converges on the target grammar and is not led astray.
Inhibition and children’s productions of medial wh-phrases
- Collaborators: Elaine Grolla and Jeff Lidz
- Keywords: acquisition syntax wh movement
- Output: This project is currently being written up. The latest version of this research is available as a talk.
- Overview: We report evidence that children’s non-adult-like productions of questions with medial wh-phrases, such as Who do you think who was chasing the boys, are the result of immature inhibition control, not the result of having temporarily settled on a non-target grammar (like some have argued).
There is a simplicity bias when generalising from ambiguous data
- Collaborator: Karthik Durvasula
- Keywords: phonotactics learnability artificial language learning
Output: This project has resulted in 1 publication.
Durvasula, Karthik & Adam Liter. 2020. There is a simplicity bias when generalising from ambiguous data. Phonology 37(2). 177–213.
- Overview: How exactly do learners generalize in the face of ambiguous data? While there has been a substantial amount of research studying the biases that learners employ, there has been very little work on what sorts of biases are employed in the face of data that is ambiguous between phonological generalizations with different degrees of complexity. We show, via three artificial language learning experiments, that, at least for phonotactic sequence patterns, learners are able to keep track of multiple generalizations related to the same segmental co-occurrences; however, the generalizations they learn are only the simplest ones consistent with the data.
Reevaluating the Maratsos Effect
- Collaborators: Tess Huelskamp, Susima Weerakoon, and Alan Munn
- Keywords: acquisition syntax
- Output: This project is currently being written up. Older versions are available upon request.
- Overview: Maratsos et al. (1985) found that children were later in acquiring the passives of “nonactional” verbs than passives of “actional” verbs; this has since been robustly replicated in the literature and has come to be known as the Maratsos Effect. Looking back at the literature, we question whether the “nonactional” class of verbs was ever a linguistically homogenous class. Based on verbal diagnostics for agentivity and eventivity, the “nonactional” class of verbs that have been tested can be shown to be quite heterogenous, consisting of both nonagentive eventive verbs and nonagentive noneventive (i.e., stative) verbs. We report two suggesting that the Maratsos Effect depends on agentivity, not eventivity.
Obligatory number and plural morphology
- Collaborators: Chris Heffner, Tess Huelskamp, and Cristina Schmitt
- Keywords: semantics number language universals artificial language learning
Output: This project has resulted in 2 publications.
Liter, Adam, Tess Huelskamp, Christopher C. Heffner & Cristina Schmitt. 2018. Grammaticalized number, implicated presuppositions, and the plural. Glossa 3(1): Acquisition of quantification. Magda Oiry (ed.)
Liter, Adam, Christopher C. Heffner & Cristina Schmitt. 2017. The interpretation of plural morphology and (non-)obligatory number marking: An argument from artificial language learning. Language Learning and Development 13(4). 451–480.
- doi pdf
- Overview: Languages like English allow inclusive interpretations (‘one or more’) of plural morphology in downward entailing contexts. For example, there are not books on the table is false even if there is one book on the table. Other languages, such as Korean, require exclusive interpretations (‘more than one’) even in downward entailing contexts. We investigate whether these differeing interpretations are mere historical accidents or actually follow from properties of the grammar. With data from two artificial langueage learning experiments, we provide suggestive evidence that non-obligatory number marking necessarily precludes inclsuive interpretations of the plural morpheme.