Research
My work aims to characterize the nature of the language system in autism. A large percentage—estimated between 17% and 30%—of autistic individuals remain low-verbal or nonverbal (i.e., producing little or no language, regardless of modality) into school age and beyond. I am interested in studying language development, language knowledge, and language processing specifically in low-verbal autistic individuals to discover what factors play a role in the development of linguistic generativity, and to learn what this can tell us about language in the mind and its relationship to other areas of cognition. I see my work as reaching from a cognitive perspective toward the development of better clinical strategies to support the language and communication of autistic individuals.
I use approaches and methods like longitudinal analysis, language sample analysis, eye gaze (preferential looking; Visual World Paradigm), and parent report to study language in autistic and neurotypical children and adults.
Areas of interest:
Abstract Rule Knowledge in Low-Verbal Autism | Scripting | Predictors of Language Outcomes | Verb Development
Abstract Rule Knowledge in Low-Verbal Autism
Some autistic individuals do not demonstrate linguistic generativity in their speech, producing single words and/or scripts (rote, inflexible language repeated over time) but not regularly combining words creatively. Some proposals assume that for these low-verbal individuals, linguistic knowledge consists solely of memorized chunks (lexical/item-specific knowledge), without abstract linguistic rules that guide the generation and interpretation of compositional sequences. However, these proposals are based on production data alone, and rule-based knowledge may be obscured by production-specific limitations in low-verbal individuals.
To come closer to determining where the breakdown occurs in low-verbal autism, I investigate abstract linguistic knowledge in school-age low-verbal autistic individuals via tests of language comprehension. I am currently running two eye gaze studies: the first tests incremental integration of the meanings of words (i.e., integration of the meaning of color adjectives like red with the meaning of nouns like shoe to understand the phrase red shoe), and the second tests knowledge of the abstract link between syntactic position (subject, object) and thematic relation (agent, patient) in subject-verb-object structures.
Recruitment, data collection, coding, and analysis ongoing.
Scripting
Many autistic individuals use scripts, or phrases repeated over time in relatively inflexible ways. For example, someone might say, “Good catch, Max!” whenever anyone catches anything. The nature of these scripts is yet to be determined; are these memorized, unanalyzed chunks akin to lexical items, or are they composed phrases? Gestalt language processing (GLP) proposals claim that autistic individuals who script (and who do not use much language other than scripts) do not have linguistic knowledge of the scripts they use. I test this claim.
In a current project with with 8- to 22-year-old low-verbal autistic individuals who script, participants’ parents and speech-language pathologists (SLPs) report any scripted language they use, and imageable nouns are extracted from their scripts to create individualized eye gaze experiments testing comprehension of these nouns. Success in these experiments would contradict GLP claims that these individuals do not have knowledge of the internal structure of their scripts, while failure has the potential to support GLP claims. Since many SLPs use GLP approaches in therapy, finding or failing to find support for GLP claims has clinical implications.
Recruitment, data collection, coding, and analysis ongoing.
Predictors of Language Outcomes
Given the large percentage of autistic individuals who remain low-verbal or nonverbal, it is critical to determine what factors influence language trajectories and outcomes for autistic children. One line of my research has centered around longitudinal samples, investigating linguistic factors that predict language outcomes or that differentiate those who will progress linguistically from those who will not.
We have found that verb diversity and noun diversity (the number of different verbs/nouns produced) in toddlerhood predicts language outcomes two years later, but only early verb growth, and not noun growth, predicts later language outcomes. Verb production remains important at a later developmental stage; verb diversity (and not noun diversity, mean length of utterance, or proportion of socially motivated utterances) at age five predicts adult language and communication outcomes. This could be related to the role of verbs in the transition from single word production to combinatorial speech; verbs allow a speaker to convey relations among entities in the environment, and they are the only part of speech that is syntactically required to form a sentence.
We have also studied autistic and neurotypical children who begin at the same (low) level of morphosyntactic development to determine factors that differentiate children who progress linguistically from those who do not; these factors include overall vocabulary size, receptive language scores, and response to joint attention bids. Group differences were clearest in non-expressive language measures (receptive language, joint attention); perhaps skills like language comprehension and tendency to join social interactions are more indicative of future linguistic progress than expressive language features in this initially low-verbal subgroup.
LeGrand, K.J., & Naigles, L. (2023). Can expressive language features indicate whether low-verbal children with ASD will progress linguistically? Poster presented at the Society for Research in Child Development (SRCD) Biennial Meeting.
Verb Development
We have found that early verb production predicts language outcomes in autism. Given this, another line of my research has focused on characterizing verb development via analysis of longitudinal language samples from autistic and neurotypical children (ages 20-32 months at study start). In a series of analyses, we have found that:
- Some (around 30%), but not all, autistic children seem to have difficulty with verb production compared to their neurotypical peers, and some of these children showed progress in noun production alongside lack of progress in verb production, suggesting verb-specific difficulty
- Autistic and neurotypical children show the same developmental pattern of result verb (e.g., clean) and manner verb (e.g., wipe) production, supporting the hypothesis that autistic children use similar strategies to their neurotypical peers to learn and produce result and manner meanings
- Neurotypical children produce a higher proportion of internal state verbs than autistic children, which may relate to social differences in autism
- Autistic and neurotypical children rely on high-frequency verbs (e.g., go, do, make) to a similar degree, contrary to the idea that autistic children may over-rely on high-frequency verbs (supposedly easier to acquire and produce) when faced with the social and cognitive constraints inherent to interaction with others
- On parent report measures of vocabulary (compared to children’s spontaneous production), parents of neurotypical and autistic children underestimate a higher proportion of their child’s verb vocabulary when vocabulary size is smaller; this suggests that parent report may be more problematic for children who produce few verbs
- Neurotypical children produce verbs in progressive aspect earlier than and more than verbs in perfect aspect, in line with structure-building theories of language acquisition; in future work, we will broaden this sample to determine whether the predictions of such theories extend to autistic language acquisition
LeGrand, K.J., & Naigles, L. (2023). Early use of manner and result verbs in autism and typical development. Poster presented at the International Society for Autism Research (INSAR) Annual Meeting.
LeGrand, K.J., & Naigles, L. (2023). Parent underestimations of verb vocabulary in ASD and TD. Talk presented at the International Society for Autism Research (INSAR) Annual Meeting.
Wing, E., LeGrand, K., & Diercks, M. (2026). Progressive aspect is acquired before perfect aspect. Poster to be presented at the 51st Boston University Conference on Language Development (BUCLD).