@inproceedings{malaia:26027:sign-lang:lrec,
  author    = {Malaia, Evie A. and Krebs, Julia and Harbour, Eric and Martetschl{\"a}ger, Julia and Schwameder, Hermann and Roehm, Dietmar and Wilbur, Ronnie B.},
  title     = {The Displacement-Velocity Dissociation in Sign Language Learning: Kinematic Signatures of Event Structure in Novice {{\"O}GS} Signers},
  pages     = {324--332},
  editor    = {Efthimiou, Eleni and Fotinea, Stavroula-Evita and Hanke, Thomas and Hochgesang, Julie A. and Mesch, Johanna and Schulder, Marc},
  booktitle = {Proceedings of the {LREC2026} 12th Workshop on the Representation and Processing of Sign Languages: Language in Motion},
  maintitle = {15th International Conference on Language Resources and Evaluation ({LREC} 2026)},
  publisher = {{European Language Resources Association (ELRA)}},
  address   = {Palma, Mallorca, Spain},
  day       = {16},
  month     = may,
  year      = {2026},
  isbn      = {978-2-493814-82-1},
  language  = {english},
  url       = {https://www.sign-lang.uni-hamburg.de/lrec/pub/26027.html},
  abstract  = {This study investigates how adult learners acquire linguistically contrastive movement patterns in Austrian Sign Language ({\"O}GS), focusing on the telic/atelic distinction predicted by the Event Visibility Hypothesis. Telic verbs (bounded events) are produced by proficient Deaf signers with shorter duration and temporally precise, low-entropy velocity profiles, whereas atelic verbs (unbounded processes) show more continuous motion. Using 3D motion capture (300 Hz), we compared 8 novice learners (6--12 weeks of instruction) with 6 proficient Deaf signers across 71 verbs. Linear mixed-effects models revealed a dissociation between gross movement patterning and fine-grained velocity profile structure in learner productions. Learners correctly reproduced the proportional path-length contrast between telic and atelic verbs, replicating the gross spatial distinction of proficient signers. However, temporal marking of the telic/atelic contrast was underproduced: learners showed a significantly smaller duration difference between verb types than proficient signers, while total path length did not differ significantly between verb types or groups. Temporal control showed significant between-group differences: learners exhibited elevated sample entropy, with non-proficient velocity profiles within individual sign productions, though spatial consistency across trials (STI) was comparable to that of proficient signers. Peak velocity did not differ between groups, suggesting that learners can reach target speeds but cannot yet modulate temporal structure reliably. These findings support distinct learning trajectories for gross movement patterning and fine-grained motion complexity, and demonstrate that velocity profile structure within signs constitutes a core linguistic target in sign language learning.}
}

@inproceedings{sazonov:26056:sign-lang:lrec,
  author    = {Sazonov, Dmitriy and Gurbuz, Sevgi and Malaia, Evie A. and Martetschl{\"a}ger, Julia and Schwameder, Hermann and Roehm, Dietmar and Wilbur, Ronnie B.},
  title     = {Lost in Expression: Diagnosing Systemic Challenges with Non-Manual Generalization in Sign Language Understanding Tasks},
  pages     = {438--449},
  editor    = {Efthimiou, Eleni and Fotinea, Stavroula-Evita and Hanke, Thomas and Hochgesang, Julie A. and Mesch, Johanna and Schulder, Marc},
  booktitle = {Proceedings of the {LREC2026} 12th Workshop on the Representation and Processing of Sign Languages: Language in Motion},
  maintitle = {15th International Conference on Language Resources and Evaluation ({LREC} 2026)},
  publisher = {{European Language Resources Association (ELRA)}},
  address   = {Palma, Mallorca, Spain},
  day       = {16},
  month     = may,
  year      = {2026},
  isbn      = {978-2-493814-82-1},
  language  = {english},
  url       = {https://www.sign-lang.uni-hamburg.de/lrec/pub/26056.html},
  abstract  = {Incorporation of non-manual information is one of the most challenging aspects of Sign Language Understanding (SLU), as these features contribute to the semantic, syntactic, and pragmatic structure of signed communication as a critical feature of compositional meaning at sign, phrase and sentence level. Despite their key linguistic role, non-manuals are often an afterthought in SLU model and dataset design, with many recent models still neglecting to implement non-manual analysis or evaluate how articulators beyond the hands are contributing to the model prediction. In this work, we identify and analyze the challenges relating to recognition of non-manuals and generalization of their linguistic roles encountered by SLU models, offering new explanations for failures to properly model non-manual behavior. We perform a case study on the subtasks of Continuous Sign Language Recognition and Sign Language Translation by applying the Uni-Sign model to Isharah-1000, a Saudi Sign Language dataset. Using controlled partitioning and feature attribution, we further analyze model behavior and failure cases. With this work we hope to set the stage for the creation of diagnostic frameworks for generalization of non-manuals.}
}