Expressive Go:
A Rule-Based AAC Architecture for Supporting Spontaneous Novel Utterance Generation

Technical Overview and Clinical Rationale for Dynamic Grammatical Computation in AAC Systems

The development of Augmentative and Alternative Communication (AAC) systems is central to supporting individuals with complex communication needs. Expressive Go introduces a paradigm shift in high-tech AAC by embedding a sophisticated, rule-based computational grammar engine into the user interface. Rather than relying on static phrases, this approach prioritizes the user’s ability to generate Spontaneous Novel Utterances (SNUG) and produce linguistically accurate messages in real time.

1.1 Problem Statement: Limits of Static AAC Systems

Traditional grid-based AAC systems often present significant barriers to linguistic development, compelling the need for advanced automated grammatical support.

  • Grammatical Inaccuracy: Symbol-based message construction frequently leads to omission of obligatory morphology and syntactic errors. As Sutton, Soto, and Blockberger (2002) explain, such errors limit the user’s ability to construct complex, well-formed messages and reduce opportunities for natural language learning.
  • Cognitive Load Barrier: The manual steps required to assemble grammatically correct sentences—searching for and selecting tense markers or plural forms—impose a high cognitive and motor workload. Research confirms that these demands can significantly slow conversational flow and impede spontaneity.
  • Static Content Restriction: Systems built on pre-stored utterances inherently restrict users to messages anticipated in advance by others (Todman et al., 1995). This design prevents the user from achieving genuine generative language development.

1.2 Solution Overview: Dynamic Grammatical Computation

Expressive Go addresses these technical and clinical limitations through the Smartest Language Engine (SLE), a computational architecture designed for real-time linguistic processing.

  • Contextual Morphological Processing: The SLE actively and automatically applies rules governing tense, aspect, agreement, and pluralization based on the immediate word context. This feature directly addresses the clinical finding that children who use AAC frequently omit grammatical morphemes (Binger and Light, 2008).
  • Rule-Based Accuracy: Computational linguistic supports have been shown to improve grammatical accuracy and reduce the cognitive demands of message formulation (Venkatagiri, 1999). Expressive Go’s internal rule libraries and irregular verb lexicons operationalize this need, ensuring accurate inflections automatically.
  • Visual–Linguistic Linkage: The platform utilizes color-coding, which, as Mirenda (2014) notes, provides an important scaffold for teaching language structure. Expressive Go integrates a Fitzgerald Key–based color system to reinforce the linguistic patterns applied by the SLE, supporting both comprehension and learning.