Working memory visual world linking hypothesis

This is a schematic I created that depicts a linking hypothesis for the visual world paradigm (Tanenhaus et al., 1995) proposed by Huettig, Olivers, & Hartsuiker (2011). They propose that items in the visual display must enter working memory to be bound to semantic and phonological representations. They provide a schematic (their Figure 2), but it is incomplete and suggests that spoken input directly enters the visuospatial working memory buffer. This is my attempt to make explicit the minimal components required to depict this view.
Version 2 better conforms to the proposal of Huettig et al. (2011).

Some elements (grids, silhouettes) are from publicdomainvectors.org, where all posted items are available with a CC0 license
(https://publicdomainvectors.org/en/public-domain/)