"Modality-neutral LRs". For each word (or morpheme etc.) there will be one LR in a code neutral with respect to modality. In keeping with the terminological conventions adopted in this essay, such a LR will look like this: "cat". Under this model, modality-specific information will have to be translated into the neutral code before the LR can be accessed.
- Butterworth (2004), Pag. 25