The present paper aims at re-applying Kibrik's Framework (1999, 2000, 2009, 2011) with some considerable modifications as a new cognitive linguistic channel for evaluating certain textual/discoursal representations. This is achieved through adding the acoustic dimension (with several sub-divisions) to Kibrik's Model/Framework, and applying the new version to certain speeches. These speeches are so selected as to represent outputs by native, near-native and non-native speakers of English. Major conclusions exhibit that, in the case of native-speaker discourse, the missing category of FULL NP OR PRONOUN can be compensated by the WM operating as control or maximally in full capacity to accommodate the excessive dependence on cataphora. This is not the case for near-native discourse. The net result is a WM operating on its capacity level, and not in control of the discourse, since little discourse segments, so to say, remain active. An even less coherent discourse is produced by non-native speakers. Both categories of FULL NP ONLY and PRONOUN ONLY are not there, which highlights the incomplete operations of the WM.