The only proposed analysis that can be extended to accommodate soliloquial ne is
Takubo and Kinsui’s (1997) and Kinsui and Takubo’s (1998) Discourse Management
Model. Although they assume that ne is used only interactionally, i.e. in the presence of
an addressee, they nevertheless attempt to explain its function without recourse to an
addressee’s assumed knowledge about a given topic. To this end, they posit a level of
representation that works as a cognitive interface between speech forms and the
speaker’s knowledge stored in his/her memory, which is conceived as a database in their
model. This interface is analogous to a buffer in a computer, i.e. special memory used to
temporarily store input or output data, and is divided into two psychological domains:
The direct experience domain (D-domain) and the indirect experience domain
(I-domain).
At the beginning of each discourse, information about the discourse situation (a
part of direct experience) and general information that the speaker considers relevant to
the coming discourse stored in his/her permanent memory become highlighted, and
indices of, or pointers to, such data are temporarily stored in the D-domain. At the same
time, a unique I-domain is constructed specifically for the purpose of each discourse. In
order to speak, Takubo and Kinsui argue, the speaker interprets information in the
D-domain into conceptual terms and stores it in the I-domain. (It is the indices that are
stored in these domains, but for the sake of exposition, the idea is simplified and stated
as if information itself is stored in them.)