5.2 First-Person Epistemic Parentheticals in PDE 133their primary meaning, not only as a pragmatic inference” raises two ques-
tions in respect to fi rst- person epistemic parentheticals: (a) to what extent are
epistemicity and evidentiality separate categories? and (b) is the meaning of
(un)certainty or that of mode of knowing primary in the fi rst- person epistemic
parentheticals?
The distinction between epistemicity, which concerns matters of certainty
and confi dence, and evidentiality, which concerns sources of knowledge and
modes of knowing, is a vexed one (see Cornillie 2009 : 46– 49). For Palmer
( 2001 ) “epistemic modality” and “evidential modality” are part of a larger cat-
egory of “propositional modality,” but all is classifi ed as “modality.” Chafe
( 1986 : 263) includes source of knowledge (evidence, language, or hypothesis)
and mode of knowing (belief, induction, hearsay, or deduction), as well as the
range from reliable to unreliable, into a larger domain, which he calls “evi-
dentiality” (also Biber and Finegan 1989 : 93– 94). Anderson ( 1986 : 310), too,
includes epistemic modals within the category of evidentiality. However, many
see the two categories as distinct, while admitting that evidentiality impinges
on epistemicity. Willett ( 1988 : 86; his emphasis) concludes that “the source
of a speaker’s information can skew the relation between his/ her conception
of the truth of a situation and the strength of his/ her assertion about that situa-
tion.” And Chafe ( 1986 : 266) admits that “mode of knowing implies something
about reliability but not vice versa.” The diffi culty in distinguishing evidential-
ity and epistemicity arises on the formal level because the same morphological
forms can encode both ( Chung and Timberlake 1985 : 245), but on the concep-
tual level because “reference to the knowledge that leads to a proposition is
often interpreted as a kind of evaluation of it,” i.e., epistemic speaker commit-
ment is confused with reliability of knowledge (Cornillie 2009 : 57– 59).
The second question, whether evidential or epistemic meaning is primary
in the fi rst- person epistemic parentheticals, is more diffi cult to answer. For
reasons that will become clearer in the discussion of the development of fi rst-
person epistemic parentheticals (see Section 5.6.3 ), I believe that although the
forms are in origin evidential, with epistemic meaning merely conversationally
implied, by Middle English the epistemic meaning had become conventional-
ized. In Present- day English we see that the know - verbs are frequently used
not with their original senses to refer to specifi c types of cognitive acts, but can
be interchanged quite freely. And this apparent “looseness” of meaning has
evoked prescriptive criticism. For example, the use of guess in the informal
sense ‘to suppose, imagine, think’ rather than in its strict sense ‘to form an
approximate judgment without suffi cient information’ was formerly objected
to, but is now considered acceptable ( Webster’s : s.v. guess ). The OED notes
“secondary information” and that it is this coding of secondary information which makes evi-
dentials “grammatical.”