128 Epistemic Parentheticals
and should be accorded to, the statement to which they are conjoined” (485).
Similarly, Benveniste ( 1971 [1958]) emphasizes that the verbs in this usage do
not indicate an “operation of thought” but rather “a certain attitude with regard
to the utterance that follows” (228). Benveniste further notes that when the
verbs are followed by that they denote the real thought operation, not the per-
sonal verb form (228– 229). Hooper ( 1975 ) discusses a class of what she calls
“weak assertives,” including think , believe , suppose , guess , expect , appear ,
imagine , and fi gure , which allow “complement preposing ” (see Section
1.4.1.2 ), or the formation of parentheticals, especially in the fi rst person, pres-
ent tense, where the semantic content “makes no assertion in itself but, rather,
only describes the speaker’s attitude toward the truth of the asserted proposi-
tion, which lies in the complement” (101). Evidence for the function of these
verbs derives from the behavior of tag questions, which are formed on the
complement not the main clause ( I think this car needs a tune- up, doesn’t it?
don’t I ), from the existence of neg- raising ( I think these living conditions are
not suitable = I don’t think these living conditions are suitable ), from use of
the sentence pronominal so not it , and from transparency in regard to sentence
adverbials ( Fortunately I think he’s already gone = It’s fortunate that he’s gone
not It’s fortunate that I think ) (103– 111). Quirk et al. ( 1985 : 1112– 118) treat
such forms as “parenthetical disjuncts” or “comment clauses ,” whose function
is to hedge, to express (un)certainty, to convey emotional attitude, or to claim
the hearer’s attention. Biber et al. ( 1999 ) likewise refer to parentheticals such
as I think / guess/ bet/ suppose/ believe and it seems/ appears as comment clauses
(197) or as “fi nite clause stance adverbials” (864– 865) used to express “per-
sonal feelings, attitudes, value judgments, or assessments” (966).
Thompson ( 2002 ) discusses what she calls “Complement- Taking-
Predicates ” (CTP), the fi ve most common verbs being think , know , see , guess ,
and remember. They typically occur with a fi rst- person subject and constitute
“formulaic fragments” (131) that express evidentiality , epistemicity, or evalu-
ation. Thompson argues that there is no evidence that the complement clause is
an object, nor that it is subordinate: “in the majority of cases, the complement
‘overrides’ the ‘main clause’, and the ‘main clause’ is there to provide speaker
stance towards the assessments, claims, counterclaims, and proposals” (134).
Especially when used as an independent fragment, there is evidence for the
“discourse marker status par excellence” of CTPs (144). These forms are now
typically designated by the term coined by Thompson and Mulac ( 1991 ), i.e.,
“fi rst- person epistemic parentheticals .” First- person epistemic parentheticals
such as I think or I guess are standardly included among the set of pragmatic
markers in Present- day English.^2 In response to Thompson , Boye and Harder
2 However, Östman distinguishes between epistemic modality markers such as I suppose , I
believe , and maybe and pure pragmatic particles such as I guess , I think , and I suspect , which,
he says, have no propositional content and function exclusively pragmatically (1982: 152, 153).