166 Epistemic Parentheticals
(epistemic/ evidential) and intersubjective (interactive, politeness) meanings.
(see Traugott and Dasher 2002 : 40).
Wischer ( 2000 ) argues – in the case of methinks – that both grammati-
calization and lexicalization are involved. She sees an originally productive
construction consisting of a RECIPIENT + impersonal verb as becoming fos-
silized, partly demotivated, and unproductive, now stored as a whole in the
lexicon (as a “symbol”). This process she understands as lexicalization – a
syntagm becoming a new lexical item or “the symbolifi cation of a former free
collocation” (364). Me thinks is stored as a whole entity in the lexicon and
classifi ed as an “adverb.” But this lexical item then immediately grammatical-
izes to become an disjunct marker of evidentiality (a grammatical function) on
the discourse level; it undergoes syntactic reanalysis, subjectifi cation , phonetic
attrition, extension of scope, and preference for certain sentence positions.
Wischer intends this scenario to account only for methinks , and notes that its
lexicalized nature is obvious when compared, for example, to I think , which is
not lexicalized. But Wischer provides no evidence to support the sequence of
lexicalization followed by grammaticalization. An equally plausible sequence
would be grammaticalization followed by lexicalization in this case.
Resting her argument on the view that “[i] n non- standardized languages
[epistemic parentheticals] are likely to form one lexical unit in the course of
time” and that morphological bonding may occur, Fischer ( 2007a : 109– 112,
2007b : 308– 311) argues for lexicalization rather than grammaticalization. She
does not provide any evidence for treating epistemic parentheticals as “lexical
units,” as they do not constitute a major lexical category. Moreover, she argues
that persistence (of the original meaning), divergence (between old and new
forms), and decategorialization , which we see in this case, are common to lexi-
calization and grammaticalization. However, while persistence of the original
meaning may occur in lexicalization, it frequently does not, as the lexicalized
form becomes semantically non- compositional (e.g., nuts- and- bolts ‘essen-
tials’ retains none of its original meaning). Divergence, in which the older and
newer form coexist, may also occur in lexicalization, but it frequently does
not, as the unlexicalized form is replaced by the lexicalized form (e.g., handi-
cap replaces hand in cap ). And decategorialization is clearly not a feature of
lexicalization, but only of grammaticalization (see Brinton and Traugott 2005:
107– 108). Fischer argues that epistemic parentheticals “lose some referen-
tial content, being narrowed down to a more epistemic, evaluative meaning”
(2007b: 311); in fact, this points to grammaticalization, not lexicalization,
since lexicalization is characterized by the addition rather than loss of meaning
(Brinton and Traugott 2005 : 108). Fischer ’s argument rests, ultimately, on her
sense that epistemic parentheticals retain “more meaning than is usual in the
grammaticalization of discourse markers” (2007b: 309) – a claim that would