296 Concluding Remarks: Pathways of Change
additional parameters must be added: general coherence (e.g., speech acts,
speaker information, hearer expectations, shifts in discourse), politeness (e.g.,
interactional and conversational constraints, norms of politeness and tact), and
involvement (e.g., speaker’s feelings and opinions).
There has been considerable work on diachronic Construction Grammar
over the past decade (e.g., Bergs and Diewald 2008 ; Bergs 2012 ; Hilpert 2013 ;
Traugott and Trousdale 2013 ; Barðdal , Smirnova , Sommerer , and Gildea 2015 ),
but very little attention has been devoted to the rise of pragmatic markers in this
framework. A few studies may be mentioned here. In a detailed examination
of the syntactic complementizer jestli ‘if whether’ in Czech, which comes to
function as a subjective modal particle meaning ‘I think/ guess, maybe,’ Fried
(2009) argues that this change occurs in the context of yes/ no questions: More
generally, “the meaning/ function of JESTLI cannot be determined outside of
specifi c constructions; hence, we have to take constructions as the domain of
change” (289). Fried postulates the existence of a transitional phase in which
the jestli - clause is “emancipated” and becomes a syntactically independent or
freestanding clause whose meaning and function incorporate the semantics of
the main predicate (‘lack of factual knowledge’) (274). Such a stage would not
be consistent with a straightforward grammaticalization analysis. In the fi nal
stage, jestli is fully grammaticalized as a modal particle, with narrower scope.
The relatively “holistic approach” of Construction Grammar, which focuses on
the conventionalization of complex patterns and incorporates speakers’ knowl-
edge of discourse patterns and textual organization, Fried argues, combined
with a view to internal mechanisms of change, can best explain the develop-
ment undergone by jestli.
Brinton (2008: 255– 256) suggests that from the variety of epistemic con-
structions found in Middle English (see Chapter 5 , present volume), a more
abstract meso- construction might have been abstracted, consisting of I plus a
present- tense verb of cognition. “Less grammaticalized forms (those showing
variety in their syntax), would begin to inherit the more abstract properties of
the meso- construction and be drawn into the set” (256). The meso- construction
would also license the rise of constructions incorporating new (often newly
borrowed) verbs such as expect or suspect. Here a Construction Grammar
view allows us to postulate similarities across a set of forms sharing semantic,
syntactic, and pragmatic properties. Brinton ( 2014b : 288– 290) points out that
Construction Grammar may also provide a means for understanding the devel-
opment of the comment clauses if you choose/ like / prefer / want/ wish ; as has
been discussed, these cannot be traced back to a full biclausal structure with an
explicit apodosis and a complement of the verb of ‘choice’ (see Section 10.3.2
above, example 1b), which becomes reduced through grammaticalization.
Assuming a Construction Grammar perspective, we may view the develop-
ment as follows: We begin with a variety of elliptical second- person if - clause