1.5 Processes of Change 33(i) Speakers recognize the potential inherent in certain forms for expressing
textual and interpersonal meaning and “abuse” them by using them in new
contexts.
(ii) As the textual/ interpersonal forms become more common, speakers reana-
lyze them as pragmatic markers in come contexts by conventionalizing an
invited inference.
(iii) Finally, speakers start using the pragmatic marker in additional contexts.
Ocampo ( 2006 : 317) postulates a process of “discursivization”^36 which names
“the diachronic process that ends in discourse,” either from lexicon to dis-
course or from grammar to discourse. Both grammaticalization and discursivi-
zation involve a change from concrete > abstract, autosemantic > synsemantic,
and non- relational > relational. Ocampo sees discursivization as sharing prop-
erties with grammaticalization, but remaining distinct (318).
1.5.4 Grammaticalization, Lexicalization, or Pragmaticalization?
Which process, then, best accounts for the rise of pragmatic markers?
A number of arguments can be adduced against the view that pragmatic
markers are lexicalized. The end product of lexicalization is structures that
express contentful (propositional/ referential) meaning and instantiate the
major categories – nouns, verbs, and adjectives, i.e., lexical items. Pragmatic
markers, however, do not behave like lexical items: They cannot be understood
as belonging to any major lexical category, but rather to a minor functional cat-
egory (if any category at all); they do not express lexical (propositional) mean-
ing, but rather non- lexical (non- propositional/ procedural) meaning; and, unlike
lexical items, they are syntactically and prosodically constrained. As Diewald
observes, pragmatic markers “do not display lexical semantics in the narrow
sense and therefore cannot be used to denote elements of the propositional con-
tent of the sentence” ( Diewald 2006 : 404). A number of the central features that
pragmatic markers share with the grammaticalization process are not found
in the process of lexicalization: These include decategorialization , semantic
bleaching, and (inter)subjectifi cation (see Brinton and Traugott 2005 : 107–
109; Brinton 2008 : 65– 66). Moreover, while there is typological generality
in both grammaticalization and the development of pragmatic markers – we
fi nd similar forms developing into grammatical markers cross- linguistically^37 –
such generality is not the case with lexicalization. Grammaticalization typi-
cally leads to an increase in productivity since a pattern becomes more widely
applicable; we see this to a certain extent in pragmatic markers (where, for
36 This is distinct from the notion of “discursisation” presented by Claridge and Arnovick
( 2010 : 183– 185).
37 See, e.g., the studies published in Aijmer and Simon- Vandenbergen ( 2006 ).