1.4 Pathways for Development 25long as they can speak ). Here “people making excuses” is unlikely to change,
and thus the temporal meaning is not salient. Finally, as long as will appear in
contexts where only the conditional meaning is possible (e.g., I don’t care as
long as I get paid ≠ I don’t care for the length of time I get paid). At this stage
the conditional meaning has become the new coded meaning (see Traugott and
Dasher 2002 : 36– 37).
Semantic changes in the development of the comment clause I mean can be
understood in this framework (see Brinton 2008 : 129– 130). The original lexi-
cal meanings of mean are ‘to intend [to do something]’ or ‘to signify, intend
to convey.’ When I mean is used in the context of appositional structures with
the ‘intend to convey’ meaning, a previous element is restated or reformulated,
and the invited inferences arising lead to the many different uses we see for the
comment clause. An inference of restating is that the previous utterance is being
corrected ( I’ll see you tomorrow morning, I mean, afternoon ). Another infer-
ence is that the information in the previous utterance is being more precisely
stated ( The situation could be embarrassing, I mean, politically ). One could
infer that contrastive information is being presented ( I just want to look at it,
I mean examine the evidence ). The information presented may be inferred to be
a particular instance or instantiation ( He really knows how to cook. I mean he
can even make souffl és ). The information can be inferred to be accurate ( He’s
rich, I mean, really rich ). One could infer that the speaker is evaluating the
previous information ( It was humiliating to be beaten, I mean, (truly) humiliat-
ing ). And fi nally, one could infer that the speaker is being sincere in expressing
the previous information ( I felt like walking out. I mean I wouldn’t, but ... ).
The rise of (inter)subjective meanings has also been the focus of much
attention in pragmatic- marker research. Synchronically, “subjective ” refers
to the speaker’s expression of self, including beliefs and attitudes ( Traugott
2003b : 125; Traugott and Dasher 2002 : 19– 23), while “intersubjective ” refers
to explicit attention to the self of the addressee, both epistemically and in a
social sense ( Traugott 2003b : 128; Traugott and Dasher 2002 : 23– 24). The
associated diachronic processes can be defi ned as follows: Subjectifi cation
is the process “whereby meanings tend to become increasingly based in the
[speaker/ writer]’s belief state or attitude toward what is being said and how it
is being said” ( Traugott 2003b : 125), while intersubjectifi cation is “a mech-
anism whereby meanings become more centred on the addressee” ( Traugott
2003b : 129). (On subjectivity and intersubjectivity, see López- Couso 2010 .)
Subjectifi cation and intersubjectifi cation are well attested in the develop-
ment of pragmatic markers (cf. Degand and Simon- Vandenbergen 2011 : 287).
Traugott and Dasher ( 2002 : Ch. 4) discuss the rise of subjective and intersub-
jective meaning in a number of pragmatic markers. The epistemic meanings of
indeed , in fact , and actually are clear examples of subjectifi cation, while the
hedging senses of actually and in fact are the result of intersubjectifi cation.