FG from its inception 37(1973b: 835) himself remarks, the “semantic content of a linguistic expres-
sion” is “the information which, given the role of context and of
extralinguistic knowledge, it is necessary and sufficient to assign to a lin-
guistic expression in order to explain its different final interpretations”.
This shift, I submit, lies at the origin of the problem of underlying repre-
sentations (PR2), because Dik wants the underlying representation to do
two different tasks: to structure the information that contributes to the in-
terpretation of a linguistic expression (URs as ‘semantics’) and to structure
the instructions (that is, the functional relations between the information
units) that will generate a linguistic expression (URs as ‘syntax’).^20 Dik
(1986a: 11) clearly indicates this when he states that they “...are meant to
contain everything that is needed to retrieve the semantic content of the
predication on the one hand, and to specify the form of that expression on
the other”.
The problem is illustrated by Dik’s pragmatic explanations of linguistic
phenomenon such as definiteness. Does the definiteness operator in under-
lying representations represent the intrinsic definiteness of a term (a
semantic UR) and so is included for any intrinsically definite term,^21 or
does it represent a trigger to generate the ‘definite article’ where need be?^22
Neither account seems sufficient even for English, with its fairly restrictive
use of the definite article. Consider the following three sentences, anno-
tated with both types of UR:
(7) a. Mr Smith went home grumpy.
intrinsic: (dsx: ‘Mr Smith’: ‘grumpy’)
generative: (sx: ‘Mr Smith’: ‘grumpy’)
b. The grumpy Mr Smith went home.
intrinsic: (dsx: ‘Mr Smith’: ‘grumpy’)
generative: (dsx: ‘Mr Smith’: ‘grumpy’)
c. A grumpy Mr Smith went home.
intrinsic: (dsx: ‘Mr Smith’: ‘grumpy’)
generative: (isx: ‘Mr Smith’: ‘grumpy’)
The expression rules component would obviously prefer the ‘genera-
tive’ UR as input, but this is plainly at the expense of semantic accuracy.
A further complication arises when we read Dik’s explanation of defi-
niteness (1978a: 61; cf. 1973a): “The difference between ‘definite’ and
‘indefinite’ can in my opinion only be captured in pragmatic terms: by
means of a definite term the Speaker expresses the fact that he acts on the
presupposition that the Addressee can identify the particular intended refer-