< previous page page_249 next page >

Page 249
predicates which behave syntactically as factive have this semantic property, and almost none of those which behave syntactically as nonfactive have it." (Kiparsky and Kiparsky 1970). However, those factive verbs that are at the same time verbs expressing propositional attitudes sound very embarrassing within the present framework for at least two reasons.
The first reason is that, for representing other p-terms, one can have recourse to certain primitives that, even though they should be interpreted in their turn (outside the framework of a discussion on presuppositions), can nevertheless play a certain provisional roleas if they were already analyzed. In verbs of propositional attitudes, one runs the risk of circularity, or tautology. To know means to be aware of something which is the case, to be aware means to know that something is the case, and so on. It seems that a propositional attitude cannot be interpreted except in terms of another propositional attitude.
The second reason is that these verbs seem to react differently to the negation test, according to the person by which they are expressed. It seems that (26) posits problems that (27) does not:
(26) I don't know that p.
(27) X does not know that p.
In the case of (27), there are no problems. The embedded clause p is presupposed and still taken for granted by the p-term know, even though it is denied that a given subject knows that p. On the contrary, (26) sounds odd and, as a matter of fact, a subject who, outside the lamented cases of "examplese," uttered (26) would be considered mentally disturbed. It is, however, interesting to ascertain by semiotic means what kind of mental disturbance can be imputed to the utterer of (26).
Our opinion is that the case of verbs expressing propositional attitudes cannot be solved from the point of view of the normal use of natural language, since these verbs in natural languages are taken equivocally. It is not by chance that for many centuries philosophy has been obsessed by questions such as What does it mean to know? to be aware of? to have a mental representation of? and so on. By using these verbs, language is speaking of itself, or at least of a phenomenon of which it itself is a part (be it a cause or an effect). A plausible treatment of these verbs is given within the framework of a formalized epistemic and doxastic logic, where expressions such as know or believe are taken as primitives whose conditions of use are strictly (and narrowly) made clear. As a matter of fact, such formalizations do not capture the common and ev-

 
< previous page page_249 next page >