Professional Documents
Culture Documents
Porfessor Juan Barba for the inter-university Spanish Master in Logica y Filosofa
de la Ciencia during the academic year 2015-2016. In the first part I will discuss
the themes and the related ideas that arose during the presential lessons in Sala-
manca, in the October of 2015. In the second part I will briefly discuss the subjects
of the articles written for the course and the sparks came out in the scholar forum
with my fellow students.
1
because on this perspective every contradictions has the same meaning (squared
circle, dark light, etc...)
In general terms, an intensional logic expression is a function of function, in
turn function of functions and there are not limits to its embedding and nesting.
For example, take the verb to be: it is defined as a function from the set of possible
worlds, at a specific time, to the set of entities, and equivalently in a function from
the set of possible world, at a specific time, into a set of (other) possible worlds,
at (another) specific time, which is again a function to the set of entities, into
the set of truth values. Hence Montagues grammar is one of the first practical
attempts to calculate higher order predicate. Its effect is to maintain constant
the extension of the description but constantly increasing its intension. The order
of the predicate calculus on which we rely establish the degree of comprehension
reached, for those predicates. Take for example the predicate to not exist: it
describes something, but a predicate of this predicate could precise it could exist,
or even under which circumstances. Every order refines the level of knowledge.
To solve the issues of passive forms (it is clear that in natural language P
knows M and M is known by P are equivalent) we have then used the abstractor
lambda. Adopting the tool of generalized quantifier to express nominal sintagmas
we arrived to generate a significant fragment of Spanish language. A generalized
quantifier corresponds to a set of subsets, in the sense that it associates a truth
value to every subset of a set of properties.
Standard quantifiers of first order predicate logic are related to sentences such
as the following:
(x)(John loves x)
However, it is still possible to reinterpret the predicates as expressions denoting
properties of set of individuals. Let express the property of a set to be not empty
and express the property denoting the whole domain D (equivalent to say the
set of all elements). They became now unary predicates of type << e, t >, t >, to
which we can apply a predicate of type < e, t > to obtain a well formed sentences
of type < t >. Example:
2
(x.John loves x) = 1 iff {x :John loves(x) = 1} =
6
Following this approach the standard quantifiers are of the same logical type of
generalized quantifiers. It remains a gap, thou: combines only with the predicate
type < e, t > and it gives a truth value; instead, generalized quantifiers express a
relation between two sets and it can be infinitely nesting.
We could be tempted to reduce the generalized quantifiers of natural language
into unary quantifiers of classical logic. Actually, the normal translation of sen-
tences like the following are of these sorts:
3
First, given the fact that CP has a specified content only if encored in synctactical-
semantics theories which bind their theoretic terms (meaning, constituent, syn-
tactical structure, and determine), there is still no unanimity on what should
be the right formulation of CP. Second, we still dont have a formal proof of the
validity of CP. Third, there is not yet an agreement on which role CP should cover
in the semantics of natural language: it is not clear yet, for example, wether CP
should be considered as an empirical hypothesis, to be discussed and analyzed on
the light of empirical datas and linguistic intuitions, or rather as a methodological
principle, as theoretical bind on the construction of semantics theories.
Consider the following example: suppose that the meaning of a linguistic ex-
pression is the Fregean Bedeutung; thus, suppose that the meaning of proper names
and definite descriptions is the object to which they refer; suppose furthermore
that the meaning of a predicate is a function and that the meaning of a declarative
sentence is its truth value. In this case the content of CP is specified as: the refer-
ence of a complex sentence is determined by the reference of its constituent and its
syntactical structure. But this formulation clearly fails in metaphoric expression.
The second reflection goes on the nature of CP: is it an empirical hypothesis
or a methodological principle? If we admit the former then we accept it or deny
it, like any other empirical hypothesisi; if we accept the latter there is no reason
to commit to such a drastic perspective: we will just adapt our theories in such a
way that it conform to CP.
Another theme of the course was the one regarding time, tense and aspect in
natural and formal languages. This subject let the discussion turn to the problem
of the principle of causality. The issue of time made logicians develop a more com-
plex model theory, which include the time specification. About this, except for the
little more complicated logical formalism, there is nothing obscure. The problem
arises trying to characterize perfective and imperfective aspect, say Simple Past
and Past Progressive. Think on the following two sentences:
4
In the first situation the explanation of Juan is a completed action, while in the
second the explanation is still happening and it might be not over yet. Whether the
predicate of the sentence is in perfective or imperfective aspect also affect when
the predication can be understood to hold of the subject. Unfortunately Tense
Logic Semantics doest have the tool to differentiate the two situations. To avoid
this problem we have seen two solutions: Vendlers Aktionsarten and Shanahans
Event Calculus.
Strictly connected with the conception of event as ordered triple is the one of situ-
ations. Situations, just like events, can be formulated in a partial model theory. In
opposition to Montagues grammar based on a possible worlds semantics, Barwise
and Perry attempted to characterize expressions with respect to partial, rather
than complete, worlds. With this approach we can include more phenomenolog-
ical aspect of an utterance, since the expression is always strictly related to a
situation, a subset of a possible world. Hence to provide interesting sparks to the
solution of ambiguities and vagueness main residents of our natural language.
Rebecca Cacioli