Quantification under Conceptual Covers
Maria Aloni

Abstract:
%Nr: DS-2001-01
%Author: Maria Aloni
%Title: Quantification under Conceptual Covers

Someone has killed Spiderman. After a careful investigation you
discover that John Smith is the culprit and now you want to
arrest him. He is attending a masked  ball. You go there, but you
do not know what he looks like. Is the sentence `You know who
killed Spiderman' true or false in such a situation? On the one
hand, the sentence is true, you know that John Smith did it. On
the other hand, the sentence is false. Since you do not know what
he looks like, you cannot point him out. As far as you know, this
person  here might be the culprit, or that person there. The
evaluation of this sentence seems to be dependent on the way in
which  the relevant individuals are specified. These can be
identified by a number of methods like naming (John Smith, Bill
White, and so on) or ostension (this man here, that person there,
and so on). If identification by name is assumed, the sentence is
true. If identification by ostension is assumed, the sentence is
false.

This example illustrates the central idea I defend and
investigate in this book. Different methods of identification are
operative in different conversational circumstances and the
evaluation of fragments of discourse can vary relative to these
methods. Classical semantic theory abstracts from the
ways in which individuals are identified and therefore has
difficulties in accounting for this dependence. The analysis I
propose represents different methods of identification and   is
able to account  for their impact on interpretation.

Questions, propositional attitude reports, and quantified
sentences containing epistemic modals are examples of linguistic
constructions whose interpretation depends on the ways in which
objects are given to us. In this thesis I will study these three
constructions using the partition theory for questions; modal
predicate logic for propositional attitudes; and an intensional
dynamic semantics for epistemic modals, respectively. These three
theories make crucial use of the notion of a _possible world_. 
Possible worlds are evaluation points where expressions
of the language receive a denotation. In the present context,
worlds receive an information-oriented interpretation. A world is
meant as representing an epistemic or doxastic possibility, that
is, a possible description  of what is the case which is
compatible with someone's information or belief. The
interpretation of questions,   propositional attitudes, and
epistemic modals crucially involves a shift from one world of
evaluation to another. Notions which behave in such a way are
usually called _intensional_ notions.

The context sensitive constructions that I will consider are
classically represented by logical formulae which contain some
variable occurring free in the scope of such an intensional
operator. In ordinary logical systems, variables are taken to
range over bare individuals, and for this reason these systems do
not account for the dependence of such constructions on the way in
which these individuals are identified.

The analysis I propose maintains the classical representation of
this type of sentences, but accounts for their meaning by
proposing a non-standard interpretation of variables in
intensional contexts.  One part of my proposal consists in
letting  variables range over functions from worlds to objects,
rather than over the objects themselves. These functions are
traditionally called intensional objects or _individual_ concepts},
 as they formalize (different) ways of identifying objects.
 The other part consists in making quantifiers range over sets of
concepts which (a) are contextually determined and (b) satisfy the
following natural constraint: in each world, each individual is
identified by one and only one concept in the relevant set. I
will call sets of concepts which satisfy this constraint
_conceptual covers_. A conceptual cover represents a method
of identification. Different conceptual covers represent
different ways of looking at one domain. By adopting
quantification under conceptual covers in the three previously
mentioned theories, the interpretation of questions, propositional
attitudes, and epistemic modals are made dependent on the
conceptualizations of the universe of discourse which are
pragmatically operative. I will show that such a relativization
 enable us to solve a number of traditional difficulties, and new
ones, which emerge in connection with these notions; at the same
time we avoid the specific problems which normally arise  when we
quantify over concepts rather than objects.