4.2.1 |
DEFINITION AND CRITERION |
We have called the imaginary relationship between a
proposition or system of propositions on a certain level and
the reality below that level
"correspondence". If it may be
said to exist, the proposition or system of propositions in question
is 'true'. In other words, on this view truth consists in some form of
correspondence between a proposition, belief or suchlike propositional
entity and the
factual, modal and normative conditions.
The correspondence
relationship itself is imaginary, since the real relationship is
the one between the things in the lower-level reality the
proposition (or belief) is about. It is this relationship which
is factual, modal or normative, and which exists, if the factual,
modal or normative proposition is true. For example, if A and B
are friends, there is a relationship of friendship between A and
B, and then A and B are friends is true. Hence, A and B
are friends is true, because there is a real relationship of
friendship between A and B. It is only an indirect way of
speaking to say that there is, then, a 'correspondence' between
the proposition A and B are friends and the fact that A and B
are friends, or the factual friendship between A and B. The same
applies to the fact that a thing C has an attribute D. If C is
happy, for instance, there is
a relationship of having-as-an-element
between C and an attribute of happiness, and then C has
(an attribute of) happiness or C is happy is true. The
correspondence relationship between the true proposition C is
happy and the fact that C is happy is therefore, strictly
speaking, imaginary. What counts ultimately is that there is
really the relationship of having-as-an-element between C and
the happiness attribute D.
Our 'relational theory of truth' is quite simple where it
concerns the truth of propositions which assert the existence of
one-or-more-place relationships between things, including those
between things and attributes. It is a little bit more complicated
with respect to propositions such as A exists in which A
is a name. At first sight it might seem that this proposition
does not involve any relation, even not the one of having-as-an-element.
This is misleading, however, as names are somehow
abbreviations for definite descriptions (like the person born
at that particular moment at that particular place or the
person who(se body) is the first common child of that
particular woman and that particular man). If an ordinary name
has a sense, it may thus be equated with a so-called 'co-designative
definite description', that is, a definite description which
refers to the same object. It must be admitted that a name may
then have different meanings for different people (the most
notorious one probably being God). It has also been suggested
that every name is 'loosely associated with a set of descriptions'.
This may be correct in
practise, but it will not
do when considering the question whether it is true that A exists.
Then we need one definite description or set of descriptions for A
(ultimately without using any name in these descriptions themselves).
Given the description, or set of descriptions, the
question has become whether there really is the particular kind
of relationship between the particular things, or types of
thing, mentioned in each description. If so, then it is true
that A exists, if not, then A exists is false. (It has been
denied that a definite description would give the sense of a
name. Yet, it is then still the reference of a name which must
be fixed by means of some definite description. It is such a
definite description which denotes the referent A which is said
to exist in the actual world, not just in some possible world,
unless the proposition is that A could exist or could have
existed.)
If the imaginary relationship between a proposition and a
factual, modal or normative condition on a lower
propositional level may be said to exist,
this is a 'fact' or 'factual condition of correspondence'; if it
can exist, this is a 'mode' or 'modal condition of correspondence';
if it should exist, a 'norm' or 'normative condition of
correspondence'. Anything that has to do with the relations
between propositions themselves, however, is not a condition of
correspondence, but just a first- or higher-order condition of
thought, whether factual, modal or normative. And whereas
correspondence is the key condition in the correspondence theory
of truth, it is coherence or consistency which is the key
condition in the coherence theory of truth. A synthesis of
these theories can be found by allowing correspondence the
ontological part, while using coherence as a test of truth,
playing an epistemological role. In both theories of truth there
is a more or less intimate connection with the idea of the
ultimate structure of the world, especially in the
interpretation of correspondence as a structural isomorphism,
and in the view of reality as a unified, coherent whole.
An objection against both the correspondence- and the
coherence-theories has always been that the concepts of
'correspondence' and 'coherence' are not or cannot be made adequately
clear. This objection does not concern truth in our ontological
framework, because we can exactly explain what it means that 'a
proposition corresponds with a factual, modal or normative
condition'.
We can define truth provisionally by means of correspondence,
altho we basically define
it in terms of a certain type of relationship (other than correspondence)
between two or more particular things or types of thing (of the same
propositional level) being there or not being there (while being
there itself may be factual, modal or normative).
The correspondence theory provides, then, the definition of the
word true and the coherence theory only a criterion or
test by means of which to tell whether something can be true or is
false. Since coherence is not a sufficient criterion anyhow to
determine whether a theory is true or not, it does not matter
that much that the meaning of coherence itself is vague to a
certain extent. If a theory is definitely inconsistent, because
it affirms, for example, both that a thing has a certain
relation and that it does not have that relation, then it is
false; but if it is nowhere inconsistent in this obvious way, it
still need not be true. That a proposition which coheres with a
certain system of knowledge must be consistent with that system
in that it does not imply a contradiction, is not what makes
coherence a vague notion; what makes it vague is that there
is to be some 'systematical connection' with that system of
knowledge, at least if the system is to allow for empirical
propositions as well. (Historically, coherence was mistakenly
presented as a sufficient criterion of truth for deductive
systems based on a limited number of fundamental postulates.)
Not everyone has always agreed that there is a distinction
between definition and criterion. Thus pragmatists have argued
that the meaning of a term is correctly given precisely by
supplying the criterions for its application. Of course, a
definition like that of truth cannot be identical to a
necessary criterion like coherence, but the difference is less
obvious between a definition and a sufficient criterion, or a
number of two or more criterions which are sufficient together.
In the pragmatic theory of truth 'true' is what is ultimately
satisfying to believe, either because the expectations such a
belief arouses are fulfilled or because it contributes to the
satisfactoriness of, and effectiveness in, the conduct of life.
Pragmatism does not work with correspondence in the definition
of truth, nor with coherence as a test of truth. It is claimed that
people just try to conserve their old belief set, while
restoring consistence when new experience comes in. Pragmatic
'truth' is a sort of warranted assertibility which characterizes
knowledge as a mere form of belief.
Theories of truth such as the correspondence- and coherence-theories
have been called "ontological theories of truth" (altho the
coherence theory is in fact more 'logical' than 'ontological').
The pragmatic theory is, then, a 'nonontological theory of truth',
and as such closely related to another nonontological theory,
the so-called 'consensus theory of truth'.
On this theory the truth of utterances depends on the possible
consent of 'all others' under ideal conversational conditions.
The aim is said to indicate what a 'discursive uptake' of claims
to validity based on sense perception means. This should be
prerequisite for making the meaning of truth adequately clear.
Hence it is here one particular method of verification which is
made a criterion or definition of truth. Several objections have
been brought against this consensus theory. One is that it does
not allow for a satisfying definition of falsehood. (An
utterance would have to be automatically false if there is no
agreement under ideal speaking-conditions.) Another objection is that
people may agree that p is true and that q is true, while
p logically implies that q is false (particularly
when p and q are complicated utterances between which
the connection is hard to discern). And a third objection concerns
propositions or statements about factual, modal or normative conditions
in the past. These propositions or statements would solely be true if
agreement is reached on them in the present or future. Yet, the
proposition or statement, or what is stated, was true long
before agreement was reached. (The proposition even from the
moment the thing in question did, could or should happen.)
Neither the consensus theory nor the pragmatic theory of
truth are able to distinguish what is true from the belief in
what is true, or the
propositional reality of
beliefs from the lower-level propositional or nonpropositional
reality of what the beliefs are about. They lack any
propositional hierarchy
with the accompanying, real or imaginary, relationships between
the different levels of such a hierarchy. In this respect they
are diametrically opposed to the semantic theory of truth in
which truth must be defined at every level of a propositional
hierarchy of 'languages' separately. The proposition which is
true or not true is, then, expressed in what is called "object
language", whereas the definition of truth in that language is
given in a metalanguage on the next propositional level. (If the
definition is applied to sentences, one and the same sentence
may be true in one language and false or meaningless in
another.)
On the semantic theory of truth the definition of truth must
not only be formally correct in that it is defined at one
linguistic-propositional level at a time, it must also be
materially adequate. This means that it must hold that P is true
in language L if, and only if, p (in which P is the name in the
metalanguage of a sentence in the object-language, and p the
translation in the metalanguage of that sentence in the object-language).
The underlying idea is that, for example, water is transparent
is true 'iff' (that is, if and only if) water is
transparent. The reply to those who find this trivial is that
the only question at issue here is that of the definition of
truth, not some procedure or method for verifying utterances, that
is, not questions of epistemic justification. It has even been
pointed out that the semantic conception of truth can be
embraced 'without having to give up any epistemological attitude
one had already'.
The left-hand side of the schema P is true iff p has been
interpreted as referring to the language, that is, to propositional
reality. (Thus <<water is transparent> is true> refers to
a proposition, the proposition <water is transparent>.) The
right-hand side would, then, refer to the facts. (<Water is
transparent> refers to transparent water, or rather to the
relation water bears to the property of transparence.) The
schema P is true iff p therefore fits the correspondence
theory of truth and our relational interpretation very well. The
semantic theory of truth has been said to supply a suitably
objective account of truth as a guiding or 'regulative ideal'.
On this account truth is objective or absolute in that it is not
relative to people's knowledge or belief. (It is relative tho in
that truth is only defined for one linguistic level at a time.)
