Fitch 1963
Fitch 1963
Fitch 1963
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
Association for Symbolic Logic is collaborating with JSTOR to digitize, preserve and extend access to The
Journal of Symbolic Logic.
http://www.jstor.org
135
4 A. W. Burks, The logic of causal propositious, Mind, n.s. vol. 60 (1951), pp. 363-382.
Proof. Suppose that p is true but not known by the agent. Then, since
knowing is a truth class closed with respect to conjunction elimination,
we conclude from Theorem 2 that there is some true proposition which
cannot be known by the agent.
THEOREM 5. If there is some true proposition which nobody knows (or
has known or will know) to be true, then there is a true proposition which
nobody can know to be true.
Proof. Similar to proof of Theorem 4.
THEOREM 6. If there is some true proposition about proving that nobody
has ever proved or ever will prove, then there is some true proposition about
proving that nobody can prove.
Proof. Similar to the proof of Theorem 4, using the fact that if p is a
proposition about proving, so is [p & '(ap)].
This same sort of argument also applies to the class of logically necessary
propositions, since this is a truth class closed with respect to conjunction
elimination. Thus by Theorem 1 we have the result that every proposition
of the form [P & - FIP] is necessarily not logically necessary, and hence
necessarily possibly false, where 'El' denotes logical necessity. In other
words, the proposition D',.' FII[P& N oJp] is true for every proposition p.6
In particular, if p is a true proposition which is not necessarily true, then
[p & Nop] is a true proposition which is necessarily possibly false.
I now wish to describe a relation of causation, or more accurately, partial
causation, which will be used in giving a definition of doing in terms of
striving and a definition of knowing in terms of believing, as well as some
other definitions.
I will assume that partial causation, expressed by 'C', satisfies the
following axiom schemata Cl-C4:
Cl. [[p C q] & [q C r]] -3 [p C r]. (transitivity)
C2. [p & [p C q]] -3 q. (detachment)
C3. [p & [[p & q] C r]] -3 [q C r]. (strengthening)
C4. [[p C q] & [p C r]] [p C [q & r]]. (distribution)
Here 'p q' is defined as '[p -3 q] & [q -3 p]'.
I will also employ an identity relation among propositions and will
employ the following axiom schemata I1-I9 for this identity relation:7
6 This result in slightly different form is to be found in the two papers by Anderson
cited above. He uses it in constructing a model of deontic logic in alethic modal logic
and attributes it to W. T. Parry, Modalities in the survey system of strict implication,
this JOURNAL, vol. 4 (1939), pp. 137-154, Theorem 22.8.
7 It is interesting to observe that 12-I9 may be used to serve as postulates for an
algebra like Boolean algebra but somewhat weaker, provided that the identity symbol
is regarded as a symbol for equality in such an algebra and that (in place of I 1) there
are added postulates to the effect that equality is symmetrical and transitive, and that
the negates, conjuncts, and disjuncts of equal elements of the algebra are equal.
of this striving is that p takes place. Using II, I6, C4, and properties of
existence quantification, it is easy to show that this definition gives the
result that doing is closed with respect to conjunction elimination.
A definition of knowing in terms of believing is now given:
D2. (knows p) =3q[p & q & [[p & q] C (believes [p & q])]].
This means that an agent will be said to know p provided that p and
some (possibly other) situation q are both true, and provided that the fact
that they are both true causes the agent to believe the fact that they are
both true. Thus the known fact p must be causally efficacious (as part of
the conjunction [p & q]) in bringing about the agent's belief that [p & q]
is the case, and hence that p itself is the case, since belief is assumed closed
with respect to conjunction elimination. It is easy to show that knowing,
as thus defined, is a truth class closed with respect to conjunction elimi-
nation.
Ability to do can be defined in the following way:
D3. (can do p) 3q[(strives for [p & q]) C p].
This definition can be shown to give the result that ability to do is closed
with respect to conjunction elimination.
Obligation to do can be defined in terms of doing and the concept of
obligation as expressed by the operator '0' of deontic logic, as follows:
D4. (should do p) 0(does p).
Obligation to do, as thus defined, can be shown to be closed with respect
to conjunction elimination and also with respect to conjunction introduction.
I now wish to propose a definition of desire, as follows:
D5. (desires p) 3q[(believes (can do [p & q])) C (strives for [p & q])].
This means that an agent desires a situation p if his belief that he can
achieve the conjunction of p with some (possibly other) situation causes
him to strive for that conjunction of situations. Desire as thus defined can
be shown to be closed with respect to conjunction elimination.
A concept of value, which I now wish to consider, can be defined in the
following way:
D6. (value p) 3q3r[q & [(knows q) C (strives for [p & r])]].
This means that a situation p is a value for an agent if (and only if)
there is an actual situation q and situation r such that if the agent knows q
then he will strive for the conjunction of p and r. In knowing q the agent
may be supposed to have all the knowable relevant information concerned
with the effect of his striving for the conjunction of p and r, and if this
knowledge causes him to strive for this conjunction, it must be because
this conjunction, and in particular p itself, is of value to him. To see why q
may be supposed to contain all the knowable relevant information for the
purpose at hand, let us suppose, on the contrary, that q does not contain
all such relevant information. Then there might be some additional in-
formation s such that knowledge of the conjunction of q and s would cause
the agent not to strive for any conjunction of the form [P & t]. But in the
hypothetical case that the agent knew [q & s], he would also know q because
of the fact that knowing is closed with respect to conjunction elimination,
and this knowledge of q, by assumption, would cause him to strive for
[p & r]. Thus he would be caused to strive for [P & r] and also caused not
to strive for [P & r], and the assumption that he could know such a propo-
sition as [q & s] leads to an absurdity. Hence q may be regardedas containing
all the knowable relevant information. It can be shown easily that value
as thus defined is closed with respect to conjunction elimination.
The objection might be raised against the above definition of value that
the agent must be assumed to be rational, since otherwise he might have
all the relevant knowledgeto enable him to make a choice in his own interest,
and yet, being irrational, he would be caused by this knowledge to make
some other choice and to strive for some outcome that would not be of value
to him. One way, and perhaps the only way, to attempt to meet this ob-
jection is to maintain that all irrationality is due to lack of sufficient
knowledge, so that the having of sufficient relevant knowledge already
rules out any relevant amount of irrationality. According to this view, any
sort of insanity would be curable simply by giving the patient sufficient
knowledge of himself and of the world around him. This view would not
deny that in practice there might be insuperable obstacles that prevent
the communication of this knowledge to the patient, but the existence of
such obstacles would not prove that irrationality was not essentially a
lack of knowledge.
This definition of value of course does not guarantee that there are any
values in this sense, though it seems to me not unreasonableto assume that
there may be values in this sense.
A more difficult problem is the problem of the comparison of values,
that is, the problem of greater and less among values. This problem will
not be dealt with here.
YALE UNIVERSITY