7.4.2 |
DECISION-THEORETICAL CONSEQUENTIALISM |
Moral philosophers tend to treat consequentialism, even when not
decision-theoretical in any way, as a synonym
of teleology, but we shall reserve this latter term for the type
of (consequentialistic or nonconsequentialistic) normative doctrine
in which the value of
the performatory level has become a
goal, end or purpose (the meaning of telos).
It is not until we take intentions into consideration that
a value which is simply there 'being normatively superior' (as on the
purely performatory interpretation) becomes an end towards which effort
is, or should be, directed, or a thing towards which the agent
attempts, or should attempt to advance.
Teleology shall therefore not be used by us as a synonym of
consequentialism (which is only future-regarding and need not take
intentions or motives into account) but as a synonym of
decision-theoretical value-based ethics (which may be
present- and past-regarding as well).
It is intentional when the goal aimed at may be a means
to arrive at another goal which cannot be derived from the same
ultimate goal; and motivational when the goal aimed at is not
treated as a means, but as something
perfective in itself.
(We shall pay no attention here to the descriptive, pseudodescriptive
or supernaturalist doctrines which are also called "teleologies" and
which, we are made to believe, deal with 'purpose as an explanation of
natural phenomena', 'final causes' or 'evidences of design in nature'.)
To understand the meaning of decision-theoretical consequentialist
ethics as distinct from purely performatory consequentialist ethics,
we will direct our attention for a moment at
general utilitarianism which, like rule-utilitarianism, has been
developed to meet the objections against act-utilitarianism.
(General utilitarianism is general consequentialism with happiness
or 'utility' purportedly as sole value.) Standard cases to
which these objections apply are those of people crossing a lawn
where the lawn would be destroyed, if everyone crossed it; or
of people abstaining from voting where a democratic institution
would collapse, if no-one voted. The basic general-consequentialist
idea is clear and lucid: since the consequences of
everyone's crossing the lawn or of everyone's not voting are
bad, no-one ought to cross the lawn, and no-one ought to abstain
from voting. An important characteristic of these examples is
that they have thresholds (or for the sake of argument, some
transition zone to be narrowed down to a threshold). A certain
number of people can cross the lawn in a certain period without
damaging it, and in certain kinds of electoral system additional
votes do not have any effect on the election results anymore
once the number required for passage has been reached.
Antiutilitarians or -consequentialists have argued that one
'should not forget the importance of the pattern of other people's
behavior which is the crucial factor in the production of
threshold effects', and that one 'must not only distinguish
between acts within and outside general
practises, but relate
them to the threshold in question as well'. In other words: acts
which produce, or are causally related to the production of, the
threshold effect have to be distinguished from acts which do
not. The consequences of those acts are different. Thus, someone
voting before the threshold has been reached (supposedly)
produces a good effect, whereas someone voting after it has been
reached produces at best nothing. It has been demonstrated that general
utilitarianism and act-utilitarianism are 'extensionally equivalent',
that is, that it does not matter at all whether the
utility test is applied in these threshold-related cases to
simple acts or whether an appeal is made to general practises,
or for that matter, rules, grounded in utility.
This conclusion, however, is at once true and trivial. It is trivial
because solely concerned with purely performatory (or if preferred,
'extensional') consequentialism. In that case it is obvious that
performatory act-consequentialism and performatory general
consequentialism must be 'extensionally', that is, from the
performatory point of view, equivalent. Neither the performatory
consequentialists nor the performatory anticonsequentialists are
concerned with 'the practical problems of getting the required
information and correctly inferring judgments from the principles
on the basis of that information'. But the first ones to
blame are definitely those consequentialists who themselves do
not explicitly recognize the different layers of the ethical
profile.
The consequentialist's general practise is always related to
a function, utility or institution which is (believed to be)
desirable. For example, if someone ought to vote, it is because
the institution of democracy or the support of a particular good
cause is desirable, and if someone ought to abstain from
crossing the lawn, it is because a nice, green lawn is a good
thing, and a path visibly trodden
thru it a bad thing.
The function, utility or institution which is a common good recognized
by the members of a certain group or community must be
desirable or morally acceptable on the grounds of the same
value-theory which makes an individual act or omission into a
right act or omission. In performatory consequentialism this is
an uninspiring truth, but particularly when presuppositions come in (in
decision-theoretical value-based ethics) it will turn out to be
of no little import. No doubt, one of the most important
incentives to develop general utilitarianism was the need felt
to justify on utilitarian grounds the moral 'law' that one
should never lie, should never steal. (What will happen if
everyone lies or steals?) Yet, it is precisely this kind of
acts which cannot be described in purely performatory (or
'extensional') terms at all.
A person only lies if
'e asserts something that 'e
'imself
believes to be untrue with the intent to deceive. In the
event that 'e asserts something that is untrue (while not intending
to deceive), 'e is wrong in another sense, not connected to a
moral judgment. The same holds for stealing: stealing is not
just taking what is morally speaking someone else's property,
but taking it with the intent to keep it, while appropriation
has not been sanctioned. The general utilitarianism whose
anatomy has been exposed as 'extensionally equivalent' to
act-utilitarianism only employs brute facts to render moral
judgments, while acts such as lying and stealing it was designed to
bring into moral disrepute cannot even be described in that way.
Before saying that someone lies when 'e says something untrue,
or that 'e steals when 'e takes something away that belongs to
someone else, we need to know the person's intentions against
the background of
'er information and
presuppositions. In short: the general utilitarianism which includes
acts such as lying and stealing in its generalization tests simply could
never be performatory, that is, entirely independent of intentions,
to start with.
Ideally, there should be a complete correspondence between
the agent's information and presuppositions on the one hand and
the brute facts or modal conditions on the other, but this is
only contingently so. Nonetheless, such a correspondence seems
to be covertly assumed in performatory consequentialism. One
principle of consequences, for instance, is if the consequences
of A's doing x were undesirable, then A would not have the
right to do x.
On the supposition that it only uses were and would
because the antecedent is counterfactual this formulation is
ethically inadequate, as it does not give a cue to the moral
agent in a strict sense. Only indirectly, after assuming that
the agent's information and presuppositions wholly correspond
with (past, present and future) reality, does it provide a
suitable cue for 'im. To make such an assumption beforehand is
irrational from a decision-theoretical standpoint.
Given that the value-theory determines the moral agent's
intention or goal(s) and thereby what consequences are desirable
or undesirable (if future-regarding), A ought not to do x,
if according to 'er own information and reasonable presuppositions,
'er doing x would have undesirable consequences. Or, it
ought not to be A's intention to do something that would have
undesirable consequences according to 'er theory of value(s) and
'er (own) information and presuppositions. (This might logically
be formalized as
" ( Ef )
Oty " in which E
is an epistemic operator and Ot a temporally
relative operator used in dyadic systems of deontic logic. A definition of
f Oy
is under circumstance f
it is obligatory that y in terms of
f -ideal worlds, that is, worlds
which are at least as ideal as any world in which
f is the case may be.)
Granted that there is an objective reality, facts and modal
conditions themselves cannot contradict each other, only interpretations
of the factual and modal conditions can. Therefore in
performatory consequentialism the (true) conditions with regard
to a function, utility or institution and the (true) conditions
with regard to an act or abstention will always match.
In the moral decision-theory of intentional
(and motivational) consequentialism, however, there is the additional
requirement that the agent's presuppositions (and information) with
respect to the institution, for instance, themselves cohere with 'er
presuppositions (and information) with respect to the particular
act to be performed or abstained from.
It can then be proved that voting consequentialist with good intentions
does make a difference with voting on the purely performatory schema
(assuming that the value-theory adhered to does not have something like
participation in democracy by means of voting or otherwise as a separate,
perfective value). Essential to the proof is that the agent has
accepted the institution itself as being in agreement with 'er
value-theory, or at least that its (universal) advantages are greater
than the disadvantages of supporting it. The trouble of having
to vote is then, presumably, the only disadvantage. Furthermore,
the desirability of the institution implies that its advantages
would outweigh its disadvantages even if everyone voted,
altho only a certain
minimum number of favorable votes is required.
(If this is not the case, the institution is not worth it,
or the voting system itself has to be reformed.)
We should not forget either that the person we are focusing on is not
someone acting out of plain self-interest, but is a moral
agent who rationally bases 'er decisions upon a certain nonegoistic
value-theory. In particular 'e does not yield to statistically
or otherwise unwarranted assumptions which would merely be to
'er own benefit.
A fundamental question which remains in decision-theoretical
consequentialism (and in all nonperformatory ethical doctrines) is whether
the agent is allowed to assume that other people will act from the same
value-theory (or theory of rights, or theory of
duties). This is an issue no calculus can solve, for it precedes
every value- and decision-theoretical exercise. In our discussion
of relevancy an affirmative answer to this question
corresponded to the universal version of the principle of
discriminational relevance, and a negative answer to the personal
version of that principle. (See section
5.4.4.)
Consequentialist ethics cannot bypass the realm of the
intensional, of propositional attitudes. It must take account of
the difference in kind of normative judgments on the level of
the purely performatory (nonintensional, nonpropositional or
lower-level propositional), the level of the intentional, and
the level of motives and traits of character.
If it does not make a difference --as has been
demonstrated-- whether acts are judged from the
point of view of simple or general utility, then the normative system in
which this equivalence holds is inadequate or incomplete in the first
place, even from the standpoint of monistic consequentialism, because it
sticks to judgments on the first, performatory level only.
Every adequate normative system needs at least implicitly a
moral decision-theory. By definition such a theory does not
bluntly calculate on actual performances or the actual effects
of actions. As in intentional consequentialism it elevates both
the agent and the observer watching 'im to a higher level
-- 'higher' in that it encompasses more of what is required. But
even after having recognized two or three levels of normative
judgment, a monistic consequentialist will still have to add one
or more values, or change 'er value, if generalization or rules
cannot accommodate under the heading of utility (or whatever
else it might be) everything that is believed to be moral.
Little imagination is, then, needed to introduce more criterions
of evaluation at the performatory level. But after the adoption
of a pluralistic value-theory the difficult task remains to
develop a moral decision-theory which gives directions to the
agent what to do or not to do, also when two or more values
conflict which cannot in any way be reduced to one another or to
one fundamental value. The duty to solve this problem seems to
be the main result of ruling out utilitarianism and performatory
consequentialism.
