Dispositions and Laws of Nature
Laws and Lawlessness
Stephen Mumford
1. INTRODUCTION
Problem
Are there laws of nature of the sort philosophers have discussed?
Is there a job for them to do? Do laws deliver some feature of the world
that would be otherwise lacking? Did God have to create laws once he had
put everything else in place?
Three positions
Primitivism: Laws are a distinct metaphysical category of thing, irreducible
and ineliminable.
Reductionism: There are laws but they can be accounted for entirely by
other things that are not laws, or reduced to those other things.
Eliminativism: Laws are neither reducible to other categories of things
nor a distinct category in their own right.
I explore the final position: a non-Humean version of Lawlessness that
hasn’t yet been given fair consideration.
History
Philosophy and science managed without laws before Descartes (1644), though
see Ruby (1986).
Laws have been denied before:
— by Hume and Humeans, e.g. Lewis (1973, 1986). This denial comes
from a concern to deny necessary connections in nature. There is modal
truth (Lewis) but no modal properties.
— by van Fraassen (1989). His denial comes out of his epistemology.
Laws are important features of a model. Science looks for models that
are empirically adequate (constructive empiricism), it does not aim for
truth. Symmetry principles are more basic than laws.
A third Lawless view, that interests me, is that there are enough features
in the world already to supply everything we wanted (created) laws for.
This position would be realist about modal properties (dispositions),
for instance. Those features of the world we think we need laws to supply,
can be delivered by other elements that are (a) needed anyway, but (b)
less objectionable and less mysterious.
Is this an attack on the practice and practitioners of science? No. As
van Fraassen, Giere (1995) and others have noted, that science is all
about laws has been more the interpretation of philosophers than scientists.
Laws are treated with a degree of flippancy in the science books and are
still called laws even when false (a cheap shot but might be indicative
of the attitude of science to laws).
Let us look first at the main existing Lawless view (Hume, Lewis), second
at realism about laws (Armstrong, et. al.) and third at a new Lawless
view, Realist Lawlessness (RL).
2. HUMEAN LAWLESSNESS
Metaphysical incommensurability
Standard criticism of the Regularity Theory of Laws (RTL) is usually along
lines that RTL says L is/isn’t a law when clearly L isn’t/is.
But this exploits a space between a concept of law and the RTL that its
defenders maintain defiantly is not there. No sense can be made of laws
that are stronger than theirs, they say.
A fundamental difference in metaphysics causes the disagreement.
The Lawless metaphysics of Humeanism
Some speak of the Ramsey-Lewis theory as if it were a theory of laws.
My reading is that it, like RTL, is a theory of there being no laws.
A reminder of some features of HL, the Hume-Lewis-world:
— The history of the world is a history of events.
— The fundamental, subvenient events (facts) are ‘local’:
point-sized qualities instantiated at points.
— Science can tell us what these qualities are.
— Everything (else) supervenes on them (we hope).
— There are no intrinsic modal features: all events are modally
unconnected.
The ‘laws’ of HL hardly deserve the name. They play no role
in determining or regulating what happens. Instead, they are just the
axioms of the best systematisation of all that happens. There are not
the nomological features in the world that realists claim.
The ‘regularities’ of RTL also barely deserve the name. HL
has regularities only like those created by a random number generator.
In a very long string of numbers, the sequence 1, 2, 3, 4, 5 might occur
many times. But whenever we have 1, 2, 3, 4 there would be nothing at
all, in any circumstances, that might make 5 next.
With a genuine regularity, of the kind nomological realists think exist,
there is something that would make it 5, on the occasions where it is
5, and this is the metaphysical grounding of the rational expectation
of 5.
3. NOMOLOGICAL REALISM
The Nomological Argument
Nomological realists think that there are laws. Apart from dislike of
RTL, is there an argument for NR? The Nomological Argument (NA) seems
to be:
A. There is a set S of features in the world
B. There is S because there are laws of nature.
No one has explicitly advanced NA but sometimes it seems implicit and,
in any case, it has a useful dialectical role.
The set S might be something like:
{Regularity, Universality, Objectivity, Immanence, Centrality, Explanation
and Prediction, Necessity}
Laws are then supposed to have, or deliver, these features. Can a single
thing deliver them all?
The Dretske, Tooley, Armstrong (DTA) theory says so. But there is an attempt
to marry necessity and contingency. A law is of the logical form N(F,G),
where N is a relation of natural necessitation between universals. It
is contingent which universals are so related.
Ellis has developed a necessitarian essentialist position. Laws are about
essential properties of natural kinds and ‘causal laws just describe
the natural kinds of process involved in [a causal power’s] display’
(2001, p. 4). The ‘price’ is that laws become metaphysically
necessary so could not have been otherwise than they are.
4. REALIST LAWLESSNESS
Weak and Strong RL
Let us return to the Nomological Argument. We can note that:
— NR accepts A and argues that B follows from A.
— HL is a denial of A.
RL has weak and strong variants:
— weak RL: A is true but B does not follow from A.
— strong RL: A is true but B is false.
The position is realist in virtue of accepting A, but nomological anti-realist
in rejecting B.
This might be what Armstrong called ‘one truly eccentric view …
that, although there are regularities in the world, there are no laws
of nature’ (1983, p. 5).
Evidence for NA?
Some have jumped straight from the predictable nature of the world to
laws. (Perhaps there are intermediate steps in this argument: the world
is predictable because it is regular and it is regular because there are
laws.) There seems to be little consideration of whether something other
than laws might ground regularities (and all S).
Armstrong has arguments for why S must be accounted for by relations between
universals rather than just by a universal quantification over particulars.
But laws are not primitive; they are not a new category of thing in addition
to universals.
Ellis’s new essentialism and Lowe’s (2003) fourfold ontology
might also have the resources to deal with Armstrong’s concern.
Reductionism or Eliminativism?
Those who say there are laws are usually in favour of reductionism, e.g.
in DTA laws are made from properties and the relations between them. There
is something special about N(F,G) but it is not that some extra element
has been added to the F, the G, and the N between them.
Similarly for Ellis. His metaphysic for laws requires natural kinds and
their (essential) properties. There is no extra nomological element.
The eliminativist claim of RL is that there is not a one-to-one mapping
of existing categories, or some combination thereof, on to laws as defined
by S. This ought be considered a live option, rather than ‘truly
eccentric’.
A Realist but Lawless world?
The RL world would be nothing like HL.
There would be some connections between the particulars of the world by
their essential (dispositional) properties. Our world is dyamic (the dynamic
cogito: ‘something happens’). We can gain this dynamism by
having dynamic properties rather than external relations between essentially
static properties.
These connections could explain any regularities and their concomitant
epistemic features (cf. Lombardo, 2002).
They could be objective and immanent and yet provide the natural necessitation
that philosophers have sought.
Molnar has said of causation that: ‘there is no purely descriptive
analysis of the concept. All theories are revisionary to some extent,
that is, they do not merely report/record “the concept we have”’
(2003, p. 190). The same may be true of laws. But do we really need to
bother attempting to revise the ‘concept we have’ of laws
if
a) we have other metaphysical resources that do any decent work we wanted
laws for, and
b) laws came out of a misleading picture of the world: inert, entirely
discrete units having to be governed from without.
Just as we eliminated the concept of a witch, rather than reducing it
to contemporary concepts, might we not do the same with laws?
REFERENCES
Armstrong, D. (1983) What is a Law of Nature?, CUP.
Descartes, R. (1644) Principia Philosophiae, trans. ‘Principles
of Philosophy’ in Cottingham, Stoothoff and Murdoch I, CUP (1985).
Ellis, B. (2001) Scientific Essentialism, CUP.
Giere, R. (1995) ‘The Sceptical Perspective: Science without Laws
of Nature’, in F. Weinert (ed) Laws of Nature, de Gruyter.
Lewis, D. (1973) Counterfactuals, Blackwell.
——(1986) On The Plurality of Worlds, Blackwell.
Lombardo, E. S. G. (2002) ‘Analogical versus Discrete Theories of
Possibility’, Australasian Journal of Philosophy 80.
Lowe, E. J. (2003) ‘Recent Advances in Metaphysics’, Metaphysica.
Molnar, G. (2003) Powers, Mumford (ed.), OUP.
Ruby, J. (1986) ‘The Origins of Scientific “Law”’,
Journal of the History of Ideas 47.
van Fraassen, B. (1989) Laws and Symmetry, OUP.