Natural Laws and the Problem of Provisos
Author(s): Marc Lange
Source: Erkenntnis (1975-), Vol. 38, No. 2, Special Issue on Philosophy of Science (Mar., 1993),
pp. 233-248
Published by: Springer
Stable URL: http://www.jstor.org/stable/20012470 .
Accessed: 12/11/2013 21:02
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].
.
Springer is collaborating with JSTOR to digitize, preserve and extend access to Erkenntnis (1975-).
http://www.jstor.org
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
MARC
NATURAL
LAWS AND
LANGE
THE PROBLEM
that the existence
and Giere
contend
Hempel
account
law. However,
for any regularity
of physical
a mistaken
in which
of the way
statements
conception
ABSTRACT.
difficulties
upon
OF PROVISOS
of
grave
poses
and Giere
rely
their content.
By
provisos
Hempel
acquire
and Giere
I remove
the problem Hempel
this mistake,
correcting
identify but reveal a
account
that provisos
for any account
different
indeed,
pose for a regularity
problem
to which
of physical
law according
the state of affairs described
by a law-statement
a normative
These
considerations
suggest
regularity.
analysis
are not distinguished
law-statements
this view,
from accidental
a law-statement
because
regularities
by the kind of Humean
they describe
generalizations
a law-statement
need not describe
Rather,
any Humean
says that in certain
regularity.
a Humean
presupposes
of law-statements.
On
contexts,
one
justification,
to regard
the assertion
of a given
ought
as a proper way to justify a claim of a certain
type of claim,
other kind.
if made
with
I
to the regularity account of physical law - versions of which
According
have been advocated by Ayer
(1963), Braithwaite
(1953, ch. 9), Good?
man (1983, pp. 17-27), Hempel
(1965a, pp. 264ff.), Lewis (1973, pp.
72-77;
1986), Mackie
(1962, pp. 71-73), Nagel
(1961, pp. 58ff.), and
- laws of nature are
ch.
Reichenbach
others
8), among
(1947,
regularit?
ies among events or states of affairs and a law-statement,
the linguistic
of a law, is a description
of a regularity
that is a law.
expression
The
familiar
faced by this account
is to distinguish
those
challenge
are
of
that
law-statements
from
those
that
descriptions
regularities
are accidental
a
more
to
I
wish
consider
fundamental
generalizations.
to describe no regularity at all,
that many a claim we believe
problem:
or
we
nevertheless
accidental,
nomological
accept as a law-statement.
This problem
arises from what Hempel
of
(1988) calls "the problem
provisos."
statement
of the law of thermal expansion:
of a metal bar of length L0 changes by
= k
AT, the bar's length changes by AL
L0 AT, where k is a constant
of that metal."
This statement
states a relation between
characteristic
not
AL
and
that
does
it
be
for instance,
L0, AT,
obtain;
violated,
may
if someone
is hammering
the rod inward at one end. Since this state
Consider
"Whenever
Erkenntnis
?
the familiar
the temperature
38: 233-248,
Academic
1993 Kluwer
1993.
Publishers.
Printed
in the Netherlands.
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
234
ment
MARC
does
regularity
not describe
account.
LANGE
a regularity,
it is not a law-statement,
on the
statement
hold that a complete
of the law
(1988) would
Hempel
includes the condition ". . .if the end of the rod is not being hammered
in." On this view, whereas
the familiar "law-statement"
takes the form
D
one
must
to
and
the
add
Gx),
only premise
(x) (Fx
yield the con?
the genuine
clusion Ga
is Fa,
law-statement
takes the form (x)
one
as
as Fa to infer Ga. It
to
&PxD
and
needs
add
Pa
well
(Fx
Gx),
account of laws that leads Hempel
to characterize
is the regularity
. . if Px as "essential"
to
such conditions
law-statements
(x).
(and the
as
to
Pa
"essential"
without
these
inferences);
corresponding
premises
on
the
claims
would
be
false
and
the
so,
conditions,
regularity account,
would not be law-statements.
Following Hempel
(1988, p. 23), I'll refer
to those conditions
this
that, by
(and premises)
reasoning, are necessary
to
to law-statements
inferences
from
but are
(and
law-statements),
as
"generally unstated,"
"provisos."
scientific practice. By Hempel's
Snell's
Provisos pervade
reasoning,
- when a beam of
one
to
from
medium
law of refraction
light passes
=
r
i
where
is
the
of
incidence
of
the
//sin
sin
constant,
another,
angle
r is the angle of refraction
in that
beam upon the second medium,
two
constant
of
the
of
and
is
characteristic
media
the
medium,
types
and pressures of the media as well
must require particular temperatures
or electrical potential
across
as the absence of any magnetic
difference
and non-dou
uniform optical density and transparency
the boundary,
in the two media,
and a monochromatic
beam; these
ble-refractivity
- the
are provisos. Likewise,
the law of freely falling bodies
conditions
must specify when
distance a body falls to earth in time t is (l/2)gt2
as "free"; while
true
the law can remain approximately
fall qualifies
its predictions may be
the height at which g is measured,
or other
air
when
resistance,
forces,
electromagnetic
drastically wrong
collisions affect the falling body.
to state very many
it becomes
On Hempel's
impossible
proposal,
as
it
Giere
law-statements
since,
puts
(1988, p. 40), "the
genuine
law
is
in
number of provisos
any
indefinitely
large." To state
implicit
for instance, one would need to specify
the law of thermal expansion,
the bar inward at one end, but also
not only that no one is hammering
that the bar is not encased on four of its six sides in a rigid material
away
from
that will not yield
all cases in which
as the bar is heated, and so on. For that matter,
the bar is hammered
upon constitute
exceptions
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
not
to
NATURAL
LAWS AND THE PROBLEM OF PROVISOS
235
AL = k L0 AT; the bar may be hammered
upon so softly and be on
surface that the hammering
such a frictionless
translation
produces
of the bar. One is driven to say that the only
rather than compression
is to employ some such condition
law-statement
way to utter a complete
as "...
in the absence of other relevant factors." But Hempel
deems
in a law-statement:
On the regularity
inadmissible
such an expression
states a particular relation (which must obtain
account, a law-statement
and be a law), but a claim (x) (Fx & there are no other relevant
it
factors D Gx) does not assert any determinate
relation at all because
fails to specify which other factors count as relevant,
which
i.e.,
specific
to infer Ga. Such
premises are to be added to Fa and the law-statement
a claim is no better than "The relation AL = k L0 AT holds when it
than is the familiar
holds," and so is further from being a law-statement
which at least ascribes a
expansion,
to
various
(albeit false)
quantities.
particular
sees the existence of provisos as posing a dilemma:
In short, Hempel
For many a claim that we commonly
either
accept as a law-statement,
that claim states a relation that does not obtain, and so is false, or is
shorthand for some claim that states no relation at all, and so is empty.
account must admit that many
In either case, the regularity
claims
as
are
law-statements
neither
law-state?
commonly
accepted
complete
nor even colloquial
ments
themselves
stand-ins for complete
law-state?
ments.
If we continue to regard those familiar claims as law-statements,
then we violate the regularity account. This is the problem of provisos.
(Prior to Hempel
(1988), versions of this problem were discussed
by
and Lehrer
Canfield
(1961) and Coffa
(1968), as well as by Hempel
himself (1965b, p. 167), and the issue, in general terms, was anticipated
similar to the problem of provisos have
by Scriven (1961). Difficulties
also been noted in ethics, e.g., with regard to Ross's
(1930) definition
of a "prima facie duty".)
One may be tempted to reject this problem by insisting that genuine
on this view, that many
law-statements
lack provisos;
familiar "law
are actually neither
nor abbreviations
statements"
law-statements
for
we
to
law-statements
show
that
have
few
discovered
very
only goes
statement
of the
law of
thermal
relation
To yield to this temptation
genuine laws (or nomological
explanations).
I
be
account
An
of
laws must accommodate
would,
think,
unjustified.
the fact that scientists show no reluctance
to use these familiar claims
in the manner
distinctive
of law-statements,
and
e.g., in explanations
in support of counterfactuals.
This fact would be difficult to explain if
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
236
MARC
LANGE
we held
that scientists do not consider
them to express laws. To insist
are genuine
that only claims without
nevertheless
law-state?
provisos
a claim must or must not be captured by
ments
is to hold that whether
an analysis of what it is to believe a claim to state a physical
law is not
as
or
to
whether
treat
determined
scientists
claim
able
unable
that
by
those
functions
that
from
accidental
law-statements
distinguish
perform
But, then, on what basis is the adequacy of an account
generalizations.
that
Scientific practice
of law to be evaluated?
is the only phenomenon
exists for an account of law to save; if an account is tested against not
the
actual science but science as idealized to conform to that account,
test is circular. If accounts of law are not free to disregard the fact that
scientists treat "All gold cubes are smaller than one cubic mile" as an
are smaller
and "All cubes of Uranium-235
generalization
a law, even though this fact is
as expressing
cubic mile"
I see no reason why an
for many
troublesome
accounts,
proposed
account of law should be permitted
to ignore the fact that scientists
treat as law-statements
many claims involving provisos.
some law-statements
is to argue that although
Another
temptation
law-state?
involve provisos,
these derive from other, more fundamental
accidental
than one
to describe regularities.
For
need no provisos
ments, which themselves
law
in
the
from
the
of
classical
bodies
follows
instance,
physics
falling
with
information
laws of motion
and gravitation
fundamental
along
about the earth; the proviso ". . .so long as the body is falling freely"
the gravitational-force
law
restricts the law to those cases in which
accounts
for
all
of
to
the
forces
this
the
earth
information
about
applied
should also be resisted.
this temptation
acting on the body. However,
laws: What regularity
Itmerely pushes the burden onto the fundamental
a regularity
law? If it described
is described by the gravitational-force
of two bodies and the total force
the masses
and separation
between
it would be false unless it included the
that each exerts on the other,
exert no other forces upon each
". . .so long as the bodies
proviso
law
other." But to regard this proviso as part of this "fundamental"
two
to a case in which
not only conflicts with the law's applicability
bodies
interact, but also raises the familiar prob?
charged and massive
in which other forces
lem: The proviso fails to specify the circumstances
law from setting
are present,
the gravitational-force
and so prevents
is supposed
If each of the other force-laws
relation.
forth a particular
force is present and thereby
to specify when a given non-gravitational
help
to determine
the relation
asserted
by the gravitational-force
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
law,
NATURAL
then each
of
these
LAWS AND THE PROBLEM OF PROVISOS
other
would have to apply to a body
and so could not describe a regularity
on the body. Alternatively,
for the
a regularity between
the masses
and
force-laws
by many types of forces
the total force exerted
involving
law to describe
gravitational-force
two
of
and the
bodies
separation
that each exerts on the other, a
affected
237
than total) force
force would
component
gravitational
to certain regularities.
have to be a real entity that conforms
That
a
are
is
controversial
forces
real
contention
Cart?
component
(see, e.g.,
wright 1983), and in any event, it seems to me that the nomic status of
law does not depend on it; after all, we use the
the gravitational-force
even
law in the manner
distinctive
of law-statements
Coriolis-force
gravitational
(rather
force to figure in a regularity
there to be no Coriolis
though we believe
that the law-statement
describe.
might
one would be justified in setting the problem
of provisos
Perhaps
aside if one had some account of physical
law that, except for this
problem, were entirely successful. But since I know of no such account,
I think it worth investigating whether
toward one can
greater progress
than by disregarding
them. In this
by reflecting on provisos
a
to
I'll
the
of
offer
that
response
paper,
problem
provisos
ultimately
an
account of physical
undermines
the regularity
law and suggests
of
law-statements
to
normative
which
alternative,
conception
according
to justify
they specify the claims we ought to use, in various contexts,
certain other claims. I'll argue that some claims are properly adopted as
law-statements
they are believed not to describe regularities,
although
because one who believes
that "It is F" ought to be used to justify "It
is G" need not believe
that some regularity,
such as that all Fs are G,
be made
obtains.
In Section II, I'll explain the problem of provisos more fully, and in
must
Section III, I'll attack Hempel's
view of what a "law-statement"
on which
be to qualify as complete,
the problem depends. The more
liberal criterion of completeness
that I'll defend permits us to avoid
us
to regard familiar "law-statements"
dilemma
Hempel's
by enabling
as law-statements.
I'll argue in section IV, however,
that this response
to the problem
the rejection of any regularity
of provisos
requires
account of physical
law because many a claim commonly
accepted as a
law-statement
describes no regularity; a normative
of law
conception
statements
then suggests
itself. In Section V, I'll maintain
that this
of physical
strategy can be used to argue against many other conceptions
law besides
the regularity account,
such as those of Armstrong
(1983)
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
238
and Kneale
of provisos
MARC
LANGE
I'll argue
(1949). Finally,
is superior to that offered
that my response
by Giere
(1988).
to the problem
II
as posing
does not explicitly present
the problem of provisos
Hempel
are either false or empty. Yet this
the dilemma
that "law-statements"
dilemma
It must be because he
certainly stands behind his discussion.
a "law-statement"
believes
without
would
be false, and so
provisos
would not be a law-statement,
that Hempel
defines provisos as "essen?
tial." He goes on to point out that if the complete
law-statement
includes the proviso (x). . . if Px, then among the premises of an infer?
ence from the law-statement
to testable predictions must be Pa, as well
as other auxiliary hypotheses.
This might at first appear to reduce at
to a special case of the Duhem
least part of the problem of provisos
a
not
law-statement
is
falsifiable
Quine
(at least, not in a
problem:
to make a prediction
from it that can
sense) because
straightforward
so one can preserve
be tested, one must use auxiliary hypotheses,
the
an
the
if
law-statement,
fails, by rejecting
prediction
auxiliary hypo?
are merely
thesis. If provisos
additional
distin?
auxiliary hypotheses,
in
guished from others only by the fact that they generally go unstated
scientific practice,
then (it might appear) they do not represent a novel
kind of threat to the falsifiability of individual hypotheses.
that the existence
of
however,
(1988, pp. 25f.) emphasizes,
Hempel
some
to
obstacle
that
presents
provisos
beyond
falsifiability
posed by
the Duhem-Quine
problem. The Duhem-Quine
problem assumes that
are jointly sufficient
to
the law-statements
and auxiliary hypotheses
if there are
entail the testable prediction.
But, Hempel
maintains,
then this assumption
often
among the auxiliary hypotheses,
provisos
one can state neither
fails because,
for many a familiar "law-statement,"
. . if Px needed
set of proviso
the complete
conditions
to make
(x).
true nor the complete
set of auxiliary hypotheses
that "law-statement"
to infer the testable prediction
Pa needed
from the law-statement.
to
in
the Duhem-Quine
that
Therefore,
says
comparison
Hempel
prob?
from provisos
leads rather to the stronger con?
lem, "[t]he argument
or theoretical
clusion that even a comprehensive
system of hypotheses
not
will
entail
because
the requisite
any [testable predictions]
principles
- that is to
remains
deduction
is subject to provisos"
say, always
subject
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
NATURAL
LAWS AND THE PROBLEM OF PROVISOS
239
no matter how many auxiliary premises one adds to try to
to provisos,
to the law.
exclude all factors disturbing
it is only because he considers complete
im?
law-statements
Hence,
since "the number of provisos implicit in any law is indefinitely
possible,
sees provisos as presenting
a
large" (Giere 1988, p. 40), that Hempel
difficulty that is distinct from the Duhem-Quine
problem. But suppose
believed
that a condition
such as "...
in the absence of other
Hempel
and likewise
that
relevant factors" could appear in a law-statement,
"There are no other relevant factors" could function as an auxiliary
Then he would have to admit that any law can be completely
hypothesis.
stated by a claim that includes only a finite number
of conditions
if there are no disturbing
if Px, one of which might be "...
(x)...
a
to
testable prediction
includes only
factors," and that any inference
a finite number of auxiliary hypotheses
Pa, one of which might be
"There are no disturbing
factors." The existence
of provisos would
then add nothing new to the Duhem-Quine
therefore
problem. Hempel
must regard the expression
". . .in the absence of other relevant fac?
tors" as inappropriate
for a law-statement.
Though he does not explain
account of laws, implicit throughout
this is, the regularity
his
why
= k
discussion,
L0 AT obtains
suggests an answer: The sentence "AL
in the absence of factors that disturb it" states no definite relation and
so cannot be a law-statement.
This worry is evident in Giere's
remark
is to formulate
(1988, p. 40): "The problem
without
trivial."
rendering the law completely
the needed
restrictions
Ill
. . if Px
a proviso
definition,
(x).
usually is omitted from
of the law, and the corresponding
premise Pa usually is
not mentioned
in inferences
involving that law. Hempel
regards these
as enthymemes
inferences
as in?
and these familiar "law-statements"
are
we
But
to
able
make
do with
law
complete.
why
incomplete
statements? And
is Hempel
correct in considering
them incomplete?
I'll now argue that Hempel's
criterion of completeness
is motivated
by
an incorrect view of what is necessary
in order for a sentence
to state
a determinate
relation.
I'll argue that familiar law-statements,
which
By Hempel's
a statement
clauses such as "in the absence of other relevant factors," are
as they stand.
complete
That proviso premises are distinguished
by their absence from ordin
include
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
240
MARC
LANGE
is not an incidental feature of Hempel's
definition. As
ary conversation
I've explained,
it is bound up with the fact that the number of provisos
it impossible
to offer them all as
is "indefinitely
large," which makes
But
do
scientists
find
it
to mention
any of
unnecessary
why
premises.
in order to put an end to demands
the proviso premises
for the justifi?
cation of their conclusions? Why
is it that although
it is known
that
on the bar, AL = k L0 AT need not
someone
when
is hammering
obtain, in actual practice a claim concerning
k, L0 and AT is recognized
on the bar," to put ?n end
as sufficient, without "No one is hammering
to demands
for the justification of a claim concerning AL?
on the
The answer is that in practice, when no one is hammering
of claims concerning
bar, nearly all of those who demand justifications
to consider a different
AL already believe
that this is so. (Likewise,
it is widely understood
in many fields of
by workers
proviso example,
no
cases
will
involve
that
velocities
that of light.)
physics
approaching
Of course, to someone who presents a claim concerning
k, L0 and AT
to justify a claim concerning AL, one could object, "You have not told
me that no one is hammering
on the bar, and this you must do because
if someone
is, then (you will agree) your conclusion may well be false
even though your premise
is true." But apparently,
that no one is
on
case
bar
in
the
actual
in which
it is
would,
nearly any
hammering
in
who
advance
those
demand
the
be
believed
true,
by
justifi?
might
cation of some claim concerning AL. In nearly all cases, then, someone
a justification
AL should regard
who demands
for a claim concerning
a
as
a claim concerning
and
AT
sufficient
k, L0
response.
not only explains why
to this kind of shared background
Attention
scientists needn't in practice give any of the "indefinitely
large" number
but also reveals
in order to justify their conclusions,
of proviso premises
is too high. Hempel
standard of completeness
apparently
why Hempel's
considers a "law-statement"
complete only if it suffices, in the absence
to inform one of what
it takes for
of any background
understanding,
re?
law. This ideal of completeness
nature to obey the corresponding
con?
include
all
of
the
law-statement
that
the
proviso
complete
quires
. . . if Px and, more
that none of these con?
ditions
(jc)
importantly,
this condition
ditions be "in the absence of disturbing
factors," because
"All Fs
A
to
generalization
understanding.
background
plainly appeals
are G, except when disturbing
factors are present" does not indicate,
to one who doesn't know already what consti?
in a manner
intelligible
tutes a disturbing
factor, some determinate
regularity to which nature
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
NATURAL
LAWS AND THE PROBLEM OF PROVISOS
241
In exactly the same way, someone who is told to follow a
conforms.
. . .when it is
to apply this
to the regularity
rule "Conform
appropriate
take to follow this rule only if she
rule" can understand what it would
already knows when this rule is appropriately
applied.
in the absence of implicit
But to require that a rule be intelligible
crite?
of how to apply it is not a reasonable
understanding
background
no rule can satisfy it. As Wittgenstein
of completeness
because
can
one
always conceive of alternative
(1958) suggests,
interpretations
to
the same actual past actions as conforming
of a rule that recognize
the rule but regard different hypothetical
actions as what it would take
to follow the rule. It is futile to try to avoid this by including in the
that specifies explicitly
rule an expression
how to apply the rest of
rion
are likewise
the rule, for alternative
of that expression
interpretations
a
In the same way, a law-statement
conceivable.
determinate
specifies
relation only by exploiting
of what
understanding
implicit background
it would take for nature to obey this law. This point applies to any law
or not it blatantly appeals to implicit background
statement, whether
factors."
understanding
by referring to "disturbing
That the proper way to apply a rule is not itself specified by any rule,
of how to follow
intelligible without
implicit background
understanding
seems to think, that nothing
that rule, does not imply, as Hempel
counts as a violation of the given rule. The background
understanding,
albeit implicit, enables
the rule to impose determinate
requirements.
must be capable of being taught and of
implicit understanding
the
if disputes over it ever
made
being
explicit subject of discussion
arise. Some fortune tellers explain away your failure to make accurate
by using their rules as the result of your having misapplied
predictions
to be law-statements,
the claims they believe
of your having ignored
some clause they say they neglected
to mention.
It is doubtful
that they
at all by adopting those
have undertaken
commitments
any determinate
This
"law-statements."
It
is as
if someone
says,
"I
can
run
a
four-minute
but with each failure reveals a proviso
that she had not stated
mile,"
on this track," ". . .except on sunny Tuesdays
earlier: "...except
in
and so on. It quickly becomes
that
will
this
March,"
person
apparent
not acknowledge
herself to any claim by asserting "I
having committed
can run a four-minute mile."
is distinguished
from such bunk
Science
neither by the explicit inclusion in scientific law-statements
of all con?
nor by the absence of implicit
ditions Hempel
would deem "essential"
background
understanding
of how
to apply
those
law-statements.
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
What
242
MARC
is noteworthy
is that this background
is
understanding
In
all
researchers
genuine background
understanding.
identify
general,
the same testable claims as those to which one would become committed
to a certain store of background
by adding a given lawlike hypothesis
on
to
beliefs. Because
it is subject
how
they agree
apply the hypothesis,
to honest
about
LANGE
science
test.
IV
I'll now
to the problem of provisos
argue that this attractive
response
with the view that law-statements
describe regularities
is incompatible
of a certain kind. This account of provisos
leads instead to a conception
as specifying
in a
of law-statements
the claims we ought to respect,
certain context, as able to justify certain other claims.
is the regularity stated by
What,
according to the regularity account,
of the law of thermal expansion?
the familiar expression
Presumably,
the bar has a
it is that a bar's length changes by k L0 AT whenever
certain composition,
its initial length is L0, its temperature
changes by
are
no
and
there
factors. That nearly all of us agree on
AT,
disturbing
is appropriately
said of a
factors"
"There are no disturbing
whether
given
case
saves
this
expression,
and
so
the
law-statement,
from
empti?
the law
of which predictions
ness; we share an implicit understanding
But
it
is
whether
statement
of
when
underwrites,
properly
applied.
certain scientists are correct in saying of a given case that there are no
factors depends not just on the physical features of this case
disturbing
but also on their purposes,
e.g., on the degree of approximation
they
can tolerate considering
the use they intend to make of this prediction.
as laws some uniformi?
Even if the regularity account can countenance
the law of thermal expansion
ties involving the concerns of scientists,
was surely not supposed
the subject
to be such a uniformity;
somehow,
has changed from a law of physics to a law of the science of scientific
activity.
to
that according
for the regularity account.
It gets worse
Suppose
of the law of thermal
the familiar expression
the regularity account,
states that a bar's length changes by k L0 AT whenever
expansion
its initial length is L0, its temperature
the bar has a certain composition,
no
one
on the bar hard enough to cause
is
hammering
changes by AT,
from AL = k L0 AT great enough (given our interests) to
deviations
this claim cannot qualify as a
matter
to us, and so on. Nevertheless,
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
NATURAL
LAWS AND THE PROBLEM OF PROVISOS
243
to the regularity account,
it
law-statement
for the relation
according
someone
states does not obtain.
It is violated,
for example, when
is
on the bar hard enough to cause the actual change in the
hammering
bar's length to depart from k L0 AT but lightly enough for this depart?
ure to be irrelevant to the investigator's
concerns. While
in this case it
is not true that the actual length of the bar changes by k L0 AT, it is
to predict that the length of the bar will
for such an investigator
change by k L0 AT. The relation stated by the law involves not the
in length but rather the change in length one is
bar's actual change
justified in predicting.
This suggests that the law of thermal expansion
doesn't describe a
the way a
regularity among events or states of affairs but concerns
proper
claim concerning AL ought to be
to the problem of provisos
leads
as the objective
fact
expansion
a premise
about
circumstances,
and its composition
temperature,
justified. The response I've advocated
to a conception
of the law of thermal
that under certain (partly pragmatic)
the bar's initial length, its change in
ought to be used to justify a certain
has a
length. On this view, a law-statement
it says that under certain circumstances,
used to justify certain other claims.
are
would have to show that law-statements
claim about its change in
element
because
normative
certain claims ought to be
To succeed, this account
able to explain their instances,
to support counterfactuals,
and to be
their
confirmed
instances
because
inductively by
they specify the roles
This account
that certain descriptions
should play in justifications.
would
likewise have to show that because accidental
generalizations
lack this prescriptive
import, they cannot be used in these ways. To
show this would be to break the familiar unilluminating
circle of analysis
from a law's explanatory
to its capacity
power, to its physical necessity,
to its lawlikeness,
for counterf actual support,
to its capacity
to be
to its explanatory
confirmed by its instances,
power. This
inductively
task is well beyond the scope of this paper; I begin it in my (1993). My
concern here is to show how this normative
account of law-statements
to the problem of provisos.
arises from a plausible
response
a law-state?
Let me summarize
to Hempel,
the argument. Contrary
as
are
no
ment
that includes "so long
there
factors"
is
disturbing
not thereby rendered
trivial. Like any other expression,
this condition
derives
its content from an implicit shared understanding
of how one
should use it. Hence,
the relation that a "law-statement"
that includes
this condition
claims
to obtain,
which
is determined
by the appropriate
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
244
MARC
LANGE
on proprieties
not
way to apply this statement,
ultimately
depends
to "so long as there are
codified explicitly. All there is to give meaning
no disturbing
save the "law-statement"
and thereby
from
factors,"
is
how
scientists
consider
the
"law-statement"
ap?
triviality,
properly
on this point,
there is near unanimity
the "law-state?
plied; because
is not empty. But since the statement
is properly
applied to a
circumstance
when
have
certain
inter?
given physical
only
investigators
that
law
the
constitutes
the
described
ests,
regularity
by this
supposedly
statement must involve not only physical events but also investigators'
concerns. Even
as
if the regularity account regarded such uniformities
do not obtain anyway and so, accord?
laws of physics, these uniformities
ment"
cannot be described
ing to the regularity account,
by law-statements.
a
to
it
sometimes
is and
For with
situation,
given physical
regard
is not appropriate
for us to say that AL will be k L0 AT,
sometimes
on our concerns.
But surely in a given physical
situation,
depending
the bar expands;
there is only one real amount by which
the actual
behavior
of the bar does not depend on our interests. Once
the law
a
on the
one
in
is
relation
"no
statement,
stating
involving
hammering
to
matter
it
to
bar hard enough or otherwise
us,"
disturbing
enough
turns out to involve our interests,
the other relatum, "AL," is found
to be infected by our interests as well.
It refers to the change in the
to the claim that ought to be considered
bar's length according
justified
by these premises, which depends on the purpose for which we intend
to use this claim, rather than to the bar's actual change in length, which
does not. So the law-statement
expresses a norm rather than a regularity
the presence
the
bar's
real
length. It states a relation between
involving
our
to
with
do
of certain conditions
interests) and the
(some having
one
in
to
the
bar's
way
length.
change
ought
predict
able to tell
A
law-statement,
then, informs an audience
already
that if there are none,
factors"
whether
there are "disturbing
they
ought to use a given claim to justify another. On this view, a proviso
tout court, contrary to Hempel.
There
is no fact of
is not "essential"
a law-statement
is
has or lacks provisos. There
to whether
can
discuss
that
law-statement
who
would
those
whether
typically
only
from reading it, what it prescribes
learn, merely
they do (as is the case
second law), or whether making
of Newton's
for the familiar expression
the proviso explicit is essential for that audience.
a normative
does not deny
of law-statements
Moreover,
conception
in nature. It denies only that law-statements,
that there are regularities
the matter
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
NATURAL
LAWS AND THE PROBLEM OF PROVISOS
245
are describing
some of
their distinctive
functions,
performing
in the sense that the regularity account envisions
them. Laws,
them,
need not be invoked to understand what law-statements
say.
when
V
can be used to undermine
not
if successful,
argument,
foregoing
a
account
account
of
law
but
also
that
takes
any
any
regularity
only
to describe a state of affairs that necessarily
law-statement
presupposes
it has been suggested
such a regularity. For instance,
(e.g., by Kneale
a
not
that
law-statements
describe
of
certain kind but
1949)
regularities
non-Humean
connections
of physical necessity.
Such an account implies
The
a regularity among cir?
connection
that the non-Humean
supplements
cumstances.
the problem
of provisos
takes
But with this regularity,
can be deployed
recent
accounts
hold. The same reasoning
against
to which
law
1983, and Dretske,
(see Armstrong,
1977) according
statements describe relations among universals,
such as the property of
the law-statement
by a given amount. On these accounts,
lengthening
to a certain relation
of thermal expansion
entails that events conform
"so long as there are no disturbing
factors." Precisely what this comes
some factor qualifies as "disturbing"
or not, must be
to, i.e., whether
fixed
the law-statement.
I have argued,
the law-statement
But,
a
to
set
determinate
of
conditions
it
appeals
disturbing
only because
states a relation involving not the property of expanding
in length by
a given amount but the property of being able to serve as the subject
of a justified claim attributing
In short,
by a given amount.
expansion
a
if the above reasoning
it
constitutes
goes through,
recipe for an
a
account
to
which
law
is or requires
any
argument against
according
a regularity among events or states of affairs.
a particular
A law-statement
influence, such as the Corio
concerning
lis-force law or Newton's
law, suffices to
two-body gravitational-force
tell persons having comparatively
little background
how
understanding
the magnitude
of
they ought to justify a certain claim. This is because
by
the influence
covered by the law-statement
does not depend on which
influences are at work; hence, while
there are provisos,
there is
that demands
as "in
such
significant background
understanding,
the absence of disturbing
factors" would. The gravitational-force
law,
a
one
for example,
how
should
claim
the
specifies
justify
concerning
two
force
between
whatever
the
other
influences
bodies,
gravitational
other
none
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
246
MARC
LANGE
to reach, say, the total
which
this subtotal should be combined
force on a given body.
to the problem of provisos does
Giere's
However,
(1988) response
to the gravitational-force
not capture the fact that to be committed
law
to treating a certain inference
to the component
is to be committed
force as correct, no matter what the other relevant influ?
gravitational
with
ences.
In light of the problem of provisos, Giere holds certain familiar
to be false as claims about the physical world, but he
law-statements
contends
function as
that, as law-statements,
they must nevertheless
that
tries to find something
of something. He therefore
descriptions
to
and
the
this
has
led
Dretske
describe.
While
search
Armstrong
they
with
whose
exotic realm of universals, Giere
1983,
(along
Cartwright,
that a law-statement
account is similar in all relevant respects) maintains
a scientific model.
One may hold that the relevant behavior
describes
of a given real system can be predicted
by using some model; Giere
terms such a claim a "theoretical
Since the law-statement
hypothesis."
not reality, it needs no qualification
the model,
describes
by provisos
to be accurate.
Consider,
then, what Giere
says (1988, p. 44) about two laws, each
are combined
to account for a mag?
a
that
concerning
single influence,
influenced
pendulum:
netically
We
have
discovered
a new
kind
... in which
the force of gravity
is supple?
of pendulum
the point of rest. Constructing
toward a point below
to such systems
in physics.
is a fairly easy problem
force directed
by a magnetic
a theoretical
model
that does apply
mented
a
law specifies how one ought to calculate
If the gravitational-force
influence of one body on another) no matter
subtotal (the gravitational
law does
and the magnetic-force
the other influences present,
what
likewise, then this is indeed an easy problem. Having
already accepted
the proper way to add
and having recognized
these law-statements,
we
forces and to use the net force on a body to infer its acceleration,
the bob's motion.
for predicting
to a particular procedure
not to prescribe which models
But Giere
takes these law-statements
to describe
but merely
to use (which is the job of theoretical hypotheses)
that the
admit that by accepting
certain models.
Thus, Giere must
we are not committed
law
describes
certain
models,
gravitational-force
to predict the bob's
it describes
to saying that one ought to use a model
one who accepts the theoretical hypothesis
that this
motion. Moreover,
force exerted by the
the gravitational
law should be used to calculate
are committed
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
NATURAL
LAWS AND THE PROBLEM OF PROVISOS
247
to the theoretical
bob is not thereby committed
earth on a non-magnetic
to
that
this
law
should
be
used
calculate
the gravitational
hypothesis
force exerted by the earth on a magnetically
influenced bob.
to accept this law-statement
is to recognize
On my view, in contrast,
the component
the way one should justify claims concerning
gravi?
tational force, whatever
the other relevant
influences. This view ac?
to
counts for what Giere's
That we are committed
view obscures:
some common element
in our treatments of ordinary and magnetically
an identical way of justifying a certain
(namely,
augmented
p?ndula
the magnetically
is
in virtue of which
augmented
subtotal)
pendulum
an easy problem. We became
to elements
committed
of its solution
to other problems.
when we adopted solutions
I have argued against Hempel's
law
contention
that a complete
own
on
statement must
its
without
range
specify
depending
implicit
in order to do so. I have also argued against
background
understanding
Giere's
alternative
claim that a law-statement
says
(and Cartwright's)
to specify.
about
its
it
for
theoretical
range, leaving
nothing
hypotheses
I have defended
the view that a law-statement
specifies its range in a
fashion that may involve blatant appeal to implicit proprieties
of use.
I have thereby tried to offer a way around the problem of provisos,
at
the price of abandoning
the regularity account of law in favor of a
normative
of that proposal must await an?
analysis. Further discussion
other
occasion.
REFERENCES
Armstrong,
Ayer, A.
Essays,
Braithwaite,
Canfield,
Science
D.:
J.:
1983, What
"What
1963,
Macmillan,
R. B.:
Is a Law
is a Law
London,
pp.
of Nature?,
of Nature?",
209-234.
Press, Cambridge.
Cambridge
University
in The Concept
and Other
of a Person
1953, Scientific Explanation,
Cambridge
J. and Lehrer,
K.: 1961, 'A Note on Prediction
Press, Cambridge.
University
and Deduction',
Philosophy
of
28, 204-208.
N.:
1983, How
the Laws of Physics
Lie, Clarendon
Press, Oxford.
Deductive
1968, 'Discussion:
Predictions',
35, 279-283.
Philosophy
of Science
F.: 1977, 'Laws of Nature',
Dretske,
Philosophy
of Science 44, 248-268.
and Generalizations',
in A. Grunbaum
and W. Salmon
Giere, R.: 1988, 'Laws, Theories,
The Limits
of California
Press, Berkeley,
(eds.),
of Deductivism,
University
pp. 37
46.
Cartwright,
J.:
Coffa,
Goodman,
N.:
1983,
Fact,
Fiction,
and Forecast,
4th
ed.,
Harvard
University
Cambridge.
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
Press,
MARC LANGE
248
in Aspects
'Studies
in the Logic
of Explanation",
C. G.:
1965a,
The Free Press, New York,
pp. 245-295.
Explanation,
and the Social
in the Natural
Methods
C. G.:
1965b,
'Typological
Hempel,
Ibid., pp. 155-172.
Hempel,
of Scientific
Sciences',
and W. Salmon
in A. Grunbaum
C. G.:
1988, 'Provisos',
(eds.), The Limits
Hempel,
19-36.
of
California
Press,
Deductivism,
pp.
Berkeley,
University
Oxford
W.:
and Induction.
Press, Oxford.
1949, Probability
Kneale,
University
B.:
M.
Nous,
'Lawlikeness',
1993,
forthcoming.
Lange,
of
Harvard
Press, Cambridge.
1973, Counterfactuals,
University
in
to Objective
to 'A Subjectivisms
Guide
Chance",
1986,
'Postscript
Press. Oxford,
Volume
2, Oxford
pp. 121-126.
University
Papers:
Philosophical
in R. S. Butler
and Causal Laws*,
J. L.: 1962, 'Counterfactuals
Mackie,
(ed.). Analytic
and Noble,
New York,
Barnes
pp. 66-80.
Philosophy,
New York.
and World,
E.: 1961, The Structure
Brace,
of Science, Harcourt,
Nagel,
New York.
H.:
1947, Elements
Reichenbach,
Logic. Macmillan,
of Symbolic
Oxford University
Press, Oxford.
1930, The Right and The Good.
Ross, W. D.:
in H. Feigl and
Laws
of Physical
'The Key Property
1961,
Scriven, M.:
Inaccuracy',
and
Issues
in
the
Current
Science.
Holt,
G. Maxwell
Rinehart,
Philosophy
of
(eds.),
Lewis,
D.
Lewis,
D.
Winston,
Wittgenstein,
Macmillan,
K.:
in
K.:
New
York,
pp.
91-104.
L.:
1958, Philosophical
New York.
submitted
Manuscript
received
Final version
September
April
21,
Investigations,
26,
3rd ed.,
trans. G. E. M.
1991
1992
of Philosophy
Department
Los Angeles
of California,
University
Avenue
405 Hilgard
Los Angeles,
U.S.A.
California
90024-1451
This content downloaded from 129.118.6.113 on Tue, 12 Nov 2013 21:02:48 PM
All use subject to JSTOR Terms and Conditions
Anscombe,