Laws of Nature

I

Laws of Nature, pervasive general features of the world or descriptions of those features. Laws of nature are central to scientific investigation, in part because of the role they play in the prediction and explanation of natural phenomena. From antiquity people have considered why there should be regularities in nature. The status of laws has been a central topic in the history of theology because a deity provides a possible source of regularities and because of theological interest in miracles as violations of natural law. Laws of nature are also a central topic in the philosophy of science. There are two main philosophical questions about laws: the epistemological question asks how laws are discovered, while the metaphysical question asks what sort of thing they are.

II

THE EPISTEMOLOGICAL QUESTION

The epistemological question presents difficulties because scientists cannot deduce laws of nature from their data. For example, no matter how many ravens have been observed to be black, it cannot be deduced that all ravens are black: the possibility always remains that there is an unobserved raven that is not black. The logical gap between data and laws is even wider for scientific laws that concern entities, properties, and processes that are not just unobserved in fact but unobservable in principle. Scientists cannot actually observe that some light is a stream of photons, for example, much less that all light is like this. The evidence for such theoretical laws can thus only be indirect. Much of the philosophy of science is concerned with giving an account of the form such inferences take and of the justification that scientists can have for making them.

III

THE METAPHYSICAL QUESTION

At first glance, the metaphysical question about laws seems easier to answer. Difficult though it may be to state how scientists discover the laws of nature, it seems simple enough to say what sort of thing they are discovering. According to the straightforward answer to the metaphysical question, laws are simply universal patterns in nature or true universal generalizations, such as “all ravens are black”, or general mathematical relationships that science describes, such as Boyle’s law, which claims that the pressure of a given amount of gas at a constant temperature is inversely proportional to the volume it occupies, so that if the volume is reduced, the pressure is increased.

Such an answer is misleading, however, because not all true universal generalizations are laws. Even if by some extraordinary coincidence it turns out that every philosopher has been right-handed, it would not be a law that all philosophers are right-handed, though this universal generalization would be true. Similarly, even though it is presumably true that all the gold spheres that there will ever be will have a diameter of less than 100 miles (there being only a limited amount of gold in the world), this is not a law. Universal generalizations that would not be laws even if they were true are known by philosophers as “accidental generalizations”; on the other hand, generalizations that are laws if true are known as “law-like generalizations”. To understand what it is to be a law of nature requires explaining the difference between law-like and accidental generalizations. Philosophers have not found this an easy task.

Law-like statements appear to say more than accidental generalizations by describing not just how things are, but also how they would have been under different, counter-factual conditions. If it is a law that pressure is inversely proportional to volume, that is not just to state that the pressure goes up when the volume is decreased: even if the volume of a given sample of gas is kept the same, it is still the case that the pressure would have increased, had the volume been reduced. By contrast, even if it is true that all philosophers are right-handed, it is not the case that had a left-handed person taken up the subject, he or she would have suddenly become right-handed.

A second apparent difference between law-like and accidental statements is connected to the way laws are discovered. Only laws seem susceptible to inductive support. For example, if Boyle’s law is found to hold wherever it is tested, scientists may infer that it also holds in unobserved cases. By contrast, even if all the philosophers who were asked turned out to be right-handed, it would not be inferred that the same holds for the philosophers who had not been asked. The only way to be confident that all philosophers are right-handed would be to ask every one of them.

There are thus at least two apparent differences between laws and accidents: only laws have counter-factual content and only laws are susceptible to inductive support. It is not clear, however, that these two tests for lawfulness give the same answer for every generalization or that either gives the correct answer for every generalization. In addition, even if these tests were reliable, they only provide symptoms of law-likeness, rather then confirming directly what laws of nature are like. Philosophers have thus attempted to give deeper accounts of laws of nature and of what distinguishes them from accidental generalizations.

IV

DEFINING LAWS

Although there are many different and conflicting accounts of laws, they all fall into two groups. Almost all philosophers of science agree that some universal generalizations are not law-like. However, the accounts of the first group—sometimes known as “regularity views”—all maintain that law-like statements are simply a special class of universal generalizations, whereas the accounts of the second group afford law-like statements a different sort of content.

A

Regularity Views

There are three popular types of regularity view. According to the first, the “attitude” view, what distinguishes law-like from accidental generalizations is nothing in the generalizations themselves: it is instead a difference in our attitudes towards them. Thus, a supporter of the attitude view might say that law-like statements are those that scientists are willing to use for purposes of explanation or prediction. It is not that they are used in this way because they are laws; rather, they count as laws because they are used in this way.

According to the second account, known as the “projectible predicate” view, a generalization only counts as law-like if it contains the right sort of terms: for example, predicates such as “pressure” and “volume” can figure in laws, while perhaps “philosopher” or “right-handed” cannot. Proponents of the projectible predicate view then go on to consider what distinguishes “good” from “bad” predicates. Some proponents hold that good predicates correspond to natural kinds, whose members are fundamentally similar to each other, while bad predicates lump together dissimilar things.

The third type of regularity account is the “best system” view, in which laws of nature are those true generalizations that would feature in an ideal theory of the world. Such a theory could be constructed if everything was known and an optimal balance struck between the conflicting criteria of describing as much as possible while keeping things as simple as possible. Hence, many generalizations would fall within this theory, though many would be left out, since including them would incur too great a loss of simplicity for too small a gain in strength. The universal generalizations within the system are the laws; those without are the accidental generalizations.

B

Non-Regularity Views

Philosophers who reject the regularity view of laws also offer three different accounts. There is, first, the “property” view, according to which laws do not describe patterns of particular events in nature, but rather give relations between general properties. Thus, the generalization about philosophers being right-handed is not a law, because it is only a claim about particular philosophers; Boyle’s law, by contrast, is a law, because it describes a relationship between the general properties of pressure and volume.

A second account is the “necessity” view, according to which the essential difference between accidental and law-like statements is that whereas accidentals only describe what does happen, laws describe what in some sense must happen.

The final non-regularity account is the “capacities” view, in which laws are taken to be radically different from universal generalizations, describing instead the stable tendencies or capacities of physical objects. On the capacities view, laws do not describe patterns of actual behaviour of objects but rather the forces that underlie them, forces that combine in complex ways which preclude simple patterns. This view accounts for the fact that many laws of nature, such as Boyle’s law, describe only what happens under ideal conditions, when only a single set of forces is in operation. The aim of this view is to show how simple laws may reveal truths about a complex world.