One answer which is sometimes given to the question in the title of this paper is "Sugar and spice and everything nice." Another answer which we have heard is "Animal cells of a certain sort, which are made up of various complex organic molecules, which are made up mainly of atoms of carbon, oxygen, hydrogen, nitrogen, and a few others, and which are in turn made up of elections, neutrons, and protons, plus some additional particles such as positrons, photons, neutrinos, various sorts of mesons, and many many more."

Although there is something appealing about each of these answers to our question, one has the feeling that they are incompatible, and that therefore, they cannot both be accepted. Someone overly impressed by the achievements of modern science might feel compelled to accept the latter, and thus to reject the former.

This makes us unhappy. The former answer to the question is not only much prettier, but its merit seems to us to be at least equal to the merit of the latter. That latter answer admittedly sounds more scientific, but this may just be because it sounds more complicated, more sterile, and less poetic. This of course is no reason for preferring it. Are there any reasons to prefer the atomic theory to the sugar-and-spice theory?

In a recent paper Dr. Gerald Feinberg has suggested that "current scientific theory provides an adequate answer to the question," and that furthermore, the atomic theory is true. (1) This latter reason would be the best of all possible reasons for preferring the atomic theory to the sugar-and-spice theory. The former would not. However, we do not feel that Dr. Feinberg or anyone else has demonstrated the truth of the atomic theory.

The position taken in Dr. Feinberg's paper is unfortunately a familiar one. However, for practical reasons we shall try to confine our criticism of this position to Dr. Feinberg's particular statement of it. Of course the question to which Dr. Feinberg is referring is a more general one than ours, namely "How, and of what, is the world made?", but that should not seriously affect the discussion. If the atomic theory can tell us what the world is made of, then surely it can tell us what little girls are made of.

1. Explanatory Power. One reason cited by Dr. Feinberg in support of his thesis is that "many aspects of the world around us can be understood by supposing that matter is composed of atoms" (p.6). To this we reply simply that many aspects of little girls can be understood by supposing that they are composed of sugar and spice and everything nice. For example, little girls are often very sweet things.

We might further emphasize the explanatory power of the sugar-and-spice theory by pointing out that certain differences between little girls and little boys can be quite neatly accounted for by this theory. Little boys are frisky, rowdy, and mischievous because they are composed differently from little girls: on our more general theory, they are made of frogs and snails and puppy-dogs' tails.

Furthermore, many of these properties of little girls which can be explained by the sugar-and-spice theory cannot be accounted for in terms of the atomic theory. At least we are unable to find any publication of such an account, and it seems unlikely that an explanation of the sweetness of little girls in terms of their molecular, atomic, or sub-atomic structure is forthcoming.

Admittedly, the sugar-and-spice theory has its deficiencies too. We only wish to point out that there are some areas in which it is a better explanatory tool than the atomic theory. Both theories are incomplete.

2. Evidence. Further on in his paper, Dr. Feinberg says that "In the twentieth century, through use of techniques such as x-ray diffraction, such direct evidence [for the atomic theory] was finally obtained" (p. 10). We disagree with this statement.

First of all, to what "direct evidence" is Dr. Feinberg referring? The classic x-ray diffraction experiments, those of von Laue and the Braggs of around 1912, are just not very relevant to the question of the existence of atoms. These experimenters assumed the existence of atoms, and went from there to make inferences about the nature of x-rays. More relevant to Dr. Feinberg's thesis might be the sedimentation experiments of Perrin, the work of Einstein on Brownian motion, or the experiments of Geiger and Marsden on the scattering of alpha- and beta-particles.

But certainly even these experiments provided nothing which can properly be called direct evidence for the existence of atoms, as even these experimenters themselves realized, As a matter of fact, these experiments might be cited as classic examples of the so-called indirect evidence for the atomic theory.

This "indirect evidence" is not really evidence at all. What is meant by calling the results of these experiments "indirect evidence" for the atomic theory is just that this theory can be used to give a tidy explanation of the results of the experiments. (Dr. Feinberg prefers to dodge the un tidy aspects of the atomic theory, involving some 200 additional subatomic particles, which he feels to be irrelevant to his thesis.)

But the fact that many experiments yield results for which the atomic theory can provide an explanation surely does not verify the atomic theory. It does not verify the atomic theory any more than encountering a large number of little girls who are sweet verifies the theory that little girls are made of sugar and spice and everything nice. We are brought back to the problem raised in the preceding section: the usefulness of something is not evidence for its truth.

3. Inconsistencies. In section 1, we noted that although the sugar-and-spice theory is incomplete, in that it cannot account for every property that little girls have, the atomic theory seems incomplete in this same sort of way. It is also worthwhile to point out that the atomic theory is not consistent with all of the observed data.

The most well-known sort of inconsistency in the atomic theory is the particle-wave duality exhibited by the elementary objects of the theory. Electrons, for example, have been shown to be subject to diffraction and to the creation of interference patterns--behavior which is usually thought of as foreign to particles and peculiar to waves.

If it is justifiable to discount these apparent inconsistencies and to accept the atomic theory in spite of them, it would seem equally justifiable to discount certain apparent inconsistencies in the sugar and spice theory. Thus we might simply ignore the fact that a lock of hair from a little girl does not sweeten a cup of coffee in which it is placed, although this seems inconsistent with the view that little girls are partly made of sugar.

Alternatively, we might choose to adjoin some restrictions or complications to the sugar-and-spice theory to account for such apparent inconsistencies. This is a perfectly standard sort of move in any scientific enterprise.

4. Ultimates. There are many apparent reasons for preferring the atomic theory to the sugar-and-spice theory. We can briefly mention two of the main ones. It should be noticed again however, that despite the appeal of these reasons and the weakness of certain possible replies to them, they do not involve the claim that the atomic theory is true.

The first of these involves the suggestion that atomic theory is more ultimate than the sugar-and-spice theory; the sugar-and-spice theory purports to analyze little girls, but it analyzes them as sugar and spice for the most part. And sugar and spice are themselves not ultimate simple substances, but need to be analyzed themselves. On the other hand, the atomic theory (so this line goes) gives a complete analysis of matter and a wide category of its properties in terms of ultimate particles. In fact sugar is itself subject to analysis by the atomic theory, but the situation is not reversible.

In fact the atomic theory is systematic as well as ultimate. It accounts for a wide category of properties of matter of diverse kinds, and provides a systematic and quantitative method for investigating such physical properties. The sugar-and-spice theory is not systematic in this way. Perhaps it is therefore not even entitled to be called a theory.

To this pair of objections we reply that (a) it is not clear how atoms (or their constituents) are supposed to be ultimately simple objects. In the first place, if they are not tiny bits of matter, then most talk about them is unhelpful metaphor that does not provide an understanding of what they are, and if they are tiny bits of matter then they do not tell us what matter is in the first place. And we repeat the fact that atomic scientists are continually forced to "discover" new ultimate particles to explain things that the old ultimate particles could not explain.

The number of new particles "discovered" seems to increase exponentially with time and further experiments. This of course is extremely suspicious. There is speculation among scientists that particle theory, like formalized arithmetic, might be essentially incompletable, in the sense that no finite number of particles will be able to explain everything.

Even if the prosecution proves that the defendant had motive and opportunity for the murder, that his fingerprints were all over the dagger, etc., if the defense shows that the prosecution's case also requires the assumption that the accused made his getaway with the aid of 200 trolls, ghouls, ghosts, elves, leprechauns and demons, then the jury must acquit. And any prosecutor who was forced to assume such a fantastic cast of conspirators to make most of his cases, to explain his theories of how crimes were committed, should really be replaced.

And we reply (b) it is not essential to the sugar-and-spice theory that it remain unsystematic. Someone might develop a wide-ranging and systematic account of matter in which sugar, spice, frogs, snails, and puppy-dogs' tails were taken to be the ultimate simples, and other properties of matter were quite systematically and quantitatively accounted for in terms of these things.

In any case, the issue around these objections and replies is not one of the truth of the atomic theory, but of its preferability on other grounds. We will close our discussion with a comment on the limitations of the atomic theory, and a recommendation that physicists either abandon it entirely, or at least view it more modestly.

5. Substances. We have concentrated mainly on Dr. Feinberg's statements about current physical theory. But it should be pointed out that there are more basic objections which can be raised. We feel that it is a serious error to interpret the question "How, and of what, is the world made?" to be a question about the "structure and composition of bulk matter, by which is meant the objects and substances we find around us" (p. 6). Putting this sort of interpretation on the original question limits it improperly. There are hosts of things in the world which are neither substances (in that sense) nor bulk matter. Even if the atomic theory were the truth about the composition of bulk matter, it would not ipso facto have told us the whole truth about how the world is made.

Little girls, for example, may be substances--we find it hard to say. But it would seem quite brutish to call them "bulk matter." And of course many other quite ordinary things are neither substances nor bulk matter: the square root of two, Saturday morning, middle C, the Supreme court, and San Francisco Bay. The analysis of these objects is up to mathematics, sociology, metaphysics, etc. --not physics. But they are objects, because everything is an object.

The atomic theory may be adequate for explaining many properties of "bulk matter." But to say this is to make a claim which is substantially weaker than the claim that it gives a complete account of how the world is made. And it is even weaker still than the brazen claim that the atomic theory is true.