I. Do Theoretical Entities Really Exist?

 

Some people have wondered whether theoretical entities really exist. Depending upon just what you count as a “theoretical entity”, this may seem to be a ridiculous worry. If theoretical entities are simply currently unobserved entities, then surely some theoretical things do exist. (Neptune was a theoretical entity, yet it exists.) But use the term in a more narrow way - to refer only to things that can’t be observed. And take the question to be one about what we can reasonably believe about such things. Rosenberg says that electrons, gravity, etc. are like this.

 

As Rosenberg explains, some broad theoretical concerns can generate the worry: a) science is supposed to be characterized by its ability to achieve understanding through experiment and observation. But then it’s odd that in the end that understanding rests on claims about things that cannot be observed. b) one can wonder what exactly the words for the theoretical terms mean. People have had the idea that simple words refer to observable properties, and complex words refer to complexes of these. See p. 86 in R. And see top of p. 88 for reason why this is no good. What follows largely attempts to spell out this latter point.

 

II. Instrumentalism

 

See p. 94, bottom for R’s description of this. The contrasting view is “realism”. Let’s apply this to theoretical entities. One way to think about this begins by considering sentences such as:

 

1. The average American family has 1.8 children.

 

Let’s assume that this is true. But there really is no such thing as the average American family. (1) seems to contain a term referring to that family, but it’s just shorthand for a more complex fact:

 

2. The total number children in American families divided by the total number of American families is 1.8.

 

Note: (1) is not false or misleading. No criticism is implied. You might say that the average American family is a “convenient fiction.” There is no such thing, but it is convenient to talk as if there is such a thing since it avoids more cumbersome talk like (2). One might think that theoretical entities are like this - convenient fictions, and talk of them is to be cashed out in terms of familiar things. But this seems to be wrong. The point can best be appreciated by thinking about entities that enter into theoretical explanations other than physics. Consider Freudian psychology. It says that our minds involve an id, ego, and superego. We’re told that a person whose behavior is dominated by his id engages in reckless or aggressive behavior (or something like that). [Actual details don’t matter here.] But you might wonder about the status of the id. Is it really in there somewhere? Or is it more like the average American family? Some call these entities “constructs.” If you say that “people whose behavior is dominated by their ids behave aggressively” just means that they behave in certain ways behave in those ways, then the theory seems not to explain anything - it just describes . But it is supposed to explain, not just describe. That’s why some people deny that there are ids - they think the theory does not get things right. Yet if it just described certain behavior, it seems that it does get it right and the critics are (obviously) mistaken.

 

Another example may bring out another aspect of the issue. Suppose that you notice some big foot-shaped prints in the sand. You call the thing that made them “bigfoot”. Suppose you go on to say various things about bigfoot - his size, age, looks, etc. So you have the theory that there is such a thing as Bigfoot, and it caused the prints.

 

What determines whether bigfoot exists? If you define “Bigfoot” as “whatever it is that made the prints”, then unless the prints arose spontaneously, there is bound to be such a thing as bigfoot. Your theory seems completely trivial. People who doubt the existence of bigfoot are making some elementary mistake. But if you define it more precisely, then there might not be such a thing. Suppose it turns out that the prints were made by an otherwise normal bear with a large foot. Consider two claims:

 

3. It turned out that bigfoot didn’t exist.
4. It turned out that bigfoot was a bear.

 

Which is right? We have to turn to issues about how language works to get an answer to this. Roughly, if the word “bigfoot” just means “the thing that made the prints”, then (3) is true. If the word has more content - includes the other stuff believed about it - then (4) is true. Somewhat similarly, whether the id really exists will depend in large part upon just how rich the definition of the term is. If a lot is packed in, then it is controversial (at least) whether it exists. If it just means “whatever is responsible for” some bit of behavior, then maybe it is less controversial. (Not uncontroversial, if it is required that there be one thing that causes multiple bits of behavior.)

 

So, one aspect of the issue about theoretical entities turns on issues about what is packed into the definitions of theoretical terms. For now, it’s enough for us to see this dependency.