Many dominant concepts in contemporary physics—such as string theory–are anchored in speculations that cannot be tested under present circumstances. So it is no surprise that scientists and philosophers find themselves wrestling with speculation’s proper role in science.

A significant consensus exists for scientists speculating first, and then testing any hypothesis they make. But this position is bookended by more radical views. Sir Isaac Newton famously argued speculation has no place at all in science. Austrian philosopher Paul Feyerabend saw even unfettered speculation as the indispensable engine of scientific progress.

Professor of philosophy Peter Achinstein’s new book, Speculation: Within and About Science, is a powerful intervention in this timely debate. Scientists must speculate, says Achinstein. But speculation has always possessed a certain slipperiness.

“Speculation is a concept that needs to be clarified,” says Achinstein. “And in analytic philosophy, that’s what we try to do. No philosopher or scientist has defined it. And, then, how do we evaluate it? I say: Let’s evaluate a speculation as a speculation. And you can evaluate a speculation. You can say of a speculation, for instance: Why did you make that assumption?”

In works such as Particles and Waves (which won the Lakatos Award in 1993), The Book of Evidence, Scientific Evidence, and Evidence and Method, Achinstein developed a standard for scientific evidence that requires the high probability of an explanatory connection between evidence and hypothesis.

String theorists and other researchers seeking unifying explanations for phenomena in the physical universe favor a different standard of evidence, rooted in the work of 18th-century statistician and philosopher Thomas Bayes.

“Bayesian evidence is anything that increases the probability of your theory,” Achinstein explains. “It’s a crazy idea. And it’s a dangerous idea. If string theorists think that anything that increases the probability of their theory is evidence, then they probably won’t worry about the absence of empirical tests. They will probably think, and in some cases write, that an appeal to the simplicity, beauty, or necessity of a theory will increase its probability and hence count as evidence in its favor.”

The claim that nature is simple, and hence that a simple theory is much more likely to be true than a complex one, has been made by many scientists, including Isaac Newton and Albert Einstein. Achinstein demurs.

“It is unclear what it means to say that nature is simple,” he says. “A simple object such as an electron can be subject to many different laws of considerable complexity. Moreover, there is absolutely no reason to think that simplicity, whatever it is taken to mean, is a sign of truth.”

In Speculation, Achinstein challenges an important claim made by string theorists and many other scientists. The claim is that there exists some fundamental theory, a so-called “theory of everything” (TOE), that will explain and unify all the phenomena in the universe by appeal to one set of basic laws and objects. He points to an assertion by philosopher Thomas Nagel, who seeks to provide such a theory, that in order to do science, you have to presuppose that the world is simple, unified, and completely understandable.

Achinstein takes issue with that. “Why couldn’t it be the case that, as God says to Job: ‘It’s too complicated for you?’ The claim that there is a TOE, or that scientists must presuppose that there is, is itself a speculation, indeed a very grand one.”

If there is no TOE to explain everything, or if scientists need not suppose there is, what kind of understanding should scientists attempt to provide, particularly when they speculate?

Achinstein proposes the concept of “Newtonian intelligibility”—a form of theorizing modeled on Newton’s approach to gravity in his work Principia. This type of theorizing does not seek to explain everything. Instead, it offers useful explanatory insights within the scope of the particular (and narrower) questions being raised, such as “what keeps the planets and their moons in the orbits they have?” and “why do unsupported bodies fall the way they do?”

In Speculation, Achinstein also urges a closer look at the history of science. He says that one of his heroes, James Clerk Maxwell, is an exemplar of the pragmatic flexibility and definitional rigor he seeks in scientific speculation.

“Maxwell’s kinetic theory is still taught today,” observes Achinstein, despite the fact this it was, by any definition, speculation when the physicist worked it out in the late 19th century.

“Science has many different activities,” Achinstein adds. “We don’t lionize Maxwell because of his experimental results. It’s part of science to speculate.”

Ultimately, Achinstein sees the book as a way to decouple an essential scientific pursuit from the burdens of having to explain, well, everything.

“These days, scientists don’t read much philosophy,” says Achinstein. “I hope this book will give them a defense they need to pursue speculative ideas that have not been tested and perhaps will not be for a long time, if ever.”