The Anatomy of a Good Explanation: Beyond Falsifiability

In the discourse of scientific philosophy, there is a common misconception that a theory’s value is derived primarily from its falsifiability. However, as Brett argues, falsifiable theories are "a dime a dozen." Merely being testable does not equate to being a high-quality explanation. To understand what truly constitutes a robust scientific theory, we must look beyond basic testability and examine the structural integrity of the explanation itself.

The Trap of Untethered Testability

The fundamental problem with relying solely on falsifiability is illustrated by the "grass cure for the common cold" example. If someone proposes that eating 10 kilograms of grass cures a cold, they have presented a "testable theory." Yet, this theory is worthless because it lacks an explanation of the underlying mechanism.

This lack of mechanism allows for infinite, post-hoc adjustments. If the initial test fails, the proponent can simply shift the goalposts—claiming "11 kilograms might do it," or suggesting a different type of grass or a different time of day is required. While the theory remains "always testable," it is stagnant. You are "not making any progress" because the theory is not anchored to a causal logic; it is infinitely flexible and, therefore, explanatory of nothing.

The Criterion of Being 'Hard to Vary'

A central pillar of a good explanation is that it must be "hard to vary." A theory is intellectually robust only when its components are so precisely interconnected that changing any part of the explanation would require a complete overhaul or lead to the collapse of the entire framework.

This is contrasted by the ancient Greek mythological explanation for the seasons. The story of Persephone leaving Hades is "very easy to vary." One could replace Persephone with Nike or Hades with Zeus without changing the outcome or the logic of the narrative. Because the variables are arbitrary and interchangeable, the theory lacks explanatory power.

In contrast, the "axis tilt theory" regarding Earth’s seasons is "hard to vary." By positing that the Earth is angled at 23 degrees, the theory creates a rigid structure. If you were to change the degree of the tilt, the predictions regarding the sun's position and the length of summer and winter would break. The precision is not arbitrary; it is a necessary feature of the mechanism itself.

Risky Predictions and Scientific Progress

Beyond being hard to vary, a good explanation must make "risky and narrow predictions." A theory that explains everything explains nothing. True scientific progress occurs when a theory stakes its reputation on specific, measurable outcomes that would be difficult to achieve by chance.

Brett highlights the confirmation of Einstein’s theory of relativity as the gold standard for this. When Eddington conducted the experiment observing that "star light gets bent around an eclipse," he was testing a "risky prediction." This was not a vague observation that could be retroactively fitted to the data; it was a precise, narrow expectation derived from the theory of relativity. Because the prediction was so specific and difficult to observe, its success served as a powerful validation of the theory.

Conclusion

To summarize, a scientific theory is a specific subset of a good explanation. It must go beyond being merely falsifiable. It must provide a causal mechanism that makes the theory "hard to vary," ensuring that the explanation cannot be easily manipulated after the fact. Finally, it must venture into the realm of "risky and narrow predictions," providing clear, falsifiable boundaries that, when confirmed, demonstrate the theory’s genuine insight into the fabric of reality.