When Darwin proposed natural selection, he didn't just describe patterns in nature—he revealed a mechanism that made sense of countless biological observations. Similarly, when Newton explained planetary orbits through gravitational forces, he transformed astronomical data into genuine understanding. But what exactly separates a mere description from a true scientific explanation?

This question has puzzled philosophers of science for decades because it touches the heart of what science achieves. We don't just want to know that something happens; we want to understand why it happens. Yet different scientific fields seem to explain things in remarkably different ways—physicists invoke universal laws, biologists describe mechanisms, and some explanations seem powerful precisely because they unify diverse phenomena.

The covering law model, developed by Carl Hempel, treats explanation as logical deduction. To explain why a specific piece of copper conducts electricity, we cite the general law that all copper conducts electricity, note that this object is copper, and deduce that it must conduct. The event is 'covered' by the general law, making its occurrence logically inevitable given the circumstances.

This approach works beautifully for many physical phenomena. Why did the mercury in the thermometer rise? Because metals expand when heated (general law), mercury is a metal (initial condition), and it was heated (triggering condition). The explanation shows the event wasn't random but followed necessarily from natural laws. It's the same logical structure whether explaining planetary orbits through Newton's laws or chemical reactions through thermodynamics.

Yet the covering law model struggles with common explanations. When we explain a forest fire by citing a dropped cigarette, we're not invoking a law that 'all dropped cigarettes cause forest fires'—most don't. The model also makes problematic symmetries: if we can explain a flagpole's shadow length from its height, the model suggests we can equally explain the height from the shadow. But only one direction seems genuinely explanatory.

The mechanistic approach focuses on revealing how things work rather than deducing from laws. To explain how DNA replicates, biologists describe the actual process: how helicase unzips the double helix, how polymerase reads each strand, how complementary bases attach. Understanding comes from seeing the step-by-step causal process, not from subsumption under general laws.

This approach dominates biology, neuroscience, and many social sciences where universal laws are rare but mechanisms abound. We explain memory formation through synaptic strengthening, disease spread through transmission pathways, and economic crashes through cascading failures. Each explanation succeeds by exposing the hidden machinery—the gears and pulleys beneath surface phenomena.

Mechanisms provide understanding even without perfect predictability. We grasp how smoking causes cancer through DNA damage and failed repair mechanisms, despite being unable to predict which smokers will develop tumors. The explanation's power comes from revealing causal pathways, not from enabling deduction. This matches our intuitive sense that understanding means knowing how things work, not just knowing regularities.

The most celebrated scientific explanations often succeed by unifying phenomena that seemed unrelated. Darwin's natural selection explained not just species diversity but also geographical distribution, embryological similarities, and vestigial organs. Maxwell's electromagnetic theory unified electricity, magnetism, and light. These explanations transform our worldview by revealing that apparently different things are manifestations of the same underlying process.

Unification provides understanding through pattern recognition at the deepest level. When we see that planetary orbits and falling apples obey the same gravitational principle, or that heat and motion are both forms of energy, scattered facts crystallize into coherent pictures. The explanation's power isn't just in covering individual cases but in reducing the number of independent assumptions needed to understand nature.

This suggests explanation is partly about cognitive economy—helping limited minds grasp nature's complexity by revealing simplicity beneath diversity. The best theories explain more while assuming less. String theory's appeal, despite limited evidence, stems partly from its promise to unify all forces. Whether through laws, mechanisms, or unification, good explanations make the mysterious manageable by showing how much follows from how little.

Good scientific explanation isn't monolithic—it takes different forms for different purposes. Sometimes we need the logical certainty of covering laws, sometimes the concrete detail of causal mechanisms, and sometimes the revelatory power of unification. Each approach transforms mere description into genuine understanding by answering different aspects of 'why.'

Perhaps the diversity of explanatory styles reflects the complexity of nature itself. Physical systems might be best understood through laws, biological systems through mechanisms, and fundamental reality through unification. Rather than seeking one true model of explanation, we should appreciate this plurality—recognizing that understanding nature requires multiple tools for revealing its varied forms of order.