Induction, Hume, and the Ancient Rocks

Midway through the essay, Gould’s argument descends from geology into philosophy, or perhaps reveals that geology had been philosophical all along. He writes that methodological uniformitarianism functions as a “warrant for inductive inference.” This is a compact way of saying something enormous: geology depends on the possibility of reasoning from what is observed to what is not observed, from present causes to past events, from finite evidence to historical explanation.

Induction is the intellectual engine that allows such movement. If modern glaciers scratch bedrock in recognizable ways, geologists infer that ancient scratches may have been made by ancient glaciers. If modern sedimentary environments produce certain structures, geologists infer similar conditions from similar ancient structures. If modern organisms leave tracks, burrows, shells, or chemical signatures, paleontologists use those relations to interpret fossil evidence. The present does not hand over the past directly. It offers patterns from which the past can be inferred.

Gould’s point is that this inferential movement requires an assumption. We must assume some “spatial and temporal invariance of natural laws.” Without that, there is no stable bridge between present and past. Ancient striations might not mean what modern striations mean. Fossil shells might not bear any lawful relation to living shells. Chemical signatures might not follow familiar principles. Every trace would become an isolated curiosity rather than evidence.

This is where Hume enters. Gould notes that the invariance assumption cannot be proved without circularity. Hume famously challenged the rational foundation of induction. We expect the future to resemble the past because it has done so before, but that argument already uses induction. We cannot step outside experience to prove that experience will continue to be reliable. Gould writes that efforts to validate induction were largely abandoned after Hume showed their futility.

For geology, this philosophical problem has a special flavor. Geologists do not simply predict future occurrences from past ones. They infer vanished causes from surviving traces. The event is gone. The glacier that scratched the ancient rock is gone. The river that laid the sandstone is gone. The organism that made the track is gone. The ancient ocean is gone. Induction gives geology its reach across absence.

This does not make geology weak. It makes geology honest about the kind of reasoning it performs. Historical science is not inferior because it cannot replay the past. It is inferentially rich because it reconstructs the past from multiple converging lines of evidence. A glacial interpretation may be supported by striations, till deposits, landforms, regional patterns, isotope data, and stratigraphic relationships. No single trace bears the whole burden. Induction in historical science often works through consilience, the coming together of independent clues.

Gould’s essay uses this philosophical background to clarify methodological uniformitarianism. The principle is necessary not because geology has a sacred slogan, but because empirical reasoning itself requires stability. If laws can change arbitrarily from one age to another, induction collapses. If ancient causes need not resemble any lawful relations known from the present, historical science dissolves into storytelling.

But Gould is careful not to overstate what induction can do. The assumption of invariant law does not prove that rates were uniform. It does not prove that ancient conditions resembled modern ones. It does not prove that the present contains analogues for every past state. It simply makes inference possible. This is the central difference between methodological and substantive uniformitarianism. Methodological uniformitarianism is about the reliability of law; substantive uniformitarianism is about the sameness of historical tempo or condition.

A useful way to phrase this distinction is: induction gives science a bridge, not a map. It lets us cross from present evidence toward past explanation, but it does not tell us in advance what landscape we will find on the other side. The bridge is lawful continuity. The landscape may be strange.

That strangeness matters. Ancient Earth included atmospheres, oceans, climates, continental arrangements, and biological communities unlike those of the present. If we mistake methodological uniformitarianism for substantive uniformitarianism, we may use induction too narrowly. We may say, “Because this is what we observe now, this is what must have happened then.” Gould resists that. He wants induction disciplined by law, not imprisoned by familiarity.

The post should also consider the philosophical discomfort Gould leaves unresolved. If induction cannot be proven, does science rest on faith? Not exactly, or at least not faith in the theological sense. Science rests on methodological commitment, practical success, and self-correction. It assumes enough regularity to investigate the world, and that assumption has been extraordinarily fruitful. But Gould’s point is that methodological uniformitarianism is not a special geologic discovery. It is the ordinary risk of empirical reasoning.

This humility is valuable. Science is powerful, but it does not float above philosophy. It depends on assumptions about regularity, causation, evidence, and inference. These assumptions are not arbitrary, but neither are they proven from nowhere. They are justified through practice, coherence, and the success of inquiry. Geology, with its vanished events and ancient traces, makes this structure unusually visible.

Gould’s discussion of Hume also prevents methodological uniformitarianism from becoming triumphalist. If we say “natural laws are invariant,” we are not announcing an absolute metaphysical proof. We are naming the working condition under which science proceeds. The alternative is not a better science, but no stable science at all. If nature’s laws change unpredictably, evidence loses its memory.

The beauty of geology is that it turns induction into a craft. A geologist stands before an outcrop and reads sequence, interruption, deformation, exposure, burial, and time. The reasoning may begin with present processes, but it is not mere comparison. It is an art of lawful reconstruction. Hume’s problem hums in the background, but the fieldwork continues, boot soles on shale, hand lens at the ready.

Gould’s article reminds us that every ancient rock is also a philosophical object. To interpret it is to trust that the world has enough continuity for traces to bear meaning. That trust is methodological uniformitarianism, stripped of its old disciplinary costume. It does not say the past was like the present in pace. It says the past is not sealed away from reason. The rocks are ancient, but the logic that lets them speak belongs to science itself.