At Cajamarca in 1532, Pizarro's 168 men surrounded an Inca emperor with 80,000 soldiers. The emperor Atahualpa was captured. The Inca Empire collapsed. Why? Proximate cause: guns, steel swords, horses, disease, writing (for communication), political organization (coordinated command structure). Pizarro had superior technology and organization. Cortés conquered the Aztecs with 600 men for the same proximate reasons. These causes are proximate—close in time and space to the outcome, immediately visible, causally direct.1
But why did Pizarro have guns, steel, horses, disease? Why did the Inca not have them? That requires ultimate causation: the structural, deep-time reasons that produced those proximate advantages. Ultimate cause: Eurasia had domesticable horses (none in Americas). Eurasia had iron ore deposits accessible with bronze tools. Eurasia had domesticated animals creating epidemic diseases selecting for resistant populations. Eurasia had 13,000 years of agricultural development producing surplus for specialization enabling metallurgy and weaponry. The Inca had 4,500 years of agriculture and lacked domesticable large animals, lacking epidemic diseases (so no selective pressure for immunity), lacking iron sources they could work with available technology. The proximate factors (guns vs. spears) were symptoms of ultimate factors (geography, time available for development).1
This distinction—between the immediate cause visible to participants and the deep structural cause explaining why that immediate cause existed—is what transforms history from narrative into science. It's the method allowing explanation rather than description.
Proximate causation describes the immediate, observable factors directly producing an outcome. At the tactical level, Pizarro's victory came from: guns frightened horses unaccustomed to noise, cavalry could charge through infantry, steel armor resisted obsidian swords, disease killed epidemic-naive populations, Spanish communication via writing vs. Inca runners. These are proximate—visible to participants, explicable mechanically, determinative of immediate outcomes.
Ultimate causation describes the structural, deep-time factors explaining why the proximate factors existed. Why did Spain have guns? Because it had: (1) domesticated iron-working animals (horses for plowing, food for workers); (2) endemic disease environment selecting for resistance in population; (3) 10,000 years of agricultural development producing surplus for specialization enabling weapon innovation; (4) writing systems enabling knowledge transmission and military organization; (5) geographic access to iron ore deposits requiring bronze technology to extract and work. None of these factors were chosen strategically—they resulted from geographic distribution of domesticable species and time available for technological accumulation.
The distinction becomes testable when you ask: could the Inca have defeated Pizarro with different proximate factors but identical ultimate conditions? Probably yes—better Inca metallurgy or larger standing armies might have resisted. Could Pizarro have conquered with identical proximate factors but different ultimate conditions? Unclear—if the Americas had horses, the Inca would have had horses too, changing the tactical balance entirely. This reveals the distinction: proximate causes determine immediate outcomes, but ultimate causes determine available options. Geography constrained which proximate factors were possible. Once the constraint existed, immediate causes mattered.
Evidence for ultimate causation as the more fundamental level comes from comparative analysis. Look at every conquest of Eurasia-into-Americas/Australia by 1500 CE: Spanish conquistadors, Portuguese traders, later English/French colonists. All shared proximate factors: steel, guns, writing, disease. All succeeded at conquest. The uniformity of success despite different conquistadors/companies suggests proximate advantages were powerful enough to determine outcomes. But those proximate advantages appeared everywhere because they traced to identical ultimate causes (Eurasian geography). If ultimate causes had been different (Americas had horses), proximate factors would have been different, and conquests would have failed. This is why ultimate causation is more fundamental: it explains why certain proximate factors existed in the first place.1
Diamond treats history like evolutionary biology, not like narrative. In evolutionary biology, you can't run experiments (can't evolve humans under controlled conditions). But you can extract causal conclusions by comparing natural experiments—different species in different environments, all with identical time depths, showing how environment shapes outcome. History works the same way: you can't run experiments (can't re-run 1532 with different technologies). But you can compare natural experiments—different continents with different environments, developing for similar timeframes, showing how environment shapes outcomes.1
This is what makes the proximate/ultimate distinction testable rather than merely narrative. You can ask: do all continents show similar patterns? Did Eurasian dominance result from similar ultimate causes operating independently? Five domestication centers (Southwest Asia, China, Mesoamerica, Andes, Eastern North America) all show the same cascade: domestication → surplus → specialization → states → metallurgy. The pattern repeats across independent cases. This is the evidence for ultimate causation's power—not any single case, but the uniformity of patterns across independent cases suggesting structural mechanisms rather than contingent luck.
Tension 1: Identifying the Ultimate Level
How deep does "ultimate" go? Is geography ultimate? What about physics of orbit creating climate zones? What about cosmic accident of Earth having the species it does? At some point, causation becomes so deep (cosmic history) that it's trivial. The useful distinction isn't between proximate and ultimate per se, but between different levels of causation operating at different scales. Pizarro's tactics matter at tactical scale (whether a specific battle wins). Proximate technology matters at military scale (whether conquest succeeds). Ultimate geography matters at continental scale (whether conquest is attempted at all). The tensions: where to draw the line, and whether causes at different scales truly determine each other or merely constrain options.
Tension 2: Agency Within Ultimate Constraints
If ultimate factors are determining, where is human agency? The answer: agency operates within constraints set by ultimate factors. Pizarro made strategic choices (which route, which allies to cultivate, how to negotiate). Those choices affected immediate outcomes. But those choices were made within a context shaped by ultimate factors (he had guns because Eurasia had domesticated iron-working animals; he faced Atahualpa because the Inca had centralized power enabled by 4,500 years of agricultural surplus). Agency didn't disappear—it became constrained. The tension: how much of any outcome traces to individual choice vs. structural constraint? The book claims structure matters more at continental scale, but allows that individual choice matters at tactical scale. How these scales interact remains unclear.
Single source (Diamond), but he contains tensions: he claims ultimate causes matter (geography determines), yet acknowledges proximate factors matter (strategy, leadership, individual choices). He wants both determinism and agency. The framework requires this: geography is destiny at continental scale, contingency is real at local scale. But he doesn't fully resolve how these coexist—how can geography be destiny AND Pizarro's genius matter AND Atahualpa's negotiation strategy matter? All three claims are present but not fully integrated.
Levels of Causation in Science — The proximate/ultimate distinction maps onto philosophy of science's broader insight: causation operates at different levels simultaneously. Physics describes particles and forces. Chemistry describes molecular reactions (emerged from but not reducible to physics). Biology describes organisms and selection (emerged from but not reducible to chemistry). Psychology describes minds and behavior (emerged from but not reducible to biology). History describes civilizations and institutions (emerged from but not reducible to geography). Each level has legitimate causes; lower levels don't determine higher levels completely. You can explain human behavior using physics (atoms), psychology (goals), or sociology (institutions), and all three are true simultaneously but at different levels. History operates at the institution/civilization level; geography operates at the environmental level. Both matter, but they matter at different scales. This is the meta-insight: proximate/ultimate isn't unique to history; it's how causation works across all sciences.
Nested Constraints in Complex Systems — Ultimate factors work through constraint, not determination. Geography constrains which domesticable species exist (hard constraint—no horses in Americas, full stop). This creates constraint on agriculture possible (hard constraint—no animal plowing available). This creates constraint on population density achievable (probabilistic—could reach X density with effort, Y density easily). This creates constraint on technology likely to develop (probabilistic—metallurgy more likely given surplus, but not certain). This creates constraint on conquest outcomes likely (probabilistic—conquest more likely with steel, but not inevitable). Each level constrains the next, but each level also has degrees of freedom. This is how history remains partly determined and partly contingent. Ultimate constraints are tight at the level they operate (geography), but their effects cascade through probabilistic constraints at each subsequent level, leaving room for agency at each level.
The Sharpest Implication
If ultimate causation matters more than proximate causation at historical scales, then celebrating individual genius (Pizarro, Columbus, Napoleon) is misplaced. Their achievements were possible because geography had arranged for them to have steel, guns, and organized empires. Different individuals in the same ultimate context would have achieved similar outcomes—perhaps through different routes, different timing, but same broad pattern. This is simultaneously more damning (great men aren't the source of historical change) and more liberating (we don't need great men to change course; we need structural change). The uncomfortable implication: treating individuals as history's drivers obscures structural drivers operating at scales where individual action is less powerful. This matters for contemporary policy: if you believe Pizarro's genius caused conquest, you'll design institutions around great leaders. If you believe geography caused conquest and Pizarro was a tool of geographic advantage, you'll design institutions around structural constraint-management. The proximate/ultimate distinction determines what kind of change is actually possible.
Generative Questions
At what timescale does proximate causation dominate and ultimate causation fade? Individual actions matter at tactical scale; do they matter at civilizational scale?
If geography determines ultimate factors, what determines geography? Cosmic accident? Can that be reduced further? At what level does reduction become uninformative?
Can proximate factors reverse ultimate constraints? If Inca had somehow developed steel and guns independently (unlikely given ultimate constraints), could they have conquered Spain? Does this test whether ultimate constraints are truly determinative?