Diamond argues environmental determinism—geography determines broad historical outcomes (Eurasia's domesticables enabled state formation; Americas lacked them). Yet history is full of contingency: Japan adopted firearms then rejected them; China abandoned oceangoing ships; societies make choices that reshape their futures. These two observations seem incompatible: if geography determines, where is room for choice? If choice matters, how is geography determinative? The resolution: geography determines possibility, not outcome. Geography makes some outcomes likely (Eurasia's domesticables enable states); it doesn't make them certain (Japan could have chosen states differently). Within geographic constraints, contingency operates. A continental zoom-out reveals determinism (Eurasia developed states because geography enabled them). A local zoom-in reveals contingency (Japan choosing which states technologies to adopt). Both are true simultaneously.1
Geography as Constraint, Not Determinant
Geography sets the possibility space—the range of outcomes geographically possible. Societies with domesticable animals can develop agriculture-based states; societies without cannot. This is hard constraint. But within that constraint, multiple outcomes are possible. A society with domesticable animals could develop states (Eurasia did), could reject states (some Pacific societies with pigs remained non-hierarchical), could adopt states then abandon them (Japan), could develop unique state forms (Inca without horses). Geography constrains but doesn't determine.
Degrees of Freedom Within Constraint
The more severe the geographic constraint, the fewer degrees of freedom. A tiny atoll (Pitcairn, 2 square miles) has no freedom to develop complex states—population size is hard constraint. The ocean makes escape impossible. That's low degrees of freedom. A large continent with diverse resources (Eurasia) has high degrees of freedom—multiple domesticables available, multiple development paths possible. The same principle applies everywhere: stronger constraints = fewer options; weaker constraints = more options. Geography determines the shape of the possibility space, not the point within it where societies land.
Proximate vs. Ultimate Revisited
Geography determines ultimate causes (what's possible); culture and choice determine proximate outcomes (what actually happens). A continent can support state development (ultimate), but a specific society must choose whether to develop states (proximate). The ultimate constraint sets the playing field; the proximate choice determines the game.1
At continental scale, the pattern is deterministic: every continent with agriculture developed hierarchical states. Eurasia had domesticables → agriculture → states. Mesoamerica had maize → agriculture → states. Americas without Eurasian domesticables didn't develop state-scale societies until post-contact. The pattern is consistent across cases. Zoom out to continental timescale, and geography seems to determine outcomes inevitably.1
But: Not all continents with agriculture developed states. Some African societies maintained egalitarian organization despite being agriculturalists in large populations. So even at continental scale, the pattern is probabilistic (states more likely with agriculture) not deterministic (states inevitable with agriculture).
Within Eurasia, which specific states developed? Which crops became dominant? Which animals were prioritized? These were contingent choices. The Roman Empire vs. European feudalism—both European, both developed from similar domesticables, but very different outcomes. Japan vs. China—different state forms, different paths, despite similar geographic advantages. Islamic empires vs. Christian kingdoms—different religions, different political structures, despite overlapping geography. At regional scale, outcomes are not determined by geography; they're shaped by contingent historical events.
When societies interact (trade, warfare, conquest), contingency is constrained by geographic realities. Pizarro with guns could conquer Inca because Eurasian geography had produced guns; Inca couldn't have resisted with pre-Columbian weapons. Geography didn't require conquest, but it made conquest possible and conquest-resistance difficult. This is contingency within constraint: the conquest wasn't determined (Pizarro could have died, Inca could have allied more effectively, contingent events could have changed outcome), but geography made Spanish victory likelier than Inca victory.
Tension 1: Is "Choice" Real If Geography Constrains It?
Japan "chose" to restrict firearms. But is this choice real if it was made possible by island geography (difficult to invade, enabling deliberate rejection of firearms for 200 years)? Would a mainland society be able to make the same choice? Probably not—invasion risk would make firearms restriction suicidal. So is Japan's choice free or is it geographically enabled? The tension: where does constraint end and choice begin?
Tension 2: How Severe Must Constraint Be to Eliminate Freedom?
Tiny atoll societies (Pitcairn) have no choice about state formation—population size makes it impossible. What's the threshold where geography's constraint becomes so severe that choice vanishes? 50,000 people? 5,000 people? The answer is fuzzy, which means the boundary between determinism and contingency is fuzzy.
Tension 3: Can Contingency Overcome Determinism?
If geography usually determines outcomes (but sometimes doesn't), is geography determinative at all? Or is geography just a probabilistic factor among many? The answer: geography is the strongest deterministic factor, but not the only factor. Other factors (culture, leadership, chance events) can override geography's push, but doing so is harder when geography pushes strongly.
Diamond's framework oscillates between determinism and contingency without fully resolving the tension. He claims geography determines, but acknowledges Japan's firearms rejection, China's ocean-going ship abandonment, and other contingencies that complicate determinism. His answer (implicit) is that geography determines broad patterns (all major domestication centers develop states), but contingency shapes specific outcomes (which states, which societies, which technologies). This is a sophisticated position: determinism + contingency operating at different scales. But it requires accepting that geography is powerful without being absolute—a probabilistic framework rather than strict determinism.1
Degrees of Freedom and Soft Determinism — Philosophy distinguishes hard determinism (no choice possible), libertarian free will (all choices free), and soft determinism (choices are free within constraints). Geography enables soft determinism: your freedom to choose is constrained by circumstances (geography, history, institutions), but within those constraints, you have genuine choice. This maps to history: geography constrains possibility space, but within that space, cultural and individual choice operates freely. The insight: determinism and freedom are not binary opposites; they operate at different levels simultaneously. You're determined by geography (constrained to certain options) and free within geography (can choose among options). Both are true.
Nested Constraints and Degrees of Freedom in Complex Systems — Complex systems have nested constraints: each level constrains the next. Physics constrains chemistry; chemistry constrains biology; biology constrains psychology; psychology constrains sociology; sociology constrains history. At each level, lower levels set constraints (hard walls) but don't determine outcomes (possibilities within walls). A psychologist can't explain human behavior purely through physics (too many intermediate levels) nor ignore physics completely (it sets ultimate constraints). Same with history: geography sets ultimate constraints, but multiple levels of contingency operate between geography and observed outcomes. The insight: causation in complex systems is multilevel: ultimate constraints + intermediate contingencies + proximate choices.
The Sharpest Implication
If geography determines possibility but not outcome, then both geographic determinism and human agency are real. This is uncomfortable for extremists: geographic determinists want to claim geography determines everything; radical libertarians want to claim humans are free from all constraint. Reality is messier: geography provides a playing field that makes some games far more likely than others, but within that field, contingency operates. This means: (1) we can't blame people for their societies' technological lag if geography constrained development (environmental determinism is true), AND (2) we can't absolve people of responsibility for their choices within geographic constraints (agency is real). This both absolves guilt (geographic constraint explains outcomes) and demands responsibility (within constraints, choices matter). The uncomfortable implication: we are neither determined by geography nor free from it; we are constrained agents making real choices within real limits.
Generative Questions