A person's command over food is their entitlement set: the bundles of goods they can legally acquire through production, trade, labor, or transfer given their endowments and the prevailing exchange conditions. Starvation is a relationship between people and food determined by this entitlement, so people can starve while food sits in the market, legally beyond their reach.
Famines can and do occur with no significant decline in aggregate food availability per head, and sometimes amid peak output. The lethal change is a collapse in some group's entitlement, not in the total stock of food.
Within a market economy, trace whose ability to command food collapses by tracking exchange entitlement: the rate at which a group's endowment (labor, livestock, craft output) trades for food. A famine typically devastates specific occupation groups whose terms of trade move sharply against them, rather than the population uniformly.
In the 1943 Bengal famine, roughly three million died even though food output was not exceptionally low; wartime urban demand and inflation bid up rice prices while rural laborers' wages lagged. Their exchange entitlement collapsed, so the famine fell on particular rural groups during what was in some sectors an economic boom.
Rising food prices are read not as a scarcity thermometer but as a redistribution mechanism: when one group's purchasing power surges, it can pull food away from poorer groups whose incomes do not keep pace. A price movement that efficiently clears a market can simultaneously be the direct cause of mass starvation among those priced out.
Because starvation stems from entitlement failure rather than sheer physical shortage, the effective remedy is to restore people's command over food, for instance through employment, cash, or targeted relief that regenerates purchasing power. Policy that focuses only on aggregate supply can miss the vulnerable groups entirely.