As physical entities that translate symbols into physical actions, computers offer insights into the nature of meaning and agency. Physical symbol systems, generically known as agents, link abstractions to material actions. The meaning of a symbol is defined as the physical actions an agent takes when the symbol is encountered. An agent has autonomy when it has the power to select actions based on internal decision processes. Autonomy offers a partial escape from constraints imposed by direct physical influences such as gravity and the transfer of momentum. Swimming upstream is an example. Symbols are names that can designate other entities. It appears difficult to explain the use of names and symbols in terms of more primitive functionality. The ability to use names and symbols, that is, symbol grounding, may be a fundamental cognitive building block. The standard understanding of causality—wiggling X results in Y wiggling—applies to both physical causes (e.g., one billiard ball hitting another) and symbolic causes (e.g., a traffic light changing color). Because symbols are abstract, they cannot produce direct physical effects. For a symbol to be a cause requires that the affected entity determine its own response. This is called autonomous causality. This analysis of meaning and autonomy offers new perspectives on free will.