Levels of Analysis

Complex information-processing systems can be understood or analyzed in terms of their function (what the system does and why), the processes and operations that achieve that function (how it does it), or the physical hardware that performs those processes and operations (how those processes and operations are implemented). These different perspectives are considered to be distinct levels of analysis and can be applied to the mind/brain or their subsystems. For example, choosing which breakfast cereal to buy can be understood from an abstract information-processing perspective as selecting from a set of options subject to perceived costs and benefits. The same choice can also be described in terms of processes that generate and evaluate options to produce a decision. In turn, those processes can be described in terms of the patterns of activity within neural circuits that transform sensory input into a specific response. Although all three perspectives capture the target phenomenon (in this case, choosing a cereal), they do so using distinct units of analysis and characteristic “languages”: at the functional level, abstract mathematical descriptions of goals and constraints; at the process and operational level, descriptions of information-processing operations and representational formats; and at the physical hardware level, descriptions couched in terms of neural and other physical mechanisms. This multilevel approach is useful because different levels are appropriate for answering different types of questions and because analyses at individual levels can often proceed, at least initially, without a fully worked-out account at the other levels.