Despite a consensus that humans are limited in their capacity for cognitive effort, there has been remarkably less agreement about the nature of that limitation, especially among attention researchers in the mid-20th century. On one side, Broadbent (1957) argued that a selection filter existed early in processing; alternatively, Deutsch and Deutsch (1963), among others, maintained that the bottleneck was located later in the processing chain. Triesman’s (1964) workaround was the suggestion that the filter was a bit leakier than initially conceived. Each of these positions was well-evidenced, making it difficult to jettison one explanation for another without careful regard for context. Kahneman (1973) posited that attention was tantamount to cognitive effort and argued that the amount of effort available for investment in mental work (i.e., capacity limits) fluctuated depending on variables such as task demands, environmental conditions, and physiological arousal. The idea that a general pool of resources could be flexibly mobilized was a strength of the model in that it provided a way to unite previously-incompatible evidence; ironically, this was also its weakness as Kahneman's theory proved unfalsifiable and eventually fell out of favor. However, researchers still lean on its descriptive utility to illustrate cognitive phenomena (e.g., dual process, ironic process, multiple resource theories).
Kahneman, D. (1973) Attention and Effort. Englewood Cliffs, NJ: Prentice-Hall.