In James' time, the only method available to study attention was introspection. Very little progress was made in quantifying the study of attention, though it was considered a major field of intellectual inquiry by such diverse authors as Sigmund Freud, Walter Benjamin, and Max Nordau. One major debate in this period was whether it was possible to attend to two things at once (split attention). Walter Benjamin described this experience as "reception in a state of distraction." Some thinkers felt that they were unable to do so, and other thinkers felt that they could. Without experiments, it was impossible to demonstrate which belief was correct.
From the 1920s to the 1950s, the field of attention was relatively inactive as the dominant psychological paradigm at the time was behaviorism. Behaviorism was rooted in positivism, an epistemology that rejects the study of and assumptions about processes that cannot be observed directly such as cognitive processes.
In the 1950s, research psychologists renewed their interest in attention when the dominant epistomology shifted from positivism to realism during what has come to be known as the "cognitive revolution" (Harré, 2002). The cognitive revolution admitted unobservable cognitive processes like attention as legitimate objects of scientific study.
Colin Cherry and Donald Broadbent, among others, performed experiments on dichotic listening. In a typical experiment, subjects would listen to two streams of words in different ears of a set of headphones, and selectively attend to one stream. After the task, the experimenter would ask the subjects questions about the content of the unattended stream.
During this period, the major debate was between early-selection models and late-selection models. In the early selection models, attention shuts down processing in the unattended ear before the mind can analyze its semantic content. In the late selection models, the content in both ears is analyzed semantically, but the words in the unattended ear cannot access consciousness. This debate has still not been resolved.
In the 1960s, Anne Treisman began developing the highly influential Feature integration theory (first published under this in a paper with G. Gelade in 1980). According to this model, attention is responsible for binding different features into consciously experienced wholes. Although this model has received much criticism, it is still widely accepted or held up with modifications as in Jeremy Wolfe's visual search paradigm.
In the 1960s, Robert Wurtz at the National Institutes of Health began recording electrical signals from the brains of macaques who were trained to perform attentional tasks. These experiments showed for the first time that there was a direct neural correlate of a mental process (namely, enhanced firing in the superior colliculus).
In the 1990s, neuroscientists began using fMRI to image the brain in attentive tasks. The results of these experiments have shown a broad agreement with the psychophysical and monkey literature.