Cognitive Psychology as counterrevolution

When you think of a college campus in the 1960’s, you probably imagine something between Woodstock and the March on Washington for Jobs and Freedom, occasionally interrupted by lectures and seminars. But it’s important to remember that students’ revolutionary energy was directed not only at the state — solidarity with the civil rights movement, activism against US intervention in Southeast Asia, and so on — but also against the institution of the university itself. The university was a rich site of struggle in large part because it was, in its normal functioning, complicit with so much of what campus radicals were fighting against: the reproduction of inequality, the development of new weapons and strategy for empire-building, and the automation and scaling up of processes designed to integrate its graduates into the very system they were protesting. 

Huey Newton, speaking at Boston College in 1970 pulls on one part of this thread:

In this country the Black Panther Party, taking careful note of the dialectical method, taking careful note of the social trends and the ever‐changing nature of things, sees that while the lumpen proletarians are the minority and the proletarians are the majority, technology is developing at such a rapid rate that automation will progress to cybernation, and cybernation probably to technocracy. As I came into town I saw MIT over the way. If the ruling circle remains in power it seems to me that capitalists will continue to develop their technological machinery because they are not interested in the people. Therefore, I expect from them the logic that they have always followed: to make as much money as possible, and pay the people as little as possible ‐ until the people demand more, and finally demand their heads. If revolution does not occur almost immediately, and I say almost immediately because technology is making leaps (it made a leap all the way to the moon), and if the ruling circle remains in power the proletarian working class will definitely be on the decline because they will be unemployables and therefore swell the ranks of the lumpens, who are the present unemployables. Every worker is in jeopardy because of the ruling circle, which is why we say that the lumpen proletarians have the potential for revolution, will probably carry out the revolution, and in the near future will be the popular majority. Of course, I would not like to see more of my people unemployed or become unemployables, but being objective, because we’re dialectical materialists, we must acknowledge the facts.

For an audience that came to listen to the Minister of Defense for the Black Panther Party, the seeming non-sequitur about seeing MIT on the way to campus would have had a clear implication: its position as one of the most active centers of military research made it a flashpoint for anti-war organizing and protest.

Of course MIT was not the only place where such research – and resistance to it – was going on. At Penn, some of the most visible and comprehensive organizing was against research into chemical and biological weapons. An undergrad working at the bookstore noted that a group called the Institute for Cooperative Research (ICR) was ordering rather a lot of books about diseases of rice crops and Vietnamese politics, and got his history professor involved. Protest by student groups and faculty alike brought national attention to the research, and as a result, the contracts were moved to a quasi-independent research institute that allowed administrators to claim they had moved the research "off campus," and allowed organizers to claim victory despite having done little to stop the development of chemical and biological warfare agents.

But the ICR was not Penn's only source of military funding at the time. The Office of Naval Research funded a wide range of activities at Penn, including the writing of Ulric "Dick" Neisser's foundational textbook Cognitive Psychology. It is hard to overstate the degree to which this textbook set the standard for the teaching of cognitive psychology. Today, most introductory texts in the area follow its general plan, and cite many of the same research papers. Aside from a chapter on visual imagery, about which I'll have more to say below, a cognitive psychologist designing an introductory course today could use much of the material as is, supplemented with findings from the half century or so since its initial publication. Viewed from this perspective, it is a very normal book, that played a key role in launching a period of normal science. Reading Cognitive Psychology in the context of the campus radicalism of its time, however, gives a very different view.

First let's consider Chapter 2, which covers “Iconic Storage and Verbal Encoding,” and traces out some fairly fundamental, and quite general results described in an “information processing” framework that is very much modeled on the modular design of contemporary computers. The first section establishes that visual information is first held in an “iconic store,” a kind of high-fidelity internal memory system whose information capacity and duration could be quantified with clever experiments pioneered by George Sperling.

In a typical experiment investigating the iconic store, participants are presented with three rows of three unique letters for a short time (50 milliseconds), and asked to report as many as they can. Normally, people can only report only 4 or 5 letters, suggesting that their memory for what was seen is limited. Sperling’s innovation was the partial report procedure, in which the visual stimulus is followed with one of three tones, indicating which row to report, with the high tone corresponding to the top row, the middle tone corresponding to the middle row, and the low tone corresponding to the bottom. These partial reports were consistently near-perfect, indicating that the bottleneck in the earlier experiments had not been the capacity of the iconic store itself — if participants don’t know which row they’re going to report beforehand, they must have stored all nine letters in order to perfectly remember a randomly selected set of three — but its duration. The duration of the iconic store could then be measured by carefully manipulating the time between the stimulus and the response cue.

It's not hard to see how this research stemmed from concerns about the integration of human operators into technological systems then under development for military use. A whole panoply of technologies designed to detect fluctuations in radiomagnetic fields, chemical deposits, and the like were being developed as a kind of extended sensory apparatus. At some point this information needed to be handled by an operator, and passed along up the chain of command to an officer in charge of making both real-time and long-term strategic decisions. The goal of experiments like Sperling's was to provide information about the technical capabilities of the operator as a component of this system. How much information can he handle? How long does it need to be displayed? How fast can it be recognized? and so forth. Arguably, the bureaucratic structure implied here is just as important as the computer as a metaphor for understanding the mind.

The chapter goes on to describe some results from studies using "masking," i.e., rapidly replacing one stimulus with another, demonstrating that information that enters the iconic store is processed even when participants are not aware of it. It then deals with how stimuli that we are aware of are transferred from the brief, literal "store" to a more durable, symbolic "memory" system not too different from the random access memory of a contemporary computer. In one of the studies reviewed in this section, Pollack and Johnson (1965), working intramurally at an Air Force base, had participants practice for 40 days to test the efficacy of a coding system for increasing short term memory span.

What's remarkable is how little any of this has to do with how people typically perceive and understand the world around us. Neisser himself would acknowledge this later in a much less influential, and frankly somewhat weird little book called, pointedly, Cognition and Reality. It turns out that studying the mind from the perspective of someone designing a system of sensing and information processing leaves out rather about how people actually think. Interestingly, his turn toward a more ecological approach was inspired by his interactions with James Gibson, whose focus on perception in the context of dynamic interactions with a constantly changing environment was in turn inspired by his work training Air Force pilots during WWII.

The one part of Cognitive Psychology that stands out as unlikely to be included in a 21st century course on cognitive psychology is Chapter 6. Throughout the text there are scattered references to phenomena related to altered states of consciousness – the phenomenon of masking, for example, was of interest in part for its potential application to subliminal messaging. But in Chapter 6, Neisser really gets into some weird shit, including several lengthy block quotes from Aldous Huxley’s Doors of Perception.

Neisser refuses to draw a bright line distinction between the type visual imagery that most people would engage in if they were asked, say, to describe a familiar person or location, and images resulting from dreams, hallucinations (both as psychotic symptoms and drug-induced), sensory deprivation, sleep deprivation, and hypnotic trances. Jolly West, whose edited volume on hallucinations is cited throughout, was famous for having given a lethal dose of LSD to an elephant, and was at the time running a crash pad in Haight Ashbury where experiments were done on unsuspecting hippies. Martin Orne, whose work on demand characteristics has been influential well beyond the psychedelic context in which he first conceived of it, is one of the most-cited authors in the book, and also the PI on the ONR contract that funded both Cognitive Psychology and a series of experiments designed to investigate the feasibility of using hypnosis in interrogations.

Hypnosis and hallucinations, effects of sensory deprivation, etc., are weird and interesting in their own right, and plenty of people were experimenting with them before the CIA took interest. It’s not entirely implausible, as Orne implied in an interview with the APA Monitor, that the researchers involved would have been pursuing that work, no matter who funded it. Further, his position on hypnosis as an interrogation technique, as made clear in his contribution to a collection called The Manipulation of Human Behavior commissioned by the Air Force, was that it would be highly inefficient, in particular because of the issue of demand characteristics.

Indeed, most citations of Orne's work throughout Cognitive Psychology extend the notion of demand characteristics Orne developed in the context of research on hypnosis and altered states of consciousness to more mundane experiments at the core of cognitive psychology research. I first learned about the notion of demand characteristics in a course on experimental design that had nothing at all to do with hypnosis. It turns out to be a useful concept when designing any experiment in which you're hoping to observe an automatic or unconscious process. Say, for example, you're interested in priming phenomena, where responses to a target word, say, "book," are more efficient after recently encountering a related "prime" word, say, "page," than when an unrelated prime, e.g., "frog" is presented. If the goal is to say something about the organization of memory for the meanings of words, or its role in processing text or speech, it's important to rule out the possibility that participants are aware that some of the words are related, and that this must be what the experiment is about. Like Orne's hypnotic subjects, the reasoning goes, they may end up giving you the results you want, due to conscious processes rather than unconscious ones.

Excursions into the weird science of mind control aside, Cognitive Psychology is primarily a catalogue of research done in very particular environments – a person sitting at a computer, taking stimuli as input and producing key presses as output – and with largely with very particular goals: characterizing the memory and processing capacity of human computer operators in terms borrowed from computer technology. It seems inevitable that the computer would replace the brass instruments of previous reductionist research programs in the study of the mind. Similarly, the idea of the mind as a kind of computer may well have arisen without the explicit entanglement of research on minds and the development of computer systems. Titchener, under the influence of contemporary advances in physics and chemistry, thought of his work as attempting to create a kind of periodic table of experience. Faced with a machine that can store and process information, someone would inevitably have come up with a theory using it as a metaphor for some aspect of cognition. Nonetheless it is hard to imagine that the first cognitive psychology textbook would not have looked different if the material basis for research into the mind had not been so firmly rooted in the development of technology for war.