The Lazy Brain

One of the things that I have always found both interesting and irksome about a great deal of philosophy, be it Cartesian epistemology, Husserlian phenomenology, or contemporary philosophy of mind, is the role that consciousness plays. In a great deal of philosophy, thought and consciousness are synonymous, making the phrase "conscious thought" redundant. Sure, there are philosophers for whom consciousness is not (always) focal, such as Nietzsche, Bergson, Heidegger, or Merleau-Ponty (and maybe even Husserl), but even today, that breed of philosopher is in the minority. Philosophers aren't alone in this misguided (in my view) focus. Psychologists tend to operate on the assumption, as John Bargh put it, "that people are consciously and systematically processing incoming information in order to construe and interpret their world and to plan and engage in courses of action." I don't mean to imply that the problems of consciousness (e.g., those that are still at the fore of most philosophy of mind) are not important. However, most of what we do, and I would even say most of what we "think," is largely unconscious, or at least not under conscious control. While this may seem like a negative, it's actually a pretty good thing. Conscious thought is effortful. Not only does it take up cognitive resources (e.g., attention) which places limits on the number and complexity of the tasks that we can consciously perform, but it also takes up a whole lot of physical resources. This makes conscious thought fairly inefficient. It's fortunate, then, that our brains are designed, as a friend of mine often says, to avoid "thinking" (where thinking is synonymous with consciousness). Most of what goes on in the brain is automatic, relatively effortless, and thus uses a much smaller portion of our cognitive and physical resources. Knowledge, in the traditional sense, is the exception, rather than the rule in the brain.

Automaticity, therefore, is the name of the cognitive game. Some of the things that we do are fairly automatic from birth, and much of what isn't quickly becomes so. Some of the more prominent automatic aspects of cognition include virtually everything about perception, motor programs, solving familiar problems (think of arithmetic), speech, reading, writing, and expertise of various sorts (be it physical or intellectual), just to name a few. In fact, virtually everything we do during the course of the day is automatic and, at least in large part, unconscious. With this in mind (conscious mind!), I thought (consciously) that I should celebrate automaticity, by posting about it.

First of all, we should get out of the way what consciousness is. For the purposes of this post, I'll again quote Bargh (from the article linked above) to define consciousness. He writes:

The defining features of what we are referring to as a conscious process... are mental acts of which we are aware, that we intend (i.e., that we start by an act of will), that require effort, and that we can control (i.e., we can stop them and go on to something else if we choose). [emphasis in the original]

If we treat awareness, intentional, effortfol, and controlable as the necessary and sufficient features of conscious processes, than any processes to which some or all of these don't apply can be considered unconscious or automatic. Now that we've got the definitions out of the way, we need a theory. The most prominent theory of automaticity, these days, is Gordan Logan's instance-based theory of automaticity. Some of the details of this theory are a bit complex for a blog post, and I'm afraid that whittling it down to its bare essentials may be beyond my capabilities, but I'm going to try.

The old view of automaticity, as expressed by Logan (follow the link), involved the following beliefs about automaticity:

Automatic processing is fast and effortless because it is not subject to attentional limitations. It is autonomous, obligatory, or uncontrollable because attentional control is exerted by allocating capacity; a process that does not require capacity cannot be controlled by allocating capacity. Finally, it is unavailable to consciousness because attention is the mechanism of consciousness and only those things that are attended are available to consciousness. (p. 492-3)

Logan and others have called all of these assumptions into question. In place of the traditional view, Logan has offered a new theory designed to explain automaticity without these assumptions. Logan's theory begins with three different assumptions. The first is that encoding is obligatory. When we attend to a stimulus, we automatically encode it into memory. The second assumption is that retrieval is also obligatory. When we attend to a stimulus, we automatically retrieve associated memories. The third assumption, which follows from the first, is that every time we attend to a stimulus, we encode it, and thus every attended instance of a stimulus is encoded (this is sometimes called the exemplar theory of memory, because it states that all attended exemplars are encoded). Unless you want to read Logan's paper, you're going to have to take my word for it that these assumptions have empirical support (though they are sometimes questioned).

These assumptions provide the beginning of a theory of automaticity. Within this theory, the processing of tasks is effortful at first, and utilizes domain-general processes (Logan calls them "algorithms") of knowledge-retrieval. After repeated exposures to a particular task, particular solutions are encoded, and retrieval becomes automatic and task (or domain) specific¹. As we learn a task, our automatic performance in that task becomes more and more efficient because it is easier to automatically retrieve past instances of successful solutions. As Logan puts it:

Memory becomes stronger because each experience lays down a separate trace that may be recruited at the time of retrieval. (p. 494)

This, according to Logan, is what constitutes the path of learning from the effortful to the automatic.

Because automatic and effortful processes differ in type, they are affected by different things. For instance, effortful processes can be influenced by attentional resources. Thus, performing more than one task at a time may hurt performance utilizing effortful processes. For the most part, automatic processes are not influenced by attentional resources (though attention does impose some limits). Instead, they are influenced by the sorts of things that affect memory retrieval, such as relevant differences in retrieval context (if you can't recognize an instance of a specific schema, you can't retrieve the schema), and concurrent retrieval processes. Since learning is a continuum from effort to automaticity, the degree to which different factors (e.g., attention and contextual factors) influence the performance of a task will differ depending on how many traces have been encoding, and thus how reliant task performance is on general, effortful algorithms.

As you can probably see, the picture of automatic processes that this theory paints forms a nice contrast with our definition of conscious processes. Processes become more and more automatic as they rely more on uncontrolable, unintentional, effortless, memory retrieval that generally occurs below the level of awareness. However, as the existence of a continuum within the theory implies, not all automatic processes exhibit all of these features. Some are at least partially intentional, for instance. This is the case in skill acquisition, in which we intentionally initiate processes that, after sufficient learning has been acheived, then continue automatically. We may even be aware that we are performing the task (though we are not always aware - imagine all of the times that you have driven from work to home and, upon ariving, realized that you have no memory how you got from there to here), but you are not aware of the task-specific processes that you are carying out.

This transition in skill acquisition also illustrates another interesting difference between effortful and automatic processes. Conscious, intentional intervention, once an automatic process has begun, can often result in less efficient task performance. Anyone who has ever played a sport that required repetitive and overlearned motor tasks has probably exeprienced this. In basketball, for example, free throw shooting is practiced for hours on end to make all of the movements automatic. However, even the most skilled player occasionally consciously interrupts the automatic motor programs, or "thinks about" the shot, and when he or she does, the result is usually a missed free throw. This is because the conscious intervention makes the task effortful, and therefore less reliant on the retrieval of successful motor programs from memory. As Logan's theory states, effortful processes are generally less officient than learned automatic ones.

So far, I've talked only about automaticity in learning. While learning accounts for much of our daily lives, from driving to reading to switching on lights and even performing complex motor and intellectual tasks (think of expert chess players), there are other automatic processes that influence our action. For instance, perceptual processing, from the perception of color, distance, and location, to object-recognition, occurs largely below the level of awareness and without intention or conscious effort. The automaticity of perception allows the environment to influence our cognitive processes without our conscious intention or awareness. As Bargh (see link above) writes:

[B]ecause perceptual activity is largely automatic and not under conscious or intentional control (the orange on the desk cannot be perceived as purple through an act of will), perception is the route by which the environment directly causes mental activity–specifically, the activation of internal representations of the outside world. The activated contents of the mind are not only those in the stream of consciousness but also include representations of currently present objects, events, behavior of others, and so on. (p. 5)

Bargh, a social psychologist, is interested in the ways in which these automatic perceptual processes influence behavioral tendendies without our awareness. He describes a great deal of research in which participants are primed with some perceptual or conceptual schema, which subsequently influences their behavior without their awareness. The most striking experimental example of this involved a variation on the famous Milgram shock experiments, designed by Carver et al.². In their experiments, Carver et al. first gave participants what they called a "language task," which contained words designed to create hostile impression ("hostility-related" words). In a secont, ostensibly unrelated task, participants were assigned the role of "teacher," and told to administer shocks to "pupils" (confederates) who answered questions incorrectly. The researchers found that participants who had been given the language task containing the hostile words administered more and longer shocks than participants who had not been exposed to those words³. Thus, unconscious perceptual and conceptual processes were influencing the participants' social behavior in ways of which they were not aware.

Finally, there is one other type of unconsciuos processing that deserves mentioning. Many of the ordinary, domain-general processes involved in ordinary cognitive tasks, such as memory retrieval, analogical reasoning, categorization and conceptual processing, induction, counterfactual reasoning, and most other everyday cognitive functions, occur below the level of awareness. While many of these processes may be (but are not always) initiated intentionally, and the output of the processes may be monitored along the way, the processes themselves are largely automatic, unintentional, and below the level of awareness. This fact is, ultimately, what makes studying cognition so hard. If we were able to introspectively analyze what it is we are doing when, for instance, we compare apples and oranges, my job would be much, much easier. Of course, I would probably be out of a job, as well. Still, the difficulty of studying cognition only serves, along with the automatic processes described above, to illustrate the importance, and even dominance of the cognitive unconscious over conscious thought. Philosophers who haven't caught on to this would do well to take note, as much of what makes us human, what determines or delimits our possible behaviors, decisions, and beliefs, is simply not conscious. Consciousness is a valuable tool, designed, as Antonio Damasio likes to say, to provide a constant bridge between the external and internal worlds, but the external and the unconscious internal worlds comprise most of the traffic.

¹ In some models, this transition is characterized as being from indirect to direct access to long-term memory. E.g., Strayer, D. L., & Kramer, A. F. (1990). An Analysis of Memory-Based Theories of Automaticity. Journal of Experimental Psychology: Learning, Memory and Cognition, 16, 29 -34.
² Carver, C. S., Ganellen, R. J., Froming, W. J. & Chambers, W. (1983). Modeling: An analysis in terms of category accessibility. Journal of Experimental Social Psychology, 19, 403-421.
³ I described this experiment, instead of the many others designed to show similar effects, because it's the most striking. However, it suffers from many methodological problems, as many of you may have noticed. Others have performed similar experiments in more methodologically sound ways, and acheived similar results, though. See e.g., Bargh, J. A., Chen, M. & Burrows, L. (1996). Automaticity of social behavior: Direct effects of trait construct and stereotype activation on action. , Journal of Personality and Social Psychology, 71, 230-244. Interestingly, theoretical explanations for these effects have not been easy to come by. We know they're automatic and unconscious, but we don't know what's going on.