9.3 Working Memory: The Active, Conscious Mind

“To find out if you’re someone who could benefit from our Memory Improvement Seminar, press 1597362225830952170610137.”

© 2005 by RandyGlasbergen

We’ve used the concept of the short-term store to refer to the metaphorical space in which information is held for brief periods and “worked on.” As we mentioned earlier, working memory refers to the process of storing and transforming information being held in the short-term store. Working memory is the center of conscious perception and thought. This is the part of the mind that thinks, makes decisions, and controls such processes as attention and retrieval of information from long-term memory.

The most influential psychological model of working memory to date is that developed by Alan Baddeley (1986, 2006), which divides working memory into a number of separate but interacting components. The components include a phonological loop, responsible for holding verbal information; a visuospatial sketchpad, responsible for holding visual and spatial information; and a central executive, responsible for coordinating the mind’s activities and for bringing new information into working memory from the sensory and long-term stores. (Baddeley added a fourth component, the episodic buffer, to the model in 2000, but we will not discuss this here.) We have already discussed one function of the central executive—attention—and will discuss some of its other functions later in this chapter and in Chapter 10. Here we’ll focus on the phonological loop.

Verbal Working Memory: The Phonological Loop

As a test of one aspect of your own working memory, read the digits at the end of this sentence and then close your eyes and try to keep them in mind for a minute or so: 2 3 8 0 4 9 7. What did you do to keep them in mind? If you are like most people, you repeated the digit names over and over to yourself in the order you read them: two three eight zero four nine seven. Some IQ tests (discussed in Chapter 10) include a measure of digit span, the number of digits that the person can keep in mind for a brief period and report back accurately. Most people have a digit span of about seven digits. More generally, the number of pronounceable items—such as digits, other words, or nonsense syllables—that a person can keep in mind and report back accurately after a brief delay is called the short-term memory span, or simply memory span. According to Baddeley’s model, it might better be called the span of the phonological loop of working memory. The phonological loop is the part of working memory that holds on to verbal information by subvocally repeating it.

Research has shown that memory span, measured this way, depends on how rapidly the person can pronounce the items to be remembered (Baddeley, 1986). Generally, people can keep in working memory about as much verbal material as they can state aloud in 2 seconds (Baddeley et al., 1975). Unrehearsed items fade quickly; some of them begin to disappear within about 2 seconds or slightly longer. People who can speak rapidly have larger spans than people who cannot speak so rapidly. The span for single-syllable words is greater than that for multiple-syllable words. Try repeating from memory the following seven-word list, with eyes closed, immediately after reading it: disentangle appropriation gossamer anti-intellectual preventative foreclosure documentation. Was that list harder than the list of digits?

Any manipulation that interferes with a person’s ability to articulate the words to be remembered interferes with verbal short-term memory (Baddeley, 2003). Try to hold seven digits in mind while repeating over and over, out loud, the word the. You probably can’t do it; the act of saying the interferes with your ability to articulate to yourself the digit names.

An analogy to the phonological loop of working memory Holding several items of information in the phonological loop is a bit like spinning several plates on the ends of sticks. Just as you have to go back to each plate and renew its spin before it falls, you have to go back to each item in the phonological loop and repeat it before it vanishes from working memory.

Richard T. Nowitz/Corbis

Evidence that the time it takes to articulate words influences memory span comes from research examining digit spans for people speaking different languages. For example, Chinese speakers have longer digit spans than English speakers do, a difference that is apparent as early as 4 years of age and extends into adulthood (Chen & Stevenson, 1988; Geary et al., 1993). This difference is due to differences in the rate with which number words (one, two, and so on) in the two languages are spoken. The digit names in Chinese are all one syllable and can be articulated more quickly than the longer digit words of the English language. A similar pattern has been found between English and Welsh, with digit span being greater for the more rapidly spoken English digits than for the longer Welsh digits. This was true even for subjects whose first language was Welsh (Ellis & Hennelley, 1980).

Keeping a list of memory items in the phonological loop is a bit like a circus performer’s keeping a set of plates spinning on the ends of sticks. As the number of plates increases, the performer must work more frantically to get back to each and renew its spinning before it falls. Performers who can move quickly can spin more plates than performers who move slowly. Larger plates take longer to set in motion than smaller ones, so the performer can’t spin as many large plates as small ones. If the performer attempts to do another task at the same time that involves his or her arms and hands—such as building a tower of cups and saucers—the number of plates he or she can spin decreases.

Of course, in everyday life we don’t normally use our phonological loop to keep nonsensical lists in mind, any more than we use our hands to keep plates spinning. Rather, we use it for useful work. We say words silently to ourselves, and we bring ideas together in the form of words, as we reminisce about our experiences, solve problems, make plans, or in other ways engage in verbal thought. We don’t just hold material in working memory; we stream material through it, in an often-logical fashion.

Working-Memory Span

As we said, the capacity of the short-term store is assessed by memory-span tasks, with subjects recalling a series of items in the order they were presented. Many years ago, George Miller (1956) declared that the capacity of the short-term-memory store was seven plus or minus two items. This means that, depending on the information one is working with, an average adult can keep between five and nine items active in consciousness. Memory span, and thus the capacity of short-term memory, increases over childhood (Dempster, 1981) and decreases in old age (Horn & Hofer, 1992).

Simulated driving while talking on hands-free cell phone In their simulated-driving experiments, Strayer and Drews (2007) found that talking on a cell phone created more driving errors, regardless of whether or not the phone was hands free.

Photo by James Moulin/Courtesy of David Strayer. Ph.D

As useful as memory span is for assessing cognitive performance, in recent years cognitive psychologists have found that an even better measure for assessing cognitive abilities is to examine how many items a person can keep in mind while performing some “work.” In working-memory span tasks, subjects are asked remember a set of items while doing something with those items. For example, in a reading-span task subjects may be asked to read a set of short sentences (for instance, “In the summer it is very hot”; “The horse jumped over the fence”). After hearing several such sentences, subjects are asked to recall the last word in each sentence, in the order they were presented. Or subjects may be given a series of simple arithmetic problems to solve (8 + 1 = ?; 4 + 3 = ?) and asked to remember the sum (9, 7) of each problem, in the order they were presented. Working-memory span is typically about two items shorter than memory span, and also shows improvements over childhood and declines in older adulthood (Cowan & Alloway, 2009; Dykiert et al., 2012). One reason for the increased interest in working-memory span tasks is that, compared to memory-span tasks, they are more strongly associated with (and predictive of) important higher-level abilities, including reading, writing, mathematics, memory strategies, and IQ, among others (Bjorklund, 2013; Kane & Engle, 2002).

One way of demonstrating the importance of working memory on task performance is to examine what happens when someone tries to engage in two tasks at once, or multitasking. Consider, for example, the dual tasks of driving a car and talking on a cell phone. Although both driving and speaking are highly developed and automated skills, they each consume a portion of working memory, and so performing one interferes with performing the other. In a correlational study involving only drivers who sometimes used their cell phones while driving, the accident rate during phone use was four times that for the same group when they did not use their phones (Redelmeier & Tibshirani, 1997). In a simulated-driving experiment in the laboratory, conversing on a phone doubled the number of driving errors made (Strayer & Johnston, 2001). Moreover, it is important to note that in both of these studies, the disruptive effect on driving was as great for hand-free phones as for hand-held phones. The interference is a mental one, involving competing uses of working memory, not a motor one, involving competing uses of the hands. Subsequent simulated-driving experiments showed that drivers whose minds were occupied with phone conversations frequently missed road signs because of inattentional blindness, the same reason that people in the basketball pass-counting experiment missed the gorilla (Strayer & Drews, 2007). Conversations with passengers did not have the deleterious effects that cell phone conversations had. Why? Because passengers, unlike phone partners, experience the driving conditions that the driver experiences, so the conversation becomes synchronized with the driving; when the driving gets difficult, the conversation temporarily stops. The effects of texting while driving may be even more substantial, as this behavior combines cognitive distraction with manual and visual distractions (CDC, 2013). Indeed, as of 2014, 41 states and the District of Columbia had enacted texting bans for all drivers.

Although developing expertise In something like driving often permits one to “dual-task”—do two things at once—texting while driving shouldn’t be one of them. In a simulation of drivers (17 to 24 years of age) in the United Kingdom, texting while driving reduced the drivers’ reaction times by 35 percent. This compares to reductions of 12 percent due to alcohol consumption and 21 percent when smoking marijuana (Reed & Robbins, 2008), making texting a greater accident risk than driving while intoxicated.

Elena Elisseeva/Alamy