CHAPTER 18 Procedural Memory and Skill Acquisition ADDIE JOHNSON DECLARATIVE MEMORY AND SKILL ACQUISITION The Roles of Attention and Intention in Memory and Skill 500 Learning to Ignore Irrelevant Information 501 IMPLICIT LEARNING 502 Implicit Learning and Awareness 502 Implicit Learning and Attention 503 The Nature of Implicit Learning 504 PROCEDURAL MEMORY 504 Evidence for Procedural Memory 505 A Procedural Memory System? 505 PROCEDURAL MEMORY, IMPLICIT LEARNING, AND SKILL 506 SKILLED PERFORMANCE 506 Phases of Skill Acquisition 506 Mechanisms of Change 507 TYPES OF SKILLS 507 Perceptual Skill 507 Cognitive Skill 508 Motor Skill 509 FACTORS INFLUENCING SKILL ACQUISITION 510 Feedback 510 Practice Schedules 511 INDIVIDUAL DIFFERENCES IN SKILLED PERFORMANCE 511 EXPERTISE 512 Studying Expertise 513 Characteristics of Expertise 513 Skill and Expertise 513 TRAINING 514 Skill Acquisition and Attentional Strategies 516 Automaticity and Training 516 Team Training 516 RETENTION AND TRANSFER OF SKILL 517 Transfer of Training 517 Long-Term Retention of Skill 517 MODELING SKILL 518 NEW DIRECTIONS 518 REFERENCES 518 499 One of the most remarkable things about human performance is the regularity, efficiency, and precision with which it commonly occurs Despite the fact that we are presented with a complex array of stimuli in a constantly changing environment with a bewildering array of choices, things usually go as planned Even in the performance of complex tasks, patterns of stimuli in the environment are grouped and reacted to in what appears to be seamless, coordinated ease Skilled performance obviously depends on prior experience, but exactly what must be learned and remembered in order to develop and exercise skill? What aspects from learning episodes are important for the development of skill, and what aspects of memory are involved in this learning? These are key issues in understanding the development, maintenance, and exercise of skill Other issues of importance are the roles of forgetting, the making of mistakes, and attention in the acquisition and execution of skilled performance In this chapter, the roles of explicit, declarative memory in skilled performance will be considered and contrasted with the role of implicit, procedural memory DECLARATIVE MEMORY AND SKILL ACQUISITION It is probably not too daring to say that all major models of skill acquisition, just as the acquisition of skill, itself, begin with declarative memory Declarative memory has been described as an episodic or recollective memory system (Squire, 1992), the characterization of which overlaps with descriptions of episodic and semantic memory (see the chapters in this volume by Nairne; McNamara & Holbrook; and Roediger & Marsh) Basically, declarative memory refers to a system that works with verbalizable knowledge In his 499 500 Procedural Memory and Skill Acquisition influential ACT* (and ACT-R) model of the development of cognitive skill, Anderson (1982, 1983, 1993) calls the first stage in the development of skill the declarative stage Anderson’s work will be more fully described in a later section At this point it is sufficient to note that the declarative stage is one in which verbal mediation is used to maintain facts in working memory so that they can be used to execute the task at hand In other words, performance at this level depends heavily on declarative memory Fitts (1962/1990, 1964; Fitts & Posner, 1967) called the first phase of skill acquisition by a different name, but his cognitive phase also depends heavily on declarative memory for comprehending instructions and maintaining a description of the cues that must be attended to and the relevance of the feedback that is provided during performance In the frameworks of both Anderson and Fitts, the development of skill is characterized by reduced dependence on declarative memory At least one account of skill acquisition, Logan’s (1988, 1990) instance theory of automaticity, suggests that memory demands of performance not qualitatively change as a function of skill, at least not once the basic instructions have been mastered Logan’s theory may not apply to skill acquisition in a broad sense, but it has been to shown to provide a good description of the development of skilled performance in a range of cognitive tasks Logan describes the development of automaticity as the shift from a dependence on general algorithms that not rely on previous experience but that are sufficient to produce solutions to problems posed by the task, to a reliance on the retrieval of performance episodes Memory plays a critical role in this model in which skilled, automatic performance entails a shift from algorithmbased performance to memory-based performance The instance theory of automaticity rests on several assumptions The first of these assumptions is that encoding is obligatory, such that attention to an object or event is sufficient for it to be encoded into memory The second assumption calls for obligatory retrieval, in which attention to an object or event is sufficient to cause things associated with it to be retrieved An additional, critical assumption is that each encounter with an object or event is encoded, stored, and retrieved separately, and on every encounter These encounters are the instances in the instance theory of automaticity As mentioned above, the instance theory assumes that automaticity involves a transition from performance based on general rules or algorithms for performing a task to performance based on the retrieval of instances Once performance is instance based, it continues to speed up because the number of instances continues to increase as long as the task is practiced This speed-up is predicted on the basis of the statistical properties of the distribution of retrieval times for Figure 18.1 Performance speed-up in various tasks illustrating the power law of practice Note: When plotted in log-log coordinates, a power function appears as a straight line CRT = choice reaction time; S-R = stimulusresponse instances: As the number of instances increases, the minimum time to retrieve an instance decreases Because retrieval is obligatory, according to the theory, performance time will decrease as a function of practice due to this faster retrieval time An important aspect of the theory is that it predicts that changes in performance will follow a power function This is consistent with the power law of practice, which reflects the finding that performance improvements in many tasks follow a power function (see Figure 18.1) It can be argued that the early dependence on an algorithm for task performance can be likened to the declarative or cognitive phase of the frameworks of Fitts (1962/1990, 1964) and Anderson (1982, 1983) At this stage, the rules or guidelines for performing a task presumably must be active in working memory, and performance is relatively deliberative and slow As a result of experience, and of paying attention to the right things at the right time, a collection of memory traces, or instances, builds up and gradually comes to dominate performance The Roles of Attention and Intention in Memory and Skill Attention has assumed a curious place in the study of skill acquisition Often, it seems that the goal of researchers has been to show that attention may not be necessary at all once a skill has been learned The traditional view of attentive processing (or “controlled” processing; Atkinson & Shiffrin, 1968) is that it is relatively slow, requires effort, and involves consciousness of one’s actions Skill is described as a gradual (or abrupt) freeing of resources and shift to a capacity-free, Declarative Memory and Skill Acquisition stimulus-driven mode of performance that is not dependent on conscious control Posner and Snyder (1975) described automatic processes as those that may occur “without intention, without any conscious awareness and without interference with other mental activity” (p 81) A great deal of research has been directed to exploring and confirming this view of dichotomous processing modes For example, W Schneider and Shiffrin (1977; Shiffrin & Schneider, 1977) performed an extensive series of hybrid memory and visual search experiments that seemed to support the idea that there are two different modes of processing and that controlled processing gives way to automatic processing if only enough practice is given The view that controlled and automatic processing are qualitatively distinct has, to some extent, fallen out of favor Within the realm of visual search, where Shiffrin and Schneider (1977) carried out their influential work supporting such a dichotomy, researchers now tend speak about the efficiency of search, rather than pre-attentive and attentive search, and the role of attention in processing remains present across search types Rather than considering it a form of processing, Neumann (1987) describes automaticity as a phenomenon arising from a conjunction of input stimuli, skill, and the desired action In his view, it is appropriate to speak of automaticity when all the information for performing a task is present in the input information (stimulus information available in the environment) or in long-term memory This view is not too different from Logan’s (1988, 1990), described above, in which automatic processing is based on memory retrieval, and attention forms the cues necessary for the retrieval processing Attention remains an important process even in highly practiced tasks As will be discussed at more length in the section on training, automatic processing, as assessed by an apparent insensitivity to attentional resources or demands, can develop with learning when the right conditions are provided The important conditions seem to be the consistency of the discrimination and interpretation of the stimuli, and the stimulus-to-response mapping (W Schneider & Fisk, 1982) The development of automaticity can be shown for a range of tasks The idea that it depends more on consistency than on properties of the stimuli, such as perceptual salience, is supported by the finding that automatic processing can also be produced by training with stimuli divided into arbitrary classes (Shiffrin & Schneider, 1977) According to the instance theory, “attention drives both the acquisition of automaticity and the expression of automaticity in skilled performance” (Logan & Compton, 1998, p 114) Selected information enters into the instances that come to drive performance, but ignored information does not Moreover, if attention is not paid to the right cues, 501 associations dependent on those cues will not be retrieved (Logan & Etherton, 1994) Logan and Compton describe attention as an interface between memory and events in the world The dependence of memory on attention means that knowing (or learning) what to attend to is a critical component in the development of skill Other authors have emphasized that learning not to attend to irrelevant information is also a component of skill acquisition Learning to Ignore Irrelevant Information One hypothesis about how learning to ignore irrelevant information contributes to performance changes with practice is the information reduction hypothesis (Haider & Frensch, 1996) According to this hypothesis, performance improvements can be attributed to learning to distinguish taskrelevant information from task-redundant (and, therefore, task-irrelevant) information and then learning to ignore the task-irrelevant information Evidence for this hypothesis comes largely from tasks in which participants verified alphabetic strings such as E [4] J K L The task is to determine whether the letters follow in alphabetic order, where the number in brackets corresponds to the number of letters left out of the alphabetic sequence In most conditions, the length of the string was varied by changing the number of letters following the digit, which always occupied the second position in the string If there was an error in the stimulus, the error was in the number of letters that was skipped (e.g., E [4] K L M) Early in practice, Haider and Frensch found an effect of string length on performance, such that verification times were slower when the number of letters after the number in brackets was increased With practice, however, the slope of the function relating performance time to string length decreased This finding suggests that participants in the study learned that the extra letters were not important for the task and should be ignored Additional evidence for this hypothesis was found in a transfer condition in which errors could occur in the letters to the right of the gap (e.g., E [4] J K M) Consistent with the supposition that participants learned to ignore the extra letters during training, the error rate in detecting these invalid sequences increased as a function of practice Haider and Frensch also showed that learning in this task was not stimulus specific by demonstrating transfer from one half of the alphabet to the other Haider and Frensch (1996) showed that learners were able to distinguish relevant from redundant task information and to limit their processing to the relevant information They also showed that learning to reduce the amount of information that is processed takes time, developing over the course of practice, and that this ability appears to be largely stimulus 502 Procedural Memory and Skill Acquisition independent Moreover, after finding that speed instructions affect whether or not people learn to ignore irrelevant information, Haider and Frensch (1999) argued that skill acquisition is neither passive nor “low-level,” but at least partly under the influence of intention It seems obvious that knowing what to attend to will increase the chance that the right events are experienced such that useful instances are created, and that the allocation of attention at encoding and retrieval determines to a large extent both the nature of what is learned and the influence of previous experiences on performance in the present There is, however, much to be said, and even more to be learned, about the interplay between intention and attention, and about how much we learn without really intending it IMPLICIT LEARNING Learning without intention, and without conscious awareness of what is being learned, is a topic that has received much attention in recent decades Models of skill typically emphasize early processing of task instructions and goal-directed learning, and paying attention to the correct elements in a task situation is considered crucial to eventual skilled performance The topic of this section is implicit learning (also referred to as incidental learning), that is, learning without intention, or the unintended by-product of experience with a task Consider a relatively simple task, that of pressing an assigned key whenever a stimulus appears at one of four particular locations on a screen The instructions are simple: Press the rightmost key when the rightmost stimulus appears, the second key to the right for a stimulus in the corresponding location, and so on One aspect of performance in such a task is that, despite the simplicity of the task, performance improves as a function of practice Reaction times become faster and error rates lower (Dutta & Proctor, 1992; Proctor & Dutta, 1993), with improvements in accuracy and reaction time typically following a power function (Newell & Rosenbloom, 1981; see Figure 18.1 and the chapter by Proctor & Vu in this volume) These improvements can be attributed to intentional learning of key and stimulus locations and of the stimulusresponse associations Performance can be considerably improved if elements are repeated within the sequence of trials One sort of repetition is just that: A particular stimulus may be repeated in two successive trials (see the chapter by Proctor & Vu for a discussion of the basis of such repetition effects) However, even when the repetition occurs across a longer sequence of trials, benefits of repetition can occur Nissen and Bullemer (1987) provided practice with the task described above, in which keys are pressed according to the spatial location of targets Within the sequence of trials, certain stimuli were repeated (designating the positions from left to right as A, B, C, and D, the repeating sequence was D-B-C-A-C-B-D-C-B-A) People who practiced this serial response time (SRT) task with the 10-element repeating sequence showed vastly more improvement than those who practiced the task with a random presentation of stimuli, even though the participants were not informed that there was a repeating sequence or instructed to look for repetitions while performing the task Implicit Learning and Awareness The participants in Nissen and Bullemer’s (1987) study evidently learned something (the repeating sequence) even though they were not instructed to so Organizing and making sense of the environment is, however, something that comes naturally to most of us The question is, then, whether participants in Nissen and Bullemer’s study either consciously looked for or somehow noticed that there was a repeating sequence and used this explicit knowledge to improve task performance In order to separate intentional and incidental learning in this task, and in order to assess the role of awareness in the performance of the task, Nissen and Bullemer asked participants whether they were aware of any sequences in the stimuli All of the participants in the repeated-sequence condition reported being aware of the sequence Thus, awareness was coupled with the improvement of performance for this group In order to address the question of whether awareness was necessary for the performance benefit to occur, Nissen and Bullemer repeated the experiment with a group of individuals characterized by a profound amnesia that prevented them from recognizing and recalling material to which they had been exposed: Korsakoff patients As predicted, the Korsakoff patients reported no awareness of the repeating sequence More interesting, their performance showed a degree of learning of the sequence comparable to that of controls (see Figure 18.2) This shows that learning can and does occur without awareness Later work (Willingham, Nissen, & Bullemer, 1989) showed that the degree of awareness of the sequence is correlated with performance for normal participants: People who showed more awareness (as indexed by explicit recall of the sequence) also showed more performance improvement However, when anticipatory responses (i.e., pressing the response key before the next stimulus appeared) were eliminated from the analysis, the difference in performance between those who reported full or partial knowledge of the sequence and those who could evidence no explicit knowledge was minimal ... development of skill, and what aspects of memory are involved in this learning? These are key issues in understanding the development, maintenance, and exercise of skill Other issues of importance... complex array of stimuli in a constantly changing environment with a bewildering array of choices, things usually go as planned Even in the performance of complex tasks, patterns of stimuli in... of importance are the roles of forgetting, the making of mistakes, and attention in the acquisition and execution of skilled performance In this chapter, the roles of explicit, declarative memory