"For this third edition of Learning in Adulthood we have paid particular attention to work published since the last edition of the book. This third edition of Learning in Adulthood builds on material in the 1999 edition, bringing together the important contributions of the past decade to our understanding of adult learning. While we have preserved important foundational material (such as a discussion of andragogy), we have also brought to bear the most recent thinking and research. We have strived to put together a comprehensive overview and synthesis of what we know about adult learning: the context in which it takes place, who the participants are, what they learn and why, the nature of the learning process itself, new approaches to adult learning, the development of theory in adult learning, and other issues relevant to understanding adult learning."
—From the Preface
Previous praise for Learning in Adulthood
"An essential volume in adult education."
"This book is extremely useful in that it provides an informed overview of issues related to adult learning. It is very readable, yet packs considerable challenge for the more experienced adult educator."
—Canadian Journal of University Continuing Education
"Learning in Adulthood's greatest strengths are its coverage of the learning context, theory-building within adult learning, and how social and ethical issues offer special challenges to adult learning. It will be useful to both experienced and novice continuing educators, as well as laypersons."
—Continuing Higher Education Review
Sharan B. Merriam is professor of adult education at the University of Georgia. Merriam's research and writing activities have focused on the foundations of adult education, adult learning, adult development, and qualitative research methods. She has won the Cyril O. Houle Award for Outstanding Literature in Adult Education for three books, including one for Learning in Adulthood.
Rosemary S. Caffarella is professor and chair of the Department of Education in the College of Agriculture and Life Science at Cornell University. Her research and writing activities have focused on adult development and learning, program planning and evaluation, and leadership development. She has authored or coauthored a number of books, including the award-winning Learning in Adulthood.
Lisa M. Baumgartner is an associate professor of adult education at Northern Illinois University, DeKalb. Her research and writing focus on adult learning and development and women's contributions to the field of adult education. She is a recipient of the W.K. Kellogg Foundation's Cyril O. Houle Scholars Research Grant for Emerging Scholars in Adult Education.
The adage that "you can't teach an old dog new tricks" haunts both instructors of adults and adult learners themselves as they set forth on new learning ventures. The image of staff members who refuse to use the office e-mail system because "it just takes too long to learn" is still a reality. So too is the existence of the young trainer, fresh out of graduate school, who secretly believes she will never be able to teach her entrenched training staff anything, let alone to work as a team. This powerful myth-that adults lose their ability to learn as they age-prevails, although for the most part it has not been substantiated in the literature.
CONCEPT OF INTELLIGENCE AND ITS MEASUREMENT
The concept of intelligence has become much more complex and multifaceted over the last two decades, often causing confusion to casual readers. Different words describing the same concept have been interchanged or slightly modified, adding to the complexity of understanding the vast array of work on adult intelligence. For example, some authors use the terms intellectual development and cognitive development interchangeably, whereas other authors using these terms are exploring very different phenomena. Schaie (1996) and Berg (1992) have sought to add clarity to discussions of what the term intelligence means. Berg (1992) highlights several perspectives that dominate current research in intelligence. Among these are the psychometric tradition, the process orientation of Piagetian and neo-Piagetian thought, information processing, and the contextual perspective. Schaie (1996) agrees that at least two of these theoretical perspectives-those of information processing and the psychometric tradition-have driven the study of intelligence, but adds a third, that of practical intelligence. More specifically, Schaie purports that "there is a natural hierarchy in the study of intelligence leading from information processing, through the products measured in tests of intelligence, to practical and everyday intelligence" (p. 266). The Piagetian and neo-Piagetian traditions were discussed in Chapter Seven and the information processing perspective is discussed in Chapter Nine. In this chapter we use work from the psychometric tradition, practical intelligence, and the contextual perspective to illustrate important ideas related to intelligence and aging.
Intelligence Testing with Adults
The first widespread use of intelligence testing with adults was by the army in World War I, with the administration of the Army Alpha Tests of Intelligence. Thorndike, Bregman, Tilton, and Woodyard (1928) followed closely behind with their pioneering work that challenged the fundamental notion that learning ability peaks very early in life. Using primarily laboratory or schoolroom tasks, Thorndike measured the speed of the performance of people from ages fourteen to fifty on a variety of tasks, from memorizing poetry to acquiring an artificial language (Kidd, 1973). Thorndike, Bregman, Tilton, and Woodyard (1928, pp. 178-179) concluded from their many studies that, "in general, teachers of adults of age twenty-five to forty-five should expect them to learn at nearly the same rate and in nearly the same manner as they would have learned the same thing at twenty." In reflecting on Thorndike's work, Kidd (1973) noted two major contributions. The first was to raise the age of onset of the downhill slide of a person's ability to learn from twenty years of age to forty-five. Second, and even more important, Thorndike "helped to stimulate colleagues to reject traditional views and formulas" (Kidd, 1973, p. 79) about learning in adulthood. Other noted studies of intelligence (Jones and Conrad, 1933; Miles and Miles, 1932) of that same era reached similar conclusions, "although they found that the decline begins at a later age and the rate of that decline is not as sharp as in 'Thorndike's curve'" (Cross, 1981, p. 158).
to pay a debt. To these researchers, the findings suggest "a strong relationship between the 'building blocks' of intelligence and abilities on real life tasks" (p. 290). Only further research will tell whether the PMA and other intelligence tests of this nature can actually be used as adequate predictors of everyday intelligence.
Major Issues and Future Concerns in Assessing Adult Intelligence
In using the traditional psychometric measures of intelligence with adults, there are two cogent issues: the tests themselves and the social and policy implications of IQ scores. As Tennant and Pogson (1995) have observed, these tests "are too culture-specific; ... [and] they are constructed from problems and tasks derived from the context or 'culture' of schooling rather than everyday life" (p. 17). In other words, in using these tests, we perpetuate the notion of intelligence as being more "academic" in nature and culture free. In addition, Thomas (1992) has pointed out that the timed nature of some of these tests is biased against older adults, in that the reaction time of many older adults is slower (see Chapter Four).
REPRESENTATIVE THEORIES OF INTELLIGENCE
Two theories of intelligence grounded in the psychometric tradition of measurement, those of Horn and Cattell and of Guilford, which are useful in helping us understand more clearly the connection between intelligence and aging, are described first in this section. These descriptions are followed by a review of the work of scholars who have challenged traditional theories, such as Gardner and Sternberg, who believe that these traditional models have little, if any, relationship to what they term "real-world" or "practical" intelligence. Theorists who advocate practical intelligence most often view intelligence as a combination of biological and environmental factors and stress genetic and environmental interactions.
Theory of Fluid and Crystallized Intelligence
One of the major conceptualizations of adult intelligence was popularized by the work of Cattell (1963, 1987) and Horn (Horn 1976, 1982, 1985). They viewed intelligence as consisting of two primary factors: fluid intelligence (Gf) and crystallized intelligence (Gc). Fluid intelligence, or the ability to perceive complex relations and engage in short-term memory, concept formation, reasoning, and abstraction, is measured by tests for rote memory, basic reasoning, figural relations, and memory span. In contrast, crystallized intelligence is normally associated with acculturated information-those "sets of skills and bits of knowledge that we each learn as part of growing up in any given culture, such as verbal comprehension, vocabulary, [and] the ability to evaluate experience" (Bee, 1996, pp. 155-156). Examples of measures of crystallized intelligence include vocabulary and verbal comprehension, numerical reasoning, and an individual's ability to extract information from the environment. Tests of fluid intelligence are primarily speeded and are viewed as "culture fair," while crystallized intelligence is more likely to be assessed by nonspeeded measures. There is no single test that researchers can use to measure both factors of intelligence; rather, researchers often label tests as measures of either fluid or crystallized intelligence (see, for example, Christensen, 1994; Kaufman, Kaufman, Chen, and Kaufman, 1996).
Guilford's Structure of Intellect Model
In contrast to the two-factor model of fluid and crystallized intelligence proposed by Cattell, Guilford's model (1967, 1985) consists of 120 theoretical factors clustered into three major categories that are independent of each other: (1) contents, referring to the type of verbal, numerical, or behavioral material being tested; (2) operations, which are the basic mental processes such as memory, reasoning, and creative thinking; and (3) product, referring to the form of information that results from the interactions of the other two categories (from a single unit to complex patterns of information). A key assumption underlying the model is that the mental operations used on a particular task area are as important as the nature of the task itself. Guilford's model offers researchers an alternative frame of reference about the human intellect on which hypotheses about new factors of intelligence can be generated (Guilford, 1967, 1985; Huyck and Hoyer, 1982). Therefore, the model continues to provide a major building block for expanding our thinking about the fundamental nature of human intelligence.
Gardner's Theory of Multiple Intelligences
Gardner is representative of theorists who broke away from the psychometric tradition of intelligence during the early 1980s. From Gardner's perspective, the concept of intelligence has been too narrowly limited to the realm of logical and linguistic abilities, primarily by the way intelligence has been measured. Rather, he argues, "there is persuasive evidence for the existence of several relatively autonomous human intellectual competencies ... that can be fashioned and combined in a multiplicity of adaptive ways by individuals and cultures" (Gardner, 1983, pp. 8-9). From a number of unrelated sources, such as studies of prodigies, brain-damaged patients, and normal children and adults, Gardner originally identified seven different forms of intelligence, with an eighth recently added. His first seven forms include "not only the standard academic ones of linguistic, logical-mathematical, and spatial (the visual skills exhibited by a painter or architect) but also musical, 'bodily-kinesthetic,' and two 'personal' intelligences involving a fine-tuned understanding of oneself and others" (Levine, 1987, p. 54). (See Gardner, 1983, for a complete description of his seven original intelligences.) Gardner calls his newest form of intelligence the "naturalist" intelligence: "The intelligence of the naturalist involves the ability to recognize important distinctions in the natural world (among flora, fauna). It can also be applied to manmade objects in our consumer society (cars, sneakers). Obviously this skill is crucial in hunting or farming cultures, and it is at a premium among biologists and others who work with nature in our own society" (Shores, 1995, p. 5).
Sternberg's Theories of Intelligence
Sternberg too has broken from the tradition of framing intelligence as primarily a measure of formal testlike problem solving to one that includes problem solving for everyday life. Unlike the "schooling world," where problems are usually highly definitive and structured, real-world issues tend to be both ill defined and contexualized. Therefore, Sternberg contends that most theories, especially the measures of intelligence, address only the "schooling" kind of intelligence and almost totally ignore the notion of practical intelligence. Sternberg has proposed an important theory of human intelligence, the triarchic theory, and a layperson's version of theory that he terms "successful intelligence."
translate thought into action, have a product orientation, complete tasks and follow through, and are not afraid to risk failure. In summary, Sternberg's work on intelligence, like that of Gardner, is very useful in informing both the theory and practice of learning in adulthood.
Goleman's Theory of Emotional Intelligence
Goleman (1995) too has challenged our traditional views of intelligence. By expanding what we mean by intelligence, Goleman (1995, p. 9), grounding his work in the new discoveries of the emotional architecture of the brain, asserts that we have two very different ways of knowing-the rational and the emotional-which are, for the most part, intertwined and "exquisitely coordinated; feelings are essential to thought, thought to feelings." Yet, in Goleman's beliefs, it is the emotional mind-in his terms, emotional intelligence-that is the major determiner of success in life. Building on the work of Salovey and Mayer (1990), Goleman describes five major domains of emotional intelligence: knowing one's emotions, managing emotions, motivating oneself, recognizing emotions in others, and handling relationships. He believes self-awareness of one's feelings is the key to emotional intelligence, but one must also be attuned to the emotions of others. His descriptions of how adults might display their emotional intelligences are similar to Gardner's concepts of personal intelligences. For example, both authors speak to the need for people to make personal connections and be empathetic as well as to have access to their own internal feelings. In addition, Goleman's ideas about emotional intelligence are echoed in Sternberg's list of the characteristics and attributes of people who display successful intelligence.
THE CONTEXTUAL PERSPECTIVE OF INTELLIGENCE
Acknowledging the contextual dimension of intelligence in adulthood has moved thinking about intelligence beyond individual learners. Within the broad framework of the contextual perspective, two major threads are important in gaining a clearer understanding of adult intelligence: intellectual abilities lie at the intersection between mind and context, and intelligence is defined differently by different social classes and cultural groups.
AGE AND INTELLECTUAL ABILITIES
Does intelligence decline with age? Responses to this question are mixed and often controversial. The classic school of thought contends that intelligence enters a process of irreversible decline in the adult years, although the hypothesized onset of that decline has been extended from the early twenties to at least the age of fifty or sixty. Others say that intelligence is relatively stable through the adult years, with substantial intellectual changes occurring only very late in life, and then primarily "in abilities that were less central to the individual's life experience and thus perhaps less practiced" (Schaie, 1996, p. 2). In essence, we have enough brain capacity to do almost anything we choose, until serious illness sets in. Still others argue that intelligence declines in some respects, remains stable in others, and may even increase in some functions (Baltes, Dittmann-Kohli, and Dixon, 1984; Kaufman, Kaufman, Chen, and Kaufman, 1996; Raykov, 1995).
What Is Meant by Age and Aging
Whether or not one believes intelligence declines with age depends on "where in the age spectrum one chooses to look" (Botwinick, 1977, p. 580). Are we talking about adults in early, middle, or later adulthood? In reviewing data on early and middle adulthood, our response would be that intelligence does not decline with age. (See Baltes, Dittmann-Kohli, and Dixon, 1984; Schaie, 1994, 1996; Schaie and Hertzog, 1983.) In fact, some intellectual functions, no matter what testing procedures are employed, seem to increase over the course of the years. Our response to whether intelligence declines in later adulthood is not as clear-cut. (See Baltes, 1993; Baltes, Dittmann-Kohli, and Dixon, 1984; Horn and Donaldson, 1976; Schaie, 1996.) Most agree that some decline in functioning occurs between age sixty and the early seventies, but the precise nature of decline and, more important, its practical effect on learning ability are still unknown. For example, intelligence does appear to drop within a few years of death. This phenomenon, labeled the terminal drop, might account at least in part for the decrease in the intellectual functioning of older adults. This explanation contrasts with the traditional view that intellectual impairment is a universal condition of advancing age (Labouvie-Vief, 1990; Troll, 1982). In line with this observation, only a few studies have addressed the intellectual abilities of healthy adults beyond the age of seventy. In one longitudinal comparison of subjects ranging in age from seventy-three to ninety-nine, researchers found that although many of the subjects showed some decline in abilities, more than half displayed no such changes, even at the older ages (Field, Schaie, and Leino, 1988). In a more recent study of eighteen people between the ages of 100 and 106, these "centenarians reported rich late-life learning experiences, ... the majority of [which] occurred through social interactions" (Fenimore, 1997, p. 57).
Definitions of Intelligence
There is no universal agreement as to what constitutes intelligence. Therefore, when we speak about changes in intellectual ability as we age, a key question must be answered: What do we mean by intelligence? When intelligence is defined as a unitary property, the research tends to confirm that intelligence does indeed decline with age, although again the point of the departure for that decline often varies (Schaie and Willis, 1986). Yet when intelligence is defined as a multifaceted entity, the response tends to be that some of our abilities decline, while others remain stable or even increase (Baltes, 1993; Baltes, Dittmann-Kohli, and Dixon, 1984; Schaie, 1979, 1987, 1996; Sternberg and Berg, 1987).
Types of Tests
Behind every intelligence test is a definition of intelligence, whether given directly or implied by the test's author. Thus, it is apparent that generalizations about intelligence and about the aging process are also affected by the tests used.
The research designs employed in investigations of changes in intelligence over the life span have also generated much discussion in the literature. Results of cross-sectional studies, those that compare one-time test scores of different age groups (twenty, forty, and sixty years old, for example), have been misinterpreted to show that as we age, our intelligence declines. (See Botwinick, 1977; Knox, 1977; Schaie and Willis, 1986.) Findings from longitudinal studies, however, usually support a very different conclusion. Based primarily on readministration of intelligence tests over time and to the same group of people, various longitudinal investigations demonstrate that intellectual abilities of groups of older people are remarkably stable over time (Ivnik and others, 1995; Rabbitt and others, 1993; Schaie, 1994, 1996). However, the group data appear to mask individual changes within each of the cohort groups. Botwinick (1977, p. 582), reflecting on this problem of research designs, has observed that "the cross-sectional method may spuriously magnify age decline, and the longitudinal method may minimize it." In cross-sectional investigations, the background and experiences of the different cohorts being studied may cloud the results due to such factors as formal educational attainment and health. Moreover younger cohorts, especially those of college students in their twenties, may be more "testwise" than those comprising people in their sixties and seventies. In longitudinal studies, on the other hand, there are problems of selective attrition and dropout.
health care that may have resulted in superior physiological brain functioning.
INTELLIGENCE, AGING, AND ADULT LEARNING
Among the many new ideas about intellectual functioning in adulthood, three ideas surface as the most intriguing and useful to educators. The first is the framing of more holistic conceptions of adult intelligence that are grounded in the real lives of adults of all colors, races, and ethnic backgrounds. Theorists such as Gardner (1983, 1995) and Sternberg (1996a, 1996b) have moved the boundaries of intelligence beyond just the minds of individuals and have challenged researchers and practitioners to consider how the individual and the context interact to shape intellectual functioning in adulthood. They have also moved intelligence out of the hallowed hallways of schools and universities, which have always placed more value on academic skills and abilities, into how people function in all aspects of life. Sternberg (1994b) has added a further idea to understanding of the mind in context: that of the "luck" or "whoops" factor. Each of us is born with different gifts and into different circumstances. Some of us are lucky enough to find places where our gifts have been prized and nurtured (luck factor), while others, no matter what the individual effort, are never recognized or are blocked by circumstances beyond their control (whoops factor).
In this chapter we discussed the concept of intelligence and how different ways of viewing intelligence can better help us understand learning in adulthood. The most often used paradigm of intelligence is the psychometric tradition, which assumes that intelligence is a measurable quantity. First conceptualized from this tradition as a single factor of general ability, the construct has broadened to include the notion that there are multiple forms of intellectual ability, such as those proposed by Horn (1976), Cattell (1963), and Guilford (1967). Commonly used tests of adult intelligence that fit within this psychometric tradition include the Wechsler Adult Intelligence Scale-Revised and the Primary Mental Abilities Test. Two issues that have surfaced with the use of these types of tests are the test items and the social and policy implications of IQ scores.