Developmental Psychology 1987, Vol. 2.1, No. 5, 611-626
Copyright 1987 by the American Psychological Association, Inc. 0OI2-l649/87/$00.75
Theoretical Propositions of Life-Span Developmental Psychology: On the Dynamics Between Growth and Decline
Paul B. Baltes Max Planck Institute for Human Development and Education
Berlin, Federal Republic of Germany
Life-span developmental psychology involves the study of constancy and change in behavior throughout the life course. One aspect of life-span research has been the advancement of a more general, metatheoretical view on the nature of development. The family of theoretical perspectives associated with this metatheoretical view of life-span developmental psychology includes the recog- nition of multidirectionality in ontogenetic change, consideration of both age-connected and discon- nected developmental factors, a focus on the dynamic and continuous interplay between growth (gain) and decline (loss), emphasis on historical embeddedness and other structural contextual fac- tors, and the study of the range of plasticity in development. Application of the family of perspectives associated with life-span developmental psychology is illustrated for the domain of intellectual devel- opment. Two recently emerging perspectives of the family of beliefs are given particular attention. The first proposition is methodological and suggests that plasticity can best be studied with a research strategy called testing-the-limits. The second proposition is theoretical and proffers that any develop- mental change includes the joint occurrence of gain (growth) and loss (decline) in adaptive capacity. To assess the pattern of positive (gains) and negative (losses) consequences resulting from develop- ment, it is necessary to know the criterion demands posed by the individual and the environment during the lifelong process of adaptation.
The study of life-span development is not a homogeneous field. It comes in two major interrelated modes. The first mode is the extension of developmental studies across the life course without a major effort at the construction of metatheory that emanates from life-span work. The second mode includes the endeavor to explore whether life-span research has specific im- plications for the general nature of developmental theory. The second approach represents the topic of this article.
Specifically, the purpose of this article is twofold. First, after a brief introduction to the field of life-span developmental psy- chology, some “prototypical” features of the life-span approach in developmental psychology are presented. Second, these fea- tures are illustrated by work in one domain: intellectual devel- opment. Although the focus of this paper is on life-span develop- mental psychology and its theoretical thrust, it is important to recognize at the outset that similar perspectives on developmen- tal theory have been advanced in other quarters of developmen- tal scholarship as well (Hetherington & Baltes, in press; Scan; 1986). There is, however, a major difference in the “gestalt” in which the features of the theoretical perspective of life-span psy- chology are organized.
This article is based on invited addresses to Division 7 of the Ameri- can Psychological Association (Toronto, Canada, August 1984) and to the International Society for the Study of Behavioral Development (Tours, France, July 1985).
Helpful suggestions by Michael Chapman and many valuable discus- sions with members of the Social Science Research Council Committee on Life-Course Perspectives in Human Development are acknowledged.
Correspondence concerning this article should be addressed to Paul B. Baltes, Max Planck Institute for Human Development and Educa- tion, Lentzeallee 94, D-1000 Berlin 3 J, Federal Republic of Germany.
Several, if not most, of the arguments presented here are con- sistent with earlier publications by the author and others on the field of life-span developmental psychology (Baltes, 1983; Baltes & Reese, 1984;Featherman, 1983;Honzik, 1984;Lerner, 1984; Sherrod & Brim, 1986). This article includes two added empha- ses. First, it represents an effort to illustrate the implications of the theoretical perspectives associated with life-span work for research in cognitive development. Second, two of the more re- cent perspectives derived from life-span work are given special attention. The first is that any process of development entails an inherent dynamic between gains and losses. According to this belief, no process of development consists only of growth or progression. The second proposition states that the range of plasticity can best be studied with a research strategy called tes- ting-the-Umits.
What is Life-Span Development?
Life-span developmental psychology involves the study of constancy and change in behavior throughout the life course (ontogenesis), from conception to death.1 The goal is to obtain knowledge about general principles of life-long development, about interindividual differences and similarities in develop- ment, as well as about the degree and conditions of individual plasticity or modifiability of development (Baltes, Reese, & Nes- selroade, 1977;Lerner, l984;Thomae, 1979).
1 In this article, the terms life span and life course are used inter- changeably. Since the origin of the West Virginia Conference Series (Goulet & Baltes, 1970), psychologists tend to prefer life span (see, how- ever. Biihler, 1933), whereas sociologists lean toward the use of the term life course.
612 PAUL B, BALTES
It is usually assumed that child development, rather than life- span development, was the subject matter of the initial scholarly pursuits into psychological ontogenesis. Several historical re- views suggest that this generalization is inaccurate (Baltes, 1983; Groffinann, 1970; Reinert, 1979). The major historical precursors of scholarship on the nature of psychological devel- opment—byTetensin 1777,Carusin 1808,Quete!etin 1835— were essentially life-span and not child-centered in approach. Despite these early origins of life-span thinking, however, life- span development has begun to be studied empirically only dur- ing the last two decades by researchers following the lead of early twentieth century psychologists such as Charlotte Biihler (1933), Erik H. Erikson (1959), G. Stanley Hall (1922), H. L. Hollingworth (1927), and Carl G. Jung (1933).
Three events seem particularly relevant to the more recent burgeoning of interest in life-span conceptions; (a) population demographic changes toward a higher percentage of elderly members; (b) the concurrent emergence of gerontology as a field of specialization, with its search for the life-long precursors of aging (Birren & Schaie, 1985); and (c) the “aging” of the sub- jects and researchers of the several classical longitudinal studies on child development begun in the 1920s and 1930s (Migdal, Abeles, & Sherrod, 1981; Verdonik & Sherrod, 1984). These events and others have pushed developmental scholarship to- ward recognizing the entire life span as a scientifically and so- cially important focus.
Added justification for a life-span view of ontogenesis and im- portant scholarly contributions originate in other disciplines as well. One such impetus for life-span work comes From sociology and anthropology (Bertaux & Kohli, 1984; Brim & Wheeler, 1966; Clausen, 1986;Dannefer, 1984; Elder, 1985;Featherman, 1983; Featherman & Lerner, 1985; Kertzer & Keith, 1983; Neugarten & Datan, 1973; Riley, 1985; Riley, Johnson, & Foner, 1972). Especially within sociology, the study of the life course and of the interage and intergenerational fabric of soci- ety is enjoying a level of attention comparable with that of the life-span approach in psychology.
Another societal or social raison d’etre for the existence of life-span interest is the status of the life course in longstanding images that societies and their members hold about the life span (Philibert, 1968; Sears, 1986). In the humanities, for example, life-span considerations have been shown to be part of everyday views of the structure and function of the human condition for many centuries. The Jewish Talmud, Greek and Roman philos- ophy (e.g., the writings of Solon and Cicero), literary works such as those of Shakespeare, Goethe, or Schopenhauer, all contain fairly precise images and beliefs about the nature of life-long change and its embeddedness in the age-graded structure of the society. Particularly vivid examples of such social images of the life span come from the arts. During the last centuries, many works of art were produced, reproduced, and modified in most European countries, each using stages, steps, or ladders as a framework for depicting the human life course (JoeriBen & Will, 1983; Sears, 1986).
These observations on literature, art history, and social im- ages of the life course suggest that the field of life-span develop- ment is by no means an invention of developmental psycholo- gists. Rather, its recent emergence in psychology reflects the per- haps belated effort on the part of psychologists to attend to an aspect of the human condition that is part and parcel of our
everyday cultural knowledge systems about living organisms. Such social images suggest that the life course is something akin to a natural, social category of knowledge about ontogenesis and the human condition.
Is Life-Span Development a Theory or a Field of Specialization?
What about the theoretical spectrum represented by life-span developmental psychology? Is it a single theory, a collection of subtheories, or just a theoretical orientation? Initial interest of- ten converges on the immediate search for one overarching and unifying theory such as Erikson’s (1959). The current research scene suggests that in the immediate future life-span develop- mental psychology will not be identified with a single theory. It is above all a subject matter divided into varying scholarly specializations. The most general orientation toward this sub- ject matter is simply to view behavioral development as a life- long process.
Such a lack of theoretical specificity may come as a surprise and be seen as a sign that the life-span perspective is doomed. In fact, the quest for a single, good theory (and the resulting frustration when none is offered) is an occasional challenge laid at the doorsteps of life-span scholars (Kaplan, 1983; Scholnick, 1985; Sears, 1980). Note, however, that the same lack of theoret- ical specificity applies to other fields of developmental special- ization. Infant development, child development, gerontology, are also not theories in themselves, nor should one expect that there would be a single theory in any of these fields. In fact, as long as scholars look for the theory of life-span development they are likely to be disappointed.
A Family of Perspectives Characterizes the Life-Span Approach
Much of life-span research proceeds within the theoretical scenarios of child developmental or aging work. In addition, however, efforts have been made by a feir number of life-span scholars to examine the question of whether life-span research suggests a particular metatheoretical world view (Reese & Over- ton, 1970) on the nature of development. The theoretical pos- ture proffered by this endeavor is the focus of the remainder of this article.
For many researchers, the life-span orientation entails several prototypical beliefs that, in their weighting and coordination, form a family of perspectives that together specify a coherent metatheoretical view on the nature of development. The sig- nificance of these beliefs lies not in the individual items but in the pattern. Indeed, none of the individual propositions taken separately is new, which is perhaps one reason why some com- mentators have argued that life-span work has little new to offer (Kaplan, 1983). Their significance consists instead in the whole complex of perspectives considered as a metatheoretical world view and applied with some degree of radicalism to the study of development.
What is this family of perspectives, which in their coordi- nated application characterizes the life-span approach? Per- haps no single set of beliefs would qualify in any definite sense. However, the beliefs summarized in Table 1 are likely to be shared by many life-span scholars. They can be identified pri-
LIFE-SPAN DEVELOPMENTAL PSYCHOLOGY 613
Table 1 Summary of Family of Theoretical Propositions Characteristic of Life-Span Developmental Psychology
Development as gain/loss
Contextualism as paradigm
Field of development as multidisciplinary
Ontogenetic development is a life-long process. No age period holds supremacy in regulating the nature of development. During development, and at all stages of the life span, both continuous (cumulative) and discontinuous (innovative) processes are at work.
Considerable diversity or pluralism is found in the directionality of changes that constitute ontogenesis, even within the same domain. The direction of change varies by categories of behavior. In addition, during the same developmental periods, some systems of behavior show increases, whereas others evince decreases in level of functioning.
The process of development is not a simple movement toward higher efficacy, such as incremental growth. Rather, throughout life, development always consists of the joint occurrence of gain (growth) and loss (decline).
Much intraindividual plasticity (within-person modifiability) is found in psychological development. Depending on the life conditions and experiences by a given individual, his or her developmental course can take many forms. The key developmental agenda is the search for the range of plasticity and its constraints.
Ontogenetic development can also vary substantially in accordance with historical-cultural conditions. How ontogenetic (age-related) development proceeds is markedly influenced by the kind of sociocultural conditions existing in a given historical period, and by how these evolve over time.
Any particular course of individual development can be understood as the outcome of the interactions (dialectics) among three systems of developmental influences: age-graded, history- graded, and nonnormative. The operation of these systems can be characterized in terms of the metatheoretical principles associated with contextualism.
Psychological development needs to be seen in the interdisciplinary context provided by other disciplines (e.g., anthropology, biology, sociology) concerned with human development. The openness of the life-span perspective to interdisciplinary posture implies that a “purist” psychological view offers but a partial representation of behavioral development from conception to death.
marily from the writings in psychology on this topic (Baltes & Reese, 1984; Baltes, Reese, & Lipsitt, 1980; Lerner, 1984; Sher- rod & Brim, 1986), but they are also consistent with sociologi- cal work on the life course (Elder, 1985; Featherman, 1983; Riley, 1985).
The family of beliefs will be illustrated below, primarily using the study of intellectual development as the forum for exposi- tion. Because, historically, the period of adulthood was the pri- mary arena of relevant research, that age period receives most coverage. It will also be shown, however, that the metatheoreti- cal posture may shed new light on intellectual development in younger age groups as well.
Empirical Illustration: Research on Intellectual Development
The area of intellectual functioning is perhaps the best stud- ied domain of life-span developmental psychology. The discus- sion and elaboration of the empirical and conceptual bases of the family of perspectives presented here is selective. The intent is not to be comprehensive but to offer examples of areas of research. (Different research findings and agendas—see, for in- stance, Keating & MacLean, in press; Perlmutter, in press; Sternberg, in press, for other similar efforts—could have been used were it not for the particular preferences of the author.) More detailed information on the topic of life-span intelligence is contained in several recent publications (Baltes, Dittmann-
Kohli, & Dixon, 1984; Berg & Sternberg, 1985; Denney, 1984; Dixon & Baltes, 1986; Labouvie-Vief, 1985; Perlmutter, in press; Rybash, Hoyer, & Roodin, 1986;Salthouse, 1985).
Intellectual Development Is a Life-Long Process Involving Multidirectionality
The first two of the family of perspectives (see Table 1) state that behavior-change processes falling under the general rubric of development can occur at any point in the life course, from conception to death. Moreover, such developmental changes can display distinct trajectories as far as their directionality is concerned.
Life-long development- The notion of life-long development implies two aspects. First, there is the general idea that develop- ment extends over the entire life span. Second, there is the added possibility that life-long development may involve processes of change that do not originate at birth but lie in later periods of the life span. Considered as a whole, life-long development is a system of diverse change patterns that differ, for example, in terms of timing (onset, duration, termination), direction, and order.
One way to give substance to the notion of life-long develop- ment is to think of the kinds of demands and opportunities that individuals face as they move through life. Havighurst’s (1948/ 1972; Oerter, 1986) formulation of the concept of developmen- tal tasks is a useful aid for grasping the notion of a life-long
614 PAUL B. BALTES
system of demands and opportunities. Developmental tasks in- volve a series of problems, challenges, or life-adjustment situa- tions that come from biological development, social expecta- tions, and personal action. These problems “change through life and give direction, force, and substance to . . . development.” (Havighurst, 1973, p. 11).
Thus, the different developmental curves constitutive of life- long development can be interpreted to reflect different develop- mental tasks. Some of these developmental tasks—like Havig- hurst’s conception—are strongly correlated with age. However, as will be shown later, such developmental tasks are also consti- tuted from certain historical and nonnormative systems of in- fluence.
Multidimensionality and muhidirectionality The terms multidimensionality and mullidirectionality are among the key concepts used by life-span researchers to describe facets of plu- rality in the course of development and to promote a concept of development that is not bound by a single criterion of growth in terms of a general increase in size or functional efficacy.
Research on psychometric intelligence illustrates the useful- ness of multidimensional and multidirectional conceptions of development. The psychometric theory of fluid-crystallized in- telligence proposed by Cattell (1971) and Horn (1970, 1982) serves as an example (Figure 1). First, according to this theory, intelligence consists of several subcomponents. Fluid and crys- tallized intelligence are the two most important clusters of abili- ties in the theory. The postulate of a system of abilities is an example of multidimensionality. Second, these multiple-ability components are expected to differ in the direction of their devel- opment. Fluid intelligence shows a turning point in adulthood (toward decline), whereas crystallized intelligence exhibits the continuation of an incremental function. This is an example of the multidirectionality of development.
Meanwhile, the Cattell-Horn theoretical approach to life- span intellectual development has been supplemented with other conceptions, each also suggesting the possibility of multi- dimensional and multidirectional change. Berg and Stemberg (1985), for example, have examined the implications of Stern- berg’s triarchic theory of intelligence for the nature of life-long development. They emphasized that the age trajectories for the three postulated components of Sternberg’s theory of intelli- gence (componential, contextual, experiential) are likely to vary in directionality.
Current work on life-span intelligence (Baltes et al., 1984; Dixon & Baltes, 1986) has also expanded on the Cattell-Horn model by linking it to ongoing work in cognitive psychology. Two domains of cognitive functioning are distinguished in a du- al-process scheme: the “fluid” mechanics and the “crystallized” pragmatics of intelligence. The mechanics of intelligence refers to the basic architecture of information processing and prob- lem solving. It deals with the basic cognitive operations and cog- nitive structures associated with such tasks as perceiving re- lations and classification. The second domain of the dual-pro- cess scheme, the pragmatics of intelligence, concerns the context- and knowledge-related application of the mechanics of intelligence. The intent is to subsume under the pragmatics of intelligence (a) fairly general systems of factual and procedural knowledge, such as crystallized intelligence; (b) specialized sys- tems of factual and procedural knowledge, such as occupa- tional expertise; and (c) knowledge about factors of perfor-
mance, that is, about skills relevant for the activation of intelli- gence in specific contexts requiring intelligent action. The explicit focus on the pragmatics of intelligence demands (like Sternberg’s concern with contextual and experiential compo- nents) a forceful consideration of the changing structure and function of knowledge systems across the life span (see also Featherman, 1983; Keating & MacLean, in press; Labouvie- Vief, 1985;Rybashetal., 1986).
New forms of intelligence in adulthood and old age? What about the question of later-life, “innovative” emergence of new forms of intelligence? It is one thing to argue in principle that new developmental acquisitions can emerge at later points in life with relatively little connection to earlier processes and quite another to demonstrate empirically the existence of such innovative, developmentally late phenomena. A classic example used by life-span researchers to illustrate the developmentally late emergence of a cognitive system is the process of reminis- cence and life review (Butler, 1963). The process of reviewing and reconstructing one’s life has been argued to be primarily a late-life phenomenon. The phenomenon of autobiographical memory is another example (Strube, 1985).
For the sample case of intelligence, it is an open question whether adulthood and old age bring with them new forms of directionality and intellectual functioning, or whether continu- ation and quantitative (but not qualitative) variation of past functioning is the gist of the process. On the one hand, there are Flavell’s (1970) and Piaget’s (1972) cognitive-structuralist accounts, which favor an interpretation of horizontal decalage rather than one of further structural evolution or transforma- tion. These authors regard the basic cognitive operations as fixed by early adulthood. What changes afterwards is the con- tent domains to which cognitive structures are applied.
On the other hand, there are very active research programs engaged in the search for qualitatively or structurally new forms of adult intelligence (Commons, Richards, & Armon, 1984). The work on dialectical and postformal operations by Basseches (1984), Labouvie-Vief(1982, 1985), Kramer (1983), Pascual- Leone (1983), and other related writings (e.g., Keating & MacLean, in press) are notable examples of this effort. This line of scholarship has been much stimulated (or even launched) by Riegel’s work on dialectical psychology and his outline of a pos- sible fifth stage of cognitive development (Riegel, 1973, 1976).
Other work on adult intellectual development proceeds from a neofunctionalist perspective (Beilin, 1983; Dixon & Baltes, 1986) and is informed by work in cognitive psychology. Models of expertise (Glaser, 1984) and knowledge systems (Brown, 1982) guide this approach in which the question of structural stages is of lesser significance. Aside from an interest in the aging of the “content-free” mechanics of intelligence (Kliegl & Baltes, in press), the dominant focus is on changes in systems of factual and procedural knowledge associated with the “crys- tallized” pragmatics of intelligence. The crystallized form of intelligence was highlighted already as an ability cluster exhibit- ing stability or even positive changes into late adulthood. The concept of crystallized intelligence, however, needs further ex- pansion to cover additional domains of knowledge and to per- mit consideration of forms of knowledge more typical of the second half of life.
Expertise, a concept currently in vogue in cognitive and de- velopmental psychology (Ericsson, 1985; Glaser, 1984; Hoyer,
LIFE-SPAN DEVELOPMENTAL PSYCHOLOGY
Different Forms of Intelligence
Examples: Language Social Intelligence /
Intelligence as Cultural KnowledgeExamples:
Memory Problem Solving (Symbols, figures)
Intelligence as Basic Information Processing
ca. 25 ca. 70 Age
Figure 1. One of the best known psychometric structural theories of intelligence is that of Raymond B. Cattell and John L. Horn, (The two main clusters of that theory, fluid and crystallized intelligence, are postulated to display different life-span developmental trajectories.)
1985; Weinert, Schneider, & Knopf, in press), can be used to illustrate this avenue of research. This concept denotes skills and knowledge that are highly developed and practiced. The sciences, chess, or job-related activities (e.g., typing) are often used as substantive domains in which performance in relation to the acquisition and maintenance of expertise is studied. The life course of many individuals most likely offers opportunity for the practice of such forms of expertise. Therefore, one would expect that further growth of intellectual functioning may occur in those domains in which individuals continue to practice and evolve their procedural and factual knowledge (Denney, 1984; Dixon & Baltes, 1986; Hoyer, I9S5; Rybashet al., 1986). In other words, expertise in select facets of the prag- matics of intelligence can be maintained, transformed, or even newly acquired in the second half of life if the conditions are such that selective optimization in the associated knowledge system can occur.
Two areas of knowledge have been identified as key candi- dates for domains in which the pragmatics of intelligence may exhibit positive changes during the second half of life: practical intelligence (Sternberg & Wagner, 1986) and knowledge about the pragmatics of life, such as is evident in research on social intelligence and wisdom (Clayton & Birren, 1980; Dittmann- Rohli & Baltes, in press; Holliday & Chandler, 1986; Meacham, 1982). Wisdom, in particular, has been identified as a prototypi- cal or exemplar task of the pragmatics of intelligence that may exhibit further advances in adulthood or whose origins may lie primarily in adulthood.
In one approach (Dittmann-Kohli & Baltes, in press; Dixon & Baltes, 1986; Smith, Dixon, & Baltes, in press), wisdom is defined as an “expertise in the fundamental pragmatics of life.” In order to capture wisdom, subjects of different age groups are asked, for example, to elaborate on everyday problems of life planning and life review. The knowledge system displayed in protocols of life planning and life review is scored against a set of criteria derived from the theory of wisdom. First evidence (Smith et al., in press) suggests, as expected by the theory, that some older adults indeed seem to have available a well-devel- oped system of knowledge about situations involving questions of life planning. For example, when older adults are asked to explore a relatively rare or odd life-planning situation involving other older persons, they demonstrate a knowledge system that is more elaborated than that of younger adults.
The study of wisdom is just beginning. Thus, it is still an open question whether the concept of wisdom can be translated into empirical steps resulting in a well-articulated psychological the- ory of wisdom. We are somewhat optimistic, inasmuch as cog- nitive psychologists are increasingly studying tasks and reason- ing problems that, like the problems to which wisdom is ap- plied, have a high degree of real-life complexity, and whose problem definition and solution involve uncertainty and rela- tivism in judgment (Dorner, in press; Neisser, 1982). In the pres- ent context, the important point is that research on wisdom is an illustration of the type of knowledge systems that develop- mental, cognitive researchers are beginning to study as their at- tention is focused on later periods of life.
616 PAUL B. BALTES
Intellectual Development as a Dynamic Between Growth (Gain) and Decline (Loss)
The next belief (see Table 1) associated with life-span work is the notion that any process of development entails aspects of growth (gain) and decline (loss). This belief is a fairly recent one and its theoretical foundation is not yet fully explored and tested.
The gain/loss argument. The gain/loss view of development emerged primarily for two reasons. The first relates to the task of defining the process of aging in a framework of development. The second reason deals with the fact of multidirectionality de- scribed earlier and the ensuing implication of simultaneous, multidirectional change for the characterization of develop- ment.
Concerning the issue of the definition of aging versus develop- ment: Traditionally in gerontology, there has been a strong push—especially by biologists (Kirkwood, 1985)—to define the essence of aging as decline (i.e., as a unidirectional process of loss in adaptive capacity). Behavioral scientists, because of their findings and expectations of some gains in old age, have had a tendency to reject this unidirectional, decline view of aging. Thus, they wanted to explore whether aging could be considered as part of a framework of development.
How could this integration of aging into the framework of development be achieved in light of the fact that the traditional definition of development was closely linked to growth, whereas that of aging was linked predominantly to decline? The sugges- tion of life-span researchers was to redefine or expand the con- cept of development beyond the biological conception of growth or progression. Specifically, the proposal was to expand the con- cept of development to include not only phenomena of growth (gain), but other directions of change as well (Bakes, 1983). As a result, development was defined as any change in the adaptive capacity of an organism, whether positive or negative. In devel- opmental psychology, this proposal was consistent with other trends. Social learning theory, for example, had suggested a sim- ilar expansion of the directional nature of ontogenesis (Ban- dura, 1982).
The second and related reason for the emergence of the theme of gain-loss relations in development is a further elaboration of the notions of multidimensionality and multidirectionality of life-span intelligence. The separate concern for multiple-ability systems associated with multidirectionality was taken one step further to the examination of the relation and perhaps even the dynamic interplay between the various subsystems. The ques- tion guiding this investigation is whether the occurrence of mul- tidirectional change, concurrently in separate components of the system (e.g., fluid vs. crystallized intelligence), requires a new conception of development itself. One possible avenue is to view development as a gain-loss relation (see also Baltes & KJiegl, 1986; Labouvie-Vief, 1980, 1982; Perlmutter, in press). According to this view, development at all points of the life course is a joint expression of features of growth (gain) and de- cline (loss). It is assumed that any developmental progression displays at the same time new adaptive capacity as well as the loss of previously existing capacity. No developmental change during the life course is pure gain.
The view of development as a gain/loss phenomenon, of course, does not imply that throughout life gain and loss exist
Life-span Development: Gain/Loss Ratios in Adaptive Capacity
c 3 O
I JO 05
Birth Old Age Figure 2. One theoretical expectation concerning the average course of gain/loss ratios is a proportional shift across the life span.
in equal strength. Systematic age-related changes in the gain/ loss proportion are likely to be present. A possible life-span sce- nario of the dynamic between gains and losses is illustrated in Figure 2. Summarizing evidence across a wide spectrum of functions, the proposition contained in Figure 2 is that the sum total of possible gains and losses in adaptive capacity shifts pro- portionally with increasing age. As a whole, life-span develop- ment proceeds within the constraints created by this dynamic.
The concern with an ongoing, developmental dynamic of positive (gains) and negative (losses) change has spurred new research in life-span work. One example is the attempt to spec- ify a general process of adaptation that would represent the life- long nature of development as a gain/loss relation. Some of this work (M. Baltes, 1987; Baltes et al., 1984; Dixon & Baltes, 1986) has outlined a theoretical framework that is aimed at making explicit the dynamic relation between gain and loss in development. In Table 2, what is described for the case of cogni- tive aging is perhaps a prototypical change mechanism of “suc- cessful aging”—selective optimization with compensation.
The process of selective optimization with compensation has three features, each indicative of a gain/loss relation: (a) contin- ual evolution of specialized forms of adaptation as a general feature of life-span development; (b) adaptation to the condi- tions of biological and social aging with its increasing limitation of plasticity; and (c) individual selective and compensatory efforts dealing with evolving deficits for the purpose of life mas- tery and effective aging.
Although the process of selective optimization and associated investments in terms of time, space, and behavior budgeting is assumed to be general and prototypical, its specific manifesta-
LIFE-SPAN DEVELOPMENTAL PSYCHOLOGY 617
Table 2 Selective Optimization With Compensation: A Process Prototypical of Adaptive Life-Span Development of Cognitive Functioning
• A general feature of life-span development is an age-related increase in specialization (selection) of motivational and cognitive resources and skills.
• There are two main features of the aging of cognitive functions: fa) The reserve capacity for peak or maximum performances in
fluid functioning (mechanics of intelligence) is reduced, (b) Some procedural and declarative knowledge systems (pragmatics
of intelligence) can continue to evolve and function at peak levels.
• When, and if, limits (thresholds) of capacity are exceeded during the course of aging for a given individual, the following developmental consequences result: (a) Increased selection (channeling) and further reduction of the
number of high-efficacy domains; (b) Development of compensatory and/or substitute mechanisms.
Note. This model is adapted from Bakes and Baltes (1980).
tion will vary depending on the individual life history. The emergence of compensatory and substitutive cognitive skills will also vary. For example, B. F. Skinner (1983) described in a personal account how and under which conditions he evolved compensatory or substitutive skills in the face of reduced effec- tiveness in select aspects of intellectual functioning.
Research by Salthouse (1984) on typing skills in old age is one persuasive example of how specialization and compensa- tion interact in older adults to produce a high level of selective efficacy. Older expert typists—although showing less efficient reaction times than young expert typists when confronted with the typing of individual characters—nevertheless display good efficacy. They cope with their loss in reaction time when typing individual letters by developing more extensive forward-pro- cessing of letter and word sequences. As a result, older typists use a different combination of component skills than younger typists to produce a comparable level of overall performance.
Relevance of the gain/loss argument for other developmental Questions. For historians interested in the study of social change and the idea of progress (Nisbet, 1980), the gain/loss argument is a truism. Few theorists would maintain that all of social change is progress.
What about ontogenesis? The notion that any ontogenetic change entails a dynamic interplay between gain and loss can be found also in biological conceptions such as Waddington’s (1975) work on ontogenesis as differentiation and specialization (canalization). Beginning on the cellular level (Cotman, 1985; Lerner, 1984), differentiation clearly implies the loss of alter- nate courses of cell differentiation. Sociologists, too, have ar- gued that the life course entails a process of specialization in the sense of commitment to and practice in selected domains of life (Featherman, 1983). Such a process of life-course special- ization is assumed to imply a loss of alternative options.
Does the consideration of gains and losses as intrinsic to any developmental process have relevance for cognitive develop- ment in younger age groups? The fact that progress is not the only feature of age-related cognitive development has been known for quite some time. Weir’s (1964) research with chil-
dren on the developmental progression from maximization to optimization strategies in solving nonperfect probability tasks is a concrete example. This work shows that higher-level cogni- tive functioning (associated with the so-called optimization strategy) can have its costs when the task in question has no perfect solution. Specifically, when a cognitive task is logically unsolvable, because it has no perfect solution, young children can outperform older children and adults. This is so because older children and adults assume the existence of a logically per- fect solution and, therefore, engage in a problem-solving behav- ior that is criterion-inadequate.
One additional example is Ross’s (1981) research on the de- velopment of judgmental heuristics. Developmentally later or more “mature” heuristics can be less efficient, depending on the problem to be solved. Another case is second-language learning. Although there is disagreement about the sources of this gain/ loss phenomenon, the fact seems accepted: The more and more efficient acquisition of a first language is associated with in- creasing difficulty in learning a second language (Davies, Griper, &Howatt, 1984; Kellerman & Smith, 1986).
The gain/loss idea is also conlained in the work of Piaget, even though Piagetian theory is likely seen by most as the hall- mark of a conception of development that features unidirec- tional growth. In the study of the age-development of visual illu- sions, for example. Piaget (1969) described illusions that in- crease with age and others that decrease with age. When visual illusions increase with age, Piaget ascribed this loss in visual accuracy to advancement in cognitive stage. A related example in Piaget’s work (see Chapman, in press) is what he called the “repression” effect associated with a dynamic between percep- tual and cognitive operations. With children of 7 to 8 years of age, Piaget found that their “veridical” perceptual knowledge was repressed by the (in this case, nonveridical) advancement in conceptual schemata. In effect, cognition “repressed” per- ception. A by-product of cognitive advancement, then, was a loss in veridical judgment involving perception.
It seems worthwhile to explore with greater force the idea that any process of cognitive development entails positive and negative aspects of change in adaptive capacity. In the long run, the treatment of this question is likely to be tied to the topic of adaptive fitness. In other words, as is true for evolutionary theories (as exemplified in the work of Gould and Lewontin [Lerner, 1984]), ontogenesis may not fundamentally be a pro- cess of general gain in the sense of a total increase in adaptive efficacy. As specific forms of mind and behavior are “selected” during ontogenesis for activation and growth, some adaptive ca- pacities may actually become reduced. Whether the reduction is manifest would depend (as is true for evolutionary change) on the criterion demands posed by the individual and the envi- ronment in subsequent phases of life, or by the experimenter for that matter.
Plasticity of Development
Another belief (Table 1) held by many life-span scholars is that there is much plasticity in the course of life-span develop- ment. Plasticity refers to within-person variability and desig- nates the potential that individuals have for different forms of behavior or development (Gollin, 1981; Lerner, 1984). Would the same individual develop differently if conditions were
618 PAUL B. BALTES
different? The question of plasticity is, of course, one that has excited developmental psychologists for a long time (e.g., Mc- Graw’s  work in the 1930s on motor development in iden- tical twins). It is of equal interest to current work in develop- mental biology (e.g., Cotman, 1985) and developmental sociol- ogy (Dannefer & Perlmutter, 1986; Featherman & Lerner, 1985).
Cognitive training research with the elderly. Whether it proves to be true, the conclusion that the plasticity of cognitive age differences found in the second half of life is larger than what is known from childhood research has impressed life-span researchers. The salience of the topic of plasticity for research- ers in life-span development and the general approach used to study it can be illustrated easily in work on intellectual develop- ment during adulthood and old age. Initially, in life-span re- search the idea of plasticity was promulgated in connection with the question of whether intellectual aging is a simple pro- cess of gradual decline (Baltes & Schaie, 1976; Horn & Donald- son, 1976). One of the attractive ways to explore this question was to conduct intervention research.
Since the early 1970s, various research programs have been underway to examine the extent to which the observed aging losses in that cluster of psychometric intelligence evincing most definite decline—fluid intelligence—could be simulated for by performance variation (plasticity) within individuals. For ex- ample, elderly people were given practice in solving the type of intelligence problems used to measure fluid intelligence (Baltes & Kliegl, 1986; Willis, 1985). The key research hypothesis was that older persons, on the average, have relatively little test expe- rience and everyday practice in fluid intelligence, but that they possess the reserve—the latent competence—to raise their level of performance on fluid intelligence tasks to that of younger adults.
The cognitive training research conducted with older adults offered strong evidence of sizeable plasticity. After a fairly brief program of cognitive practice, many older adults (age range, 60-80 years) exhibited levels of performance comparable with those observed in many “untreated” younger adults. Such findings have been replicated consistently and in other domains of cognitive functioning (e.g., Denney, 1984; Labouvie-Vief, 1985; Willis, 1985). Meanwhile, the evidence has been extended from cross-sectional to longitudinal age comparisons (Schaie & Willis, 1986).
These studies illustrate the emergence of the strong belief of life-span researchers in sizeable plasticity. Knowledge about plasticity of intelligence was judged to be as important as knowledge about the average course of age development. In or- der to achieve a comprehensive understanding of a given devel- opmental process, such as intellectual development, one must forcefully advance the study of conditions that produce differ- ences in development and highlight the potential for alternate courses of development. The resulting interpretative posture is that whatever one observes concerning the aging of intelligence is but one of many possible outcomes (Brandtstadter, 1984; Ler- ner, 1984). Knowing the range and limits of intraindividual functioning, therefore, is a cornerstone of the life-span perspec- tive.
From plasticity to limits of plasticity. Current work on plas- ticity of life-span intellectual development has added new per- spectives beyond those associated with cognitive training re-
search. The focus has changed from demonstration of plasticity toward using research on plasticity as a strategy by which limits and boundaries of development can be identified (Kliegl & Baltes, 1987). This strategy is similar to child research on the “zone” of proximal development in childhood (Brown, 1982; Ferrara, Brown, & Campione, 1986).
The resulting focus in the life-span study of plasticity is not on the normal range of intellectual functioning, but on limits of performance. The research strategy chosen to examine different aspects of plasticity and its boundary conditions is known as testing-the-Hmits (M. Baltes & Kindermann, 1985; Guthke, 1982;Wiedl, 1984). Testing-the-limits involves the sys- tematic application of (a) variations in modes of assessment, (b) methods of intervention aimed at identifying latent reserve capacity, and (c) strategies of identification of the mechanisms involved in growth and decline.
In our research efforts (Baltes et al., 1984; Kliegl & Baltes, 1987) intended to determine age-correlated changes in limits of intellectual functioning, three aspects of plasticity are distin- guished: (a) baseline performance, (b) baseline reserve capacity, and (c) developmental reserve capacity. Baseline performance indicates a person’s initial level of performance on a given task, that is, what a person can do in a specified task without inter- vention or special treatment. Baseline reserve capacity denotes the upper range of an individual’s performance potential, when, at a given point in time, all available resources are called on to optimize an individual’s performance. It is measured by tests of “maximum” performance. When conditions have been added that strengthen an individual’s baseline reserve capacity through intervention (or development), we speak of an individ- ual’s developmental reserve capacity.
Distinguishing between these three aspects of plasticity per- mits the study not only of plasticity but also of constraints (Keil, 1981) and of what behavior geneticists have called the norm of reaction (see Lerner, 1984). Concepts such as constraint and norm of reaction are intended to index those limits—biological and sociocultural—that restrict the formation of a given behav- ior and its open development. In the long run, assessment of maximum reserve capacity aspires to identify biological bound- aries of the plasticity of development. Note, however, that the range and level of maximum reserve capacity is inherently un- knowable (see also Keating & MacLean, in press); it can only be approximated. The possibility always exists that new condi- tions or agents may be found that produce new levels and forms of intellectual performance on a given task.
The general expectation of testing-the-limits research is to predict developmental differences to be most pronounced, per- haps even approaching the condition of irreversibility (Wohl- will, 1973), near maximum levels of performance. Conversely, age changes are easily masked or modified as long as they are studied within the “normal” range of functioning. Figure 3 il- lustrates how a testing-the-limits strategy is used to obtain in- formation about the range and limits of plasticity during adult- hood.
The data presented in Figure 3 are based on subjects of differ- ent adult age groups who participated in extensive longitudinal intervention studies. In these studies, high levels of cognitive functioning in the use of a mnemonic skill are engineered in the laboratory. Specifically, subjects participated in 30 sessions of training of expert memory for digits and nouns using the Method
LIFE-SPAN DEVELOPMENTAL PSYCHOLOGY 619
o • ” * •
oU l !/> Q 0£ O 3: u. O z