Bornstein_1989_sensitive Periods In Development

Psychological Bulletin 1989, Vol. 105, No. 2, 179-197

Copyright 1989 by the American Psychological Association, Inc. 0033-2909/89/$00.75

Sensitive Periods in Development: Structural Characteristics and Causal Interpretations Marc H. Bornstein National Institute of Child Health and Human Development and New \brk University The presence or absence of a particular experience at a particular time in the life cycle may exert an extraordinary and dramatic influence over structure or function well beyond that point in development. Such sensitive periods are thought to be widespread in animal and in human neurobiology and psychology. A comprehensive treatment of the sensitive period needs minimally to include information about its structural characteristics as well as an interpretation of its causes, including why the sensitive period arises in terms of the natural history of the species and how the sensitive period is regulated in terms of physical, physiological, and psychological processes. This article provides a framework for research and theory concerning specific sensitive periods and the sensitive period generally conceived. The framework delimits four sets of parameters, which encompass 14 structural characteristics that define sensitive periods, and two levels of causal interpretation that guide research and theory into sensitive periods however they may be manifested.

Mary had a little lamb Its fleece was white as snow, And everywhere that Mary went The lamb was sure to go.

Sensitive periods are meaningful in psychology, and particularly in developmental psychology, for several reasons. First, sensitive periods give concrete evidence of the continuing control that experience may exert on development and on the nature of mature structure and function. Relatedly, the influence of experience displaced over time testifies to the force of coding and retention of information in memory. Second, the fact that sensitivity to experience waxes and wanes through the life span illustrates clearly the interactional character of development. Third, manifestations of sensitive periods in development supply convincing testimony to the existence of discrete phases in the life course. Last, that specifiable early experiences may have far-reaching consequences in ontogeny renders the sensitive period of potentially great practical significance. For these reasons, presumably, sensitive periods have been the subject of popular appeal, theoretical fascination, and empirical scrutiny since their earliest discovery. Sensitive periods were first identified in ethology and in experimental embryology: Chicks hatched in the absence of a hen were observed to follow any moving object to which they were exposed in the first days after hatching, but not later (Spalding, 1873). Certain cell masses were found to be affected by intruding chemicals during a particular stage in their ontogeny but not earlier or later, and certain cells transplanted at a particular time in their ontogeny were found to assume the characteristics of host location cells and thrive, but to wither if transplanted at times preceding or following (Spemann, 1938; Stockard, 1907, 1921). Not long after these initial reports, sensitive periods came to be implicated in a plethora of fields germane to psychological inquiry: At different times, sensitive periods have been seen as relevant to studies of brain and bodily structure and function across the phylogenetic series as well as in studies of survival, social, and mental competence in infrahuman animals and in human beings.

Scientists who study structure or function from all but an entirely static perspective inevitably confront the truism that dynamic animate phenomena are shaped by endogenous and exogenous forces interacting through time. Those forces do not exercise equal effects at all times, however; indeed, it is now common to speak of unique phases in the ontogeny of many different structures and functions when evolving transactions among life forces profoundly influence development. These phases are unique in that during select times in the life cycle many structures and functions become especially susceptible to specific experiences (or to the absence of those experiences) in a way that alters some future instantiation of that (or a related) structure or function. So, during such sensitive periods in development specific experiences may exert a marked influence over future history.

This research was supported by Research Grants HD20559 and HD20807 and a Research Career Development Award (HD00521) from the National Institute of Child Health and Human Development and Research Grant BNS 84-20017 from the National Science Foundation. Preparation of this article was partially supported by a grant from the Center of Developmental Education and Research, Tokyo, and was completed during my tenure as Visiting Professor, first at the Faculty of Education of the University of Tokyo and later at the Laboratoire de Psychologic du Developpement et de 1'Education de 1'Enfant of the Sorbonne. I thank H. Bornstein for comments. Correspondence concerning this article should be addressed to Marc H. Bornstein, Child and Family Research, National Institute of Child Health and Human Development, Building 31, Room B2B15, 9000 Rockville Pike, Bethesda, Maryland 20892.

179

180

MARC H. BORNSTEIN

Given so broad a sweep of theoretical interest and scope of experimental investigation, it would be natural to expect a certain heterogeneity of focus and perhaps of definition about sensitive periods. In fact, the typical investigative style adopted by researchers of sensitive periods historically has been to center attention on the intersection of a particular phenomenon with a particular species: sensorimotor connectivity in aplysia (e.g., Montarolo, Goelet, Castellucci, Morgan, Kandel, & Schacher, 1986), cocoon preference in ants (e.g., Jaisson & Fresneau, 1978), aggression in mice (e.g., Cairns, Hood, & Midlam, 1985), imprinting in ducks (e.g., Hess, 1973), cortical cell specificity in cats (e.g., Hubel & Wiesel, 1970), sociability in dogs (e.g., Scott, 1962), emotionality in monkeys (e.g., Harlow & Harlow, 1969), or language acquisition in human beings (e.g., Lenneberg, 1967). This orientation resulted inevitably in a remarkably diverse and disparate, if not daunting, experimental literature, as noted by several previous reviewers (e.g., Bateson, 1979; Bornstein, 1987; Colombo, 1982; Immelmann & Suomi, 1981; Oyama, 1979; Scott, 1978). However, two central (and consensual) conclusions of individual investigations and of collective reviews alike are, first, that sensitive periods reflect idiosyncratic and multifaceted biological, psychological, and sociological phenomena and, second, that the sensitive period is not a monistic concept. Investigation into a sensitive period seems normally to traverse two stages. Initially, the sensitive period undergoes descriptive treatment, followed by experimentation organized to quantify its several parameters. The full details of individual sensitive periods are never unabridged however, since empirical examination tends usually to focus on single parameters. Moreover, the concept of the sensitive period per se has failed to reach a high level of theoretical development. Specifically, individual investigators of sensitive periods have been reluctant to acknowledge an underlying unified framework from which to approach the sensitive period. This conceptual shortfall is, it would seem, the unhappy by-product of two conditions. First, the sensitive period has two distinct construals: In one, normative experience in the sensitive period is necessary for or advantageous to normal development, and in the other, noxious experience in the sensitive period may harm the organism. These separate interpretations are infrequently mixed in discussions of the sensitive period, and so tend to obscure a unified view. The second rationale supporting a pluralistic view of sensitive periods is that, as stated earlier, descriptions of sensitive periods, however detailed, and experiments on sensitive periods, however defining, have tended to concentrate on the individuality of particular sensitive periods. In addition to the overrich variety of individual types and the multiplicity of facets in any one type, reports of sensitive periods appear in widely distributed and nonoverlapping literatures. Working in highly circumscribed areas and confronting the near bewildering complexity of individual sensitive periods, researchers seem therefore to have actively eschewed (indeed, many have lobbied against) generalization. These main factors have tended to impede development of a perspective of the sensitive period as unified, with the inevitable result that the overarching issues of the structural character and causal interpretation of sensitive periods in general have been insufficiently addressed.

The fact remains that the sensitive period reflects a developmental event of impressive conceptual unity and coherence even across the diversity of biological, psychological, and sociological disciplines in which it has been studied. The sensitive period possesses a common set of characteristics that define its structure and invokes a common set of interpretations that pertain to its cause. The central goal of this review is, therefore, to construct a framework that makes explicit those structural characteristics and causal interpretations essential to understanding the deeper nature of this significant and ubiquitous developmental event. Comprehensive characterization is vital, if not imperative, to meaningful analysis, and a proper descriptive framework promises significant implications for both research on and theory about the sensitive period. The ultimate intentions of this review are, therefore, to guide future investigations of individual sensitive periods and by doing so to illuminate their nature as well as that of the sensitive period per se.

A Sensitive Period Framework: I. Structural Characteristics The first question that arises in identifying a sensitive period is descriptive, that is, what are its particular structural characteristics. (I use the phase structural characteristics to stand for the list of descriptive parameters of sensitive periods.) This is as it should be: The empirical characterization of parameters of the sensitive period is a first-order prerequisite to comprehensive and systematic analysis of its function or meaning. A second prerequisite, taken up later, is investigation about cause. In one notable theoretical overview, Nash (1978) considered four parameters of sensitive period structure, including onset, terminus, and intrinsic and extrinsic factors that specify, respectively, the triggering maturational event in the organism and the influential experiential event outside the organism. On the earlier suggestion of Moltz (1973), Colombo (1982) added a fifth characteristic to Nash's four, the organismic system that is affected by experience during the sensitive period. In actuality, however, the sensitive period is comprehensively described by at least 14 distinguishing, essential, and operationally definable structural characteristics. These 14 structural characteristics may be categorized into four distinctive sets. The first set describes temporal and intensive contours of the sensitive period when it occurs in the life cycle, and links five structural characteristics: (a) developmental dating, (b) onset, (c) offset, (d) duration, and (e) asymptote. The second set unveils the mechanisms involved in sensitive-period change, and incorporates three characteristics: (f) experience, (g) system, and (h) pathway. The third set examines consequences of the sensitive period, and unites four characteristics: (i) outcome, (j) manner, (k) outcome conditions, and (1) duration. The fourth set takes into account evolutionary and ontogenetic time scales with respect to sensitive periods, and involves two characteristics: (m) variability and (n) modifiability.

Structural Characteristics of the Sensitive Period In this section, I define and illustrate each of the 14 prominent structural characteristics of sensitive periods. This taxon-

SENSITIVE PERIODS IN DEVELOPMENT omy represents one way of organizing theory and research in sensitive periods. In selecting examples to represent particular sensitive periods, I have attempted to invest this scheme with a current of continuity by referring when possible to the two most popular subjects of sensitive-period study, namely visual sensitivity and imprinting. However, perception and social attachment are not the subjects of this review; no one kind of sensitive period is. In addition, I have elected in some cases to reference special or idiosyncratic examples of sensitive periods because they constitute particularly apt or telling instances and to draw upon contrasting or complementary examples in order to throw a particular issue into relief. The text also provides some numerical values; it must be understood, however, that quantitative estimates of sensitive-period parameters are commonly acknowledged to fluctuate with species, ontogeny, methodology, and other factors.' Finally, not all features of all structural characteristics in all systems have been studied; indeed, not all structural characteristics of any one system have been exhaustively researched. This review is perforce illustrative. For purposes of this discussion, consider a system to specify the structure or function that is the main subject of sensitiveperiod analysis. What dimensions underlie sensitive-period reformation?

Temporal and Intensive Contours Five characteristics demarcate the profile of the sensitive period when it occurs. They concern timing and change associated with sensitive periods.

Developmental Dating When in the life cycle of the system the sensitive period occurs, its frequency, and quantification of its temporal component. Developmental dating of sensitive periods is requisite to first-order characterizations. Between eye-opening and about 3 months later, the kitten visual system is uniquely susceptible to visual experience, yet outside this period environmental exposure has nowhere near the equivalent effect (Sillito, 1983). In maturity, male zebra finches raised early in life by Bengalese finches almost exclusively court Bengalese finches, even if they are in later life exposed extensively to conspecific females (Immelmann, 1972). Two principal questions arise with respect to developmental dating: first, When in the life course of a system does the sensitive period occur? and, second, How often does the sensitive period repeat? A third question is how dating of the sensitive period is most meaningfully calculated. Most investigators believe that sensitive periods are standard, if not circumscribed, to early development, and many suppose that they are once-in-a-lifetime events (e.g., Bateson, 1979; Boothe, Vassdal, & Schneck, 1986; Brook, 1972; Caldwell, 1962; Colombo, 1982;Pettigrew, 1978). Much sensitive-period theory and research are predicated on these opinions. Following Child (1921), Scott (e.g., 1986) proposed the general principle that systems would be most sensitive to modification of their developmental trajectory during periods of their most rapid (re)organization, as in growth. Moreover, many sensitive-period alterations, such as adapting perceptually to the physical nature

181

of the local environment or attaching socially to a source of nurturance, ought logically and naturally to be fixed to infancy and ought also to transpire only once in the life span. In implicit support of this belief, some sensitive periods are correlated with specific developments that take place in early life and only once, as at organogenesis, at first eye-opening, or at the onset of locomotion. It is also the case, however, that sensitive periods in infancy have simply been investigated most often. Exceptions show that sensitive periods can also arise late in development: The sensitive period for sexual imprinting in maturing zebra finches demonstrates that sensitive periods are not limited to the newborn period (Bateson, 1979), as does the sensitive period for maternal responsiveness to young of the species in goats (Gubernick, 1981; Klopfer, Adams, & Klopfer, 1964), in dairy cattle (Hudson & Mullord, 1977), and in sheep (Poindron, Levy, & Krehbiel, 1988; Poindron, Martin, & Hooley, 1979). Although sensitive periods are thought normally to occur only once in the life-cycle, some sensitive-period alterations in structure and function can be expected to take place more often. Indeed, this would be expected for sensitive periods for maternal responsiveness in multiparous goats, cattle, and sheep for identification of individual young. Apparently, then, some sensitive periods may occur at sundry points in the life cycle and may recur numbers of times. How is developmental dating of the sensitive period better calculated, quantitatively or qualitatively? In terms of real time, in terms of developmental time in the life cycle of the system, or in terms of a specifiable developmental event? Is developmental timing of the sensitive period better specified in terms of conceptional age or in terms of postnatal age? In practice, most investigators use postnatal age in dating sensitive periods: Contact comfort needs to be available between the period of birth and 250 days to ensure normal social behavior in the rhesus monkey (Harlow & Harlow, 1969); normal song must be heard in the period between 10 and 50 days for normal song to develop in the song bird (Marler, 1970); familiarization with humans must obtain in the period between 18 and 63-91 days for socialization in dogs (Freedman, King, & Elliot, 1961); primary socialization must be incurred in the period between 20 and 45 days to eventuate in domestication in the silver fox (Belyaev, Plyusnina, & Trut, 1985); social isolation must be instituted in the period between 25 and 45 days for a heightened level of object contact in rats (Einon & Morgan, 1977); shelter must be secured in the period between 30 and 60 days to maintain reactivity to stimulation in the gerbil (Clark & Galef, 1979); and exposure to the visual environment must follow in the period between 30 and 90 days for modification of the functional organization of the visual cortex in the cat (Sillito, 1983). Of course, some cases mandate dating by conceptional age, 1

For example, methodological regimen influences quantitative values: Presson and Gordon (1979) report that on one methodology Blakemore and Van Sluyters placed the end of the sensitive period for binocularity in the kitten at about 8-10 weeks, that on a second Hubel and Wiesel placed it at about 12 weeks, and that on a third Timney placed it at about 16 weeks.

182

MARC H. BORNSTEIN

as when prenatal experience is critical. Hill and Przekop (1988) found that if mother rats are put on a low-sodium diet on or before 8 days postconception, their pups fail to develop taste sensitivity to sodium, but mothers who begin a low-sodium diet 10 or more days after conception give birth to normal pups. Brook (1972) provided evidence that the sensitive period for adipose-cell replication in humans to hormonal and nutritional circumstances extends from approximately 30 weeks gestation to the age of about 1 year. When choice exists in dating, controversy appears. Opinions differ, for example, with respect to computing the time of the sensitive period in imprinting in ducklings. Hess (1959) calculated the sensitive period for imprinting in terms of age in real time after hatching, whereas Gottlieb (1961) later calculated the same sensitive period in terms of age in real time after conception, claiming that such a representation reduces variation in temporal parameters of the sensitive period. Landsberg (1975) showed that the two age-dating systems contributed information valuable to understanding the ontogenetic nature of the sensitive period: developmental age for endogenous influences and posthatch age for experience-related effects. Some researchers have dated sensitive periods in terms of actual life-span events instead of age. For example, undernutrition must be endured for the period of suckling to have lasting effects on social behavior in rats (Byrne & Smart, 1980); exposure to nonmimetic eggs must fall during the period of egg laying for rejection in orioles (Rothstein, 1978); separation from the lamb immediately after parturition engenders rejection by maternal ewes (Collias, 1956); and dogs' learning of territorial boundaries coincides with the advent of sexual maturity (Tinbergen, 1951). A vulnerable period in human brain development has been postulated to coincide with the maturational spurt of axons and dendrites, glial multiplication, and myelination (Dobbing & Sands, 1973). In the same vein, Almli and Finger (1987) argued that dating by neurodevelopmental landmarks (e.g., myelination) orders the literature on sensitive periods for behavioral recovery from neural insult, which is especially useful in cross-species comparisons. In point of fact, however, both quantitative and qualitative dating systems share the same developmental events—conception or birth—as reference points, only one uses chronological time as a kind of postevent marker. It may be that researchers resort to chronological specification when, for some reason, other critical developmental events cannot be specified.

in real time, and most agree that the onsets of sensitive periods tend to be quite rapid. For example, the onset of the sensitive period for suceptibility to noise-induced cochlear damage in the mouse lasts 5 days, or less than 1% of the mouse's 527-day actuarial life span (Henry, 1983). Likewise, the onset of the sensitive period for binocularity has been reported to begin "suddenly" in the cat, coincidental with eye-opening (Hubel & Wiesel, 1970, p. 434); the kitten's rise in sensitivity has been submitted to parametric scrutiny and found to dawn between Day 9 and Day 13 (Freeman, 1978). Systems do not always or necessarily reach full susceptibility immediately at onset, however. For example, kittens who experience monocular deprivation at the start of the sensitive period may lose binocularity, but apparently loss related to deprivation at that time is only transient, because binocular experience in the balance of the sensitive period can lead to a significant degree of recovery (Olson & Freeman, 1978). In contrast with prevailing opinion about sensitive-period onsets, many researchers theorize that offsets of sensitive periods are typically gradual (e.g., Hinde, 1962). For example, the onset of the sensitive period for human binocular vision is believed to endure over a period of about 1 month, whereas the offset is believed to taper over about 5 years (Banks, Aslin, & Letson, 1975). Likewise, onset of the sensitive period for filial imprinting in the zebra finch is confined to a brief point in early life, whereas its offset is reported to linger into adolescence (Immelmann&Suomi, 1981). Several exceptions limit strict generalization, however. Hill and Przekop (1988) found the transition from maximal environmental susceptibility to no susceptibility for the fetal rat gustatory system to last only 2 days. Beatty, Dodge, Traylor, and Meany (1981) delimited the temporal period when exposure to testicular hormones affects development of rough play in male rats; castration on Days 1 or 6 reduced male play, but castration on Days 10 or 20 did not, thus circumscribing the offset of the sensitive period for demasculinizing effects of testicular androgens to a maximum of 4 days' time. Buisseret, Gary-Bobo, and Imbert (1982) assessed the offset of the sensitive period for orientation selectivity in kitten and found it to end abruptly in the rather narrow window of postnatal weeks 10-12. Similarly, Elberger (1984) investigated the role of the corpus callosum in visual development in kittens and found that animals whose posterior corpus callosum was sectioned in Week 3 showed deficits in visual acuity whereas those sectioned in Week 4 (or after) did not.

Onset and Offset The rise and decay of sensitivity in terms of setting event and time. What events govern the start and finish of the sensitive period, when and how do they come into play, and how are they described? Are onset and offset better characterized quantitatively or qualitatively? Are they better characterized in terms of real time, in terms of developmental time as a proportion of the life-cycle of the system, or in terms of a specifiable developmental event? Are the growth and decay slopes of sensitivity to and from the peak step functions or are they graded, and are they linear? Researchers usually specify the onset of the sensitive period

Duration The temporal window of susceptibility. By definition the sensitive period endures within the confines of its onset and offset, and its duration is among the most frequently described structural characteristics. How long does the system remain sensitive? The window of vulnerability is often an identifying characteristic of sensitive periods, even when little if anything else is known about them. Thus, for example, sudden infant death syndrome shows a strong, time-locked course on the basis of

SENSITIVE PERIODS IN DEVELOPMENT which it has frequently drawn the attribution of a human sensitive period (e.g., Hunt & Brouillette, 1987). Most investigators specify the temporal duration of the sensitive period in terms of real time. Examples abound and show a wide variation in the window of susceptibility; some prominent ones corresponding to the examples of developmental timing cited above (but recast by amount of time) are the following: A ewe must label her lamb in a period of less than .03-.08 days (45 minutes-2 hours) after parturition to ensure acceptance (Collias, 1956; Pissonnier, Thiery, Fabre-Nys, Poindron, & Keverne, 1985); protein and messenger RNA (mRNA) must be synthesized in a period of .04-. 12 days (1-3 hours) for long-term facilitation of sensorimotor connection in aplysia (Montarolo et al., 1986); social isolation must be instituted in a period of 20 days for a heightened level of object contact in rats (Einon & Morgan, 1977); primary socialization must be incurred in a period of 25 days to eventuate in domestication in silver foxes (Belyaev et al., 1985); shelter must be secured in a period of 30 days to maintain reactivity to stimulation in gerbils (Clark & Galef, 1979); species-normal song must be heard in a period of 40 days for normal song to develop in song birds (Marler, 1970); familiarization with humans must obtain in a period of 46 to 74 days to socialize dogs (Freedman et al., 1961); exposure to the visual environment must follow in a period of 60 days for modification of the functional organization of the visual cortex in cats (Sillito, 1983); and contact comfort needs to be available in a period of 250 days to ensure normal social behavior in rhesus monkeys (Harlow & Harlow, 1969). Some researchers, although few in number, specify duration in terms of developmental events. For example, undernutrition must be endured for the period of suckling to have lasting effects on social behavior in rats (Byrne & Smart, 1980), and exposure to nonmimetic eggs must fall around the onset of laying to provoke rejection in orioles (Rothstein, 1978).

Asymptote The nature and degree of change in sensitivity and the uniformity of sensitivity during the sensitive period. Is the sensitivity threshold always altered to the advantage of the organism during the sensitive period? By how much is sensitivity reset during the sensitive period? Is sensitivity level during the sensitive period better described as a plateau or as a peak? If a peak, where in the temporal window does the peak occur? Is sensitivity a constant, or does it fluctuate during the sensitive period? Actually, asymptotes of sensitive periods are not normally described. Most investigators seem to presume that sensitivity level during the sensitive period holds constant, as for example sensitivity to cochlear damage in mice (Henry, 1983). However, some researchers report that sensitivity is peaked, as for example for cortical susceptibility in kittens (e.g., Blakemore & Mitchell, 1973), for filial imprinting in mallard ducks (e.g., Hess, 1959), for emotional behavior in dogs (e.g., Scott, Stewart, & DeGhett, 1974), for domestication in foxes (e.g., Belyaev et al., 1985), for social behavior in monkeys (e.g., Harlow & Harlow, 1969), and for adipose-cell replication in human beings (e.g., Brook, 1972), although precise definitions for "peaked" temporal sensitivities may be lacking.

183

Further, sensitive-period peaks are often presumed to center temporally with respect to duration, and to be unimodal. However, Brook (1972) suggests that the peak of adipose-organ sensitivity to nutrition effects in man is skewed to the onset of the sensitive period. Counter-examples may also be found to the assumption that sensitive-period peaks are unimodal. Henry and Bowman (1970) acoustically primed mice from two pure strains (C57BL/6J and DBA/2J) and found a sensitive period between 12 and 26 days during which priming induced susceptibility to audiogenic convulsions in one strain and enhanced susceptibility in the other. They also found that Fl hybrids manifested a bimodal sensitive period in which either parental strain would appear dominant depending on the age at which hybrids were acoustically primed. Schutz (cited in Scott, 1962) earlier reported a bimodal sensitive period for social attachment in ducks: one peak for attachment to particular individuals formed early in the sensitive period, and a second peak for attachment to the species formed late in the sensitive period. Basic to identifying and describing a sensitive period is delimitation of structural characteristics associated with its ontogenetic timing. They include developmental dating, onset and offset, duration, and asymptote. These parameters of a sensitive period are relatively straightforward and empirically definable. Nevertheless, many serious questions envelop parameters associated with ontogenetic timing of sensitive periods. For example, rates of change in sensitivity of a system during the temporal periods of the onset and offset are rarely studied; they may be linear (as is often presumed), or they may not. Mechanisms that trigger onset and offset are infrequently documented. Likewise, questions concerning the ontogeny, temporal characterisitics, and plasticity of the duration of a sensitive period often go unraised, as well as unaddressed. The asymptote in the sensitive period is a most intriguing yet underresearched structural characteristic: By definition, sensitivity is always raised in the sensitive period, but can sensitivity sometimes be lowered to protect the organism from the environment? Would diminished sensitivity still characterize a "sensitive period"? Clearly, specific sensitive periods will be better understood, as will the concept of the sensitive period more generally, when considerations of ontogenetic timing are treated comprehensively.

Mechanisms of Change Three characteristics define structures wherein the sensitive period is shaped: the parameters of the outer and inner worlds of the organism that are most immediately involved in the sensitive period and the linkage between them.

Experience The effective stimulus event (or nonevent) in the sensitive period, its specific nature, its physical characteristics, and its origins. Many investigators endeavor to provide reasonably precise data about effective stimulation for a sensitive period. Three questions arise immediately in considering the nature of the sensitive period's special experience; they include, first, the specificity of experience, second, dimensions and features of the experience, and, third, sources of the experience.

184

MARC H. BORNSTEIN

In his classic studies of the sensitive period, Stockard (1921) found that many different inorganic chemicals would induce monstrosities in fish embryos as long as they were applied at a specific time in ontogenesis. His results called into question the specificity of the stimulus experience per se in the sensitive period. Time has shown that specificity of experience itself has a great range. Some sensitive periods admit a spectrum of experience: During the sensitive period for cortical binocularity in kittens, visual axis misalignment or monocular deprivation will be equally effective in influencing cell function (Levitt & Van Sluyters, 1982); during the sensitive period for imprinting, gulls will form a social attachment to another gull or to a cage in which they are housed (Evans, 1970); and, during the sensitive period for maternal attachment, a ewe that licks any recently born lamb, not exclusively her own, will on this basis later distinguish that lamb from others (Poindron et al., 1988; Smith, Van-Toller, & Boyes, 1966). Other sensitive-period experiences tend toward greater specificity: Gottlieb and Klopfer (1962) pointed out that ducklings imprint earlier on auditory stimuli than on visual ones; during the sensitive period for song learning, white-crowned sparrows acquire the song of their own species, but will not acquire the song of other species (Marler, 1970); and, during the sensitive period for egg recognition, orioles reject brown-headed cowbird eggs from their nest, but do not reject eggs that differ from their own only in size (Rothstein, 1978). Still other experiences are mixed in terms of specificity: During the sensitive period for orientation selectivity in kittens, exposure to verticals or to horizontals maintains exclusively vertical or horizontal sensitivity, respectively, but exposure to diagonals maintains sensitivity to diagonal, vertical, and horizontal (Leventhal & Hirsch, 1975). During the sensitive period for imprinting in the duck, experience proceeds with the greatest facility with conspecifics, next best with similar species, and least well with objects (Immelmann, 1972). Specifying physical dimensions of the experience rewards investigators invested in closely depicting the nature of a sensitive period. A first set of parameters concerns stimulus duration. How long must the stimulus be present (or absent)? (This question is intimately tied to that of duration of the sensitive period discussed above.) During the critical period for orientation selectivity (between 3 and 14 weeks), kittens express greatest sensitivity around 28 days, and at this time as little as 1 hr of selective experience will completely transform the distribution of orientation-sensitive cells in the cortex (less time was not reported); 2 or 32 hr additional experience will not exert a significantly greater effect than 1 hr(Blakemore& Mitchell, 1973). During the critical period for imprinting in the mallard, as little as 4 min 49 s of exposure will lead to a preference lasting over 103 days (Schutz, 1972, cited in Immelmann & Suomi, 1981). Additional questions are, Must the experience be present continuously, or may it be interrupted? Must it cumulate? For what proportion of the duration of sensitivity is experience effective? Pettigrew, Olson, and Barlow (1973) found that very short-term repeated stimulation of neurons in kitten cortex induced changes that mimicked permanent modifications typical of sensitive periods, but these changes were less pronounced and endured only transiently. Zajonc, Reimer, and Hausser (1973) found that object attraction in 1-day-old chicks was a mono-

tonic function of the number of 30-min exposures the chicks had experienced. Suomi and Harlow (1975) found that 2-hr periods of peer interaction just 5 days a week could mitigate atypical social behaviors of surrogate-reared monkeys. A second set of parameters specifying physical dimensions of the experience concerns stimulus intensity. How intense must the stimulus experience grow to be effective? During the sensitive period for raising avoidance-learning scores of adult rats, Denenberg (1968) found that lower levels of shock (0.2 mA) were ineffective, whereas higher levels (0.5 mA) were effective. A natural conclusion might be that stimulation and effect are linearly related, but they need not necessarily be. For example, stimulus intensity could relate to effect in an inverted-U manner, as Denenberg suggests; in such cases, there may be "optimal levels" of stimulation associated with sensitive-period experiences. Other patterns of relations are also imaginable. A third set of parameters specifying physical dimensions of the experience concerns the sensory nature of the effective stimulus. For example, Gottlieb and Klopfer (1962) found different sensitive periods for visual versus for auditory stimuli in duckling imprinting. What are the principal sources of experience? In some cases answers are clear, whereas in others answers are more ambiguous. Gottlieb (1983) found that during a sensitive period in perception duck embryos must produce their own contact-contentment call in order to ensure a normal preference for their species maternal call after hatching. Reppert (1985) contended that the developing circadian clock is initiated in the suprachiasmatic nuclei of the anterior hypothalamus by the maternal host during a brief sensitive period in utero and is so maintained until the opportunity for direct (external) photic entrainment. The distinction between initiation and maintenance of the sensitive period with regard to source is a provocative one, and arises again in the interesting and complex case of the sensitive period for maternal bonding in ungulates (Poindron et al., 1988). This sensitive period appears to be induced in the mother by the rise in estradiol associated with the term of pregnancy and is augmented by genital (vaginal-cervical) stimulation accompanying labor and expulsion; it is then sustained by exteroceptive cues in the offspring, including amniotic fluid, as well as by physical contact between mother and young through licking and interaction (Keverne, Levy, Poindron, & Lindsay, 1983; Pissonnier et al., 1985; Poindron et al., 1988). Thus, sources of this sensitive period are in the mother, in the newborn, and in interaction between the two.

System The structure or function altered in the sensitive period. The assumption that "something is changed" during the sensitive period constitutes a central tenet of sensitive-period investigation and theorizing, but the task of pinpointing the system that is actually affected and defining its nature seriously challenges most investigators. Three questions that are regularly posed concerning the nature of that system include, first, the observable or nonobservable character of the system, second, its unitary or pluralistic configuration, and, third, the level at which the affected system functions.

SENSITIVE PERIODS IN DEVELOPMENT Is the system directly observable and measurable, or must the system be inferred from change that is observed and measured later? Most investigations of sensitive periods proceed on a "first manipulate developmental timing and experience and later assess outcome" modus operandi; in this scheme, evaluation of the affected system itself may or may not take place. Some investigators actually measure intervening systems affected. Rothblat and Schwartz (1979) found that the density of dendritic spines was influenced by monocular deprivation in the visual sensitive period in rats. However, most researchers seem not to measure a system directly, so that in most cases the system that is altered in the sensitive period must be inferred. This is so, for example, in the influence that gonadal hormones exert in early development on the eventual expression of sexually dimorphic behaviors in mammals (Reinisch, 1974). As a consequence, even the basic nature of targeted systems in sensitive periods is usually not well worked out. Is the system unitary or manifold? Most discussions of systems subject to sensitive-period analysis proceed as though the system were a singular construct. This assumption may be too facile (see Scott, 1986, for a similar suggestion). Very early on, Fabricius (1964) argued that imprinting actually involved three independent components for, respectively, eliciting the following response, the response itself, and the vigor of the response. Recently, Blass (1987) studied the sequence of developing social attachments during the nursing period in rats and showed that olfactory social bonding in this system is best understood from a "multiple" sensitive-periods perspective that includes four independent yet interrelated processes. At what level is the system actually being influenced? Sensitive period phenomena may obtain at one or more levels, including prominently the physical, physiological, and psychological, and it is often advantageous to specify to which level (or levels) conclusions reached about the nature of sensitive periods apply. When the kitten experiences a particular visual environment during its sensitive period, what, if any, are the respective physical, physiological, and psychological ramifications? Even if levels are not directly comparable, it seems necessary to indicate at which level or levels change occurs.

Pathway The channels by which experience affects the system during the sensitive period. Pathway designates the physical interface between experience and system and is to be distinguished from conceptual questions of how the sensitive period is regulated. Pathway is among the least often specified of sensitive-period parameters. Either chemical (hormonal) or neural (central nervous system) pathways have been invoked to describe concrete ways through which experience and system connect. Sometimes investigators are able to specify exact means; most times they have only been able to infer a means, depending usually on a correlation between time course and pattern of susceptibility between experience and outcome. Examples of chemical bases include the proposals that tubulin synthesis serves as a pathway of sensitive-period phenomena in kitten cortical visual selectivity (Cronly-Dillon & Perry, 1976), that testosterone serves as a pathway of sensitive-period

185

phenomena in rodent social play (Beatty et al., 1981), and that plasma corticosterone serves as a pathway of sensitive-period phenomena in filial imprinting (Landsberg & Weiss, 1976). Hirsch and Tieman (1987) argued that endogenous neurotransmitters help to determine whether activity-dependent changes eventuate in synaptic organization during sensitive periods in kittens. Indeed, Lauder and Krebs (1986) have pointed out that neurotransmitters and neurohumors regulate the temporal frameworks as well as sensitivities to external stimuli that are basic to virtually all sensitive periods. Examples of neural bases include proposals that maturation of the visual system in ducklings serves as a pathway in imprinting (Paulson, 1965), that olfactory cues serve as a pathway for North Pacific salmon to identify their natal tributary (Hasler, Scholz, & Horrall, 1978), that perceptual selectivity serves as a pathway in song learning (Marler & Peters, 1980), and that dams' olfactory sensitivity to their own labels on kids serves as a pathway for maternal responsiveness in goats (Gubernick, 1981). The sensitive period reflects responsiveness to some stimulus experience outside the organism, takes place at the site of some system inside the organism, and the stimulus and system connect via some pathway. These three material entities define a second group of structural characteristics fundamental to denoting a sensitive period. Nonetheless, many difficult questions central to experience, system, and pathway dog investigators of sensitive periods. For example, are the presence and absence of the stimulus equally or differentially meaningful to the sensitive period? More elusive—and troubling—is the question of whether the stimulus is effective in terms of the phenomenology of the organism or as it is defined by the experimenter. Having to infer the system and the pathway can be problematic as well, especially since the two are believed to play such central roles in the phenomenon of sensitive periods. For example, insofar as the system is altered, it and the functional outcome of the sensitive period (see below) could be one and the same, especially in the short-term. Without distinguishing the two, the experimenter must remain in the dark as to what was affected as opposed to how. Specification of the pathway has clear and profound implications for understanding the nature of the sensitive period as well, since manipulation of effects may depend on a knowledge of the route of those effects. The place of the pathway in sensitive-period study is therefore important, but subtle and complex, since even distinguishing the pathway from the system is made difficult when the pathway is somehow also altered as the sensitive period unfolds. For these reasons, laying bare material entities of sensitive periods is key to penetrating their nature.

Consequences The third group of parameters concerns later effects of sensitive periods. This set includes four structural characteristics associated with outcomes.

Outcome The consequence of experience sustained during the sensitive period in terms of the normative development of the species.

186

MARC H. BORNSTEIN

What physical, physiological, or psychological aspects of structure or function are influenced by experience during the sensitive period? Often, the outcome serves as the major source of evidence that experience in the sensitive period has in fact had a meaningful effect on the organism. Two sets of questions typically arise with respect to outcome: The first concerns the sphere of structure or function in which the effect manifests itself, and the second concerns direct or indirect implications of the outcome for the organism. Sensitive-period phenomena have been hypothesized, and observed, to exert profound effects, sometimes across more than one sphere of structure or function, at others within a single sphere. In visual system sensitive periods, outcomes may be observed in several spheres, as in physical process (Guillery & Stelzner, 1970), in physiological function (Wiesel & Hubel, 1963), or in psychological behavior (Ganz & Fitch, 1968; Wiesel, 1982). Lenneberg (1967) proposed a sensitive period for language learning in human beings, supposing the latitude of its outcome to be equally sweeping; in a close analysis of studies intervening in the score of years following Lenneberg's proposal, however, Snow (1987) demonstrated that empirical data did not support a broad application of the sensitive-period concept in language acquisition, but rather confirmed one limited in scope to articulated accent. More usually, outcomes are specific to a single sphere of structure or function. For example, Brook (1972) found that cell number (and not even cell size) in the adipose organ in humans was the target of a nutritional sensitive period. Moreover, more than a single effect in the same sphere may obtain; an extreme example is handling of rats during the sensitive period, which, beyond physiology (Levine & Lewis, 1959), has been observed to manifest itself in a plethora of behaviors, including learning (Denenberg & Bell, 1960), exploration (Forgays & Read, 1962), and emotional expression (Denenberg, 1963). The outcome needs further to be defined with respect to its direct or indirect implications for action (Hinde, 1970). For example, the sensitive period may function for learning a behavior necessary for the later performance of that behavior (as is the following response in imprinting), or it may be necessary for some other end (as is smiling for emotional development). Manner

How the influence of experience during the sensitive period affects the outcome. Experience in the sensitive period may influence the outcome in several ways. Experience may induce the outcome, stimulating the emergence of an effect that would not otherwise ensue; experience may attune the outcome, adjusting extant structure or function; or experience may maintain the outcome, continuing the status quo of the system (see Aslin, 1981; Gottlieb, 1983). Oyama (1976) illustrated induction in the acquisition of nonnative phonology in the sensitive period in language learning; by contrast, Wiesel (1982) proposed that innate mechanisms endow the visual system with highly specific connections and that visual experience early in the sensitive period is necessary for the maintenance and full development of cellular connectivity. Of course, experience may operate to influence outcome in more than one manner:

Interpreting Pettigrew's (1978) data on the role of the sensitive period in the ontogeny of binocularity, Aslin (1981) pointed out that experience may induce or attune development at the neurochemical level, whereas experience may maintain development at the physiological level. Outcome Conditions The temporal and spatial emergence of the outcome following onsets of the sensitive period and of experience. The interconnected critical questions that concern outcome conditions of the sensitive period usually include when and under what circumstances in development the influence of experience manifests itself relative to the onset of the sensitive period and relative to the onset of the experience itself. Is the outcome immediate or eventual? Mitchell and Timney (1986) reviewed the literature in visual sensitivity and showed that orientation selectivity and binocularity are affected very soon after the onset of the sensitive period. Marler (1970) reviewed the literature for song learning and showed that vocal production often does not emerge until months after the onset of the sensitive period. Immelmann and Suomi (1981) reviewed the literature for imprinting and showed that sexual preference does not usually emerge until years after the onset of the sensitive period. The composition of outcome conditions is not always equivalently responsive to similar experiences. Money and Annecillo (1987) observed that androgen injected into a pregnant ewe between gestational Days 35 and 50 masculinizes the external genital anatomy of the lamb, whereas androgen injection after Day 50 affects only the brain. Moreover, the short-term and longterm effects of equivalent sensitive-period experiences may be similar or they may differ. Bateson (1979) analyzed filial and sexual preferences in finches in this connection. He suggested that critical early experience in the sensitive period affects filial preference in the juvenile period, but exerts effects on sexual preference in adulthood. Finally, the outcomes of sensitive periods are not always first-order; they can be second-order as well. For example, Hirsch and Tieman (1987) pointed out that nervous system changes produced as a developmental outcome of the sensitive period will systematically influence behavior long after the sensitive period has closed.

Duration The temporal nature of the outcome of the sensitive period. How long does the outcome of the sensitive period naturally endure? As for all other parameters of the sensitive period, the temporal nature of the outcome doubtlessly depends on many factors, including which outcome is under scrutiny, in which species, and so forth. Outcomes have normally been presumed to be permanent or, at least, quite long-lived (e.g., Lorenz, 1937). Immelmann (1972) observed that the natural life expectancy of the zebra finch is less than 7 years. Zebra finches raised by Bengalese finches during the sensitive period for sexual imprinting have been known to prefer Bengalese finches sexually after more than 7 years; thus, stability of the imprint seems to rival the natural life expectancy of the species.

SENSITIVE PERIODS IN DEVELOPMENT The sensitive period is hardly meaningful if it has no consequent in later development or if its consequent is hidden or shrouded in mystery—hence the significance of specifying structural characteristics associated with antecedent-consequent relations, such as outcome, manner, outcome conditions, and duration. Even though antecedent-consequent relations must be among the sensitive period's most conspicuous characteristics, many unanswered questions persist about outcomes. For example, when the outcome is displaced in time from the sensitive period, its boundary definitions are clear; but how are system and outcome distinguished when the two are closely coordinate in space and time? Further, is an outcome normative for the species? Imprinting shows that the outcome of typical experience encountered during the sensitive period will be normative. However, outcomes may be anomalous. Teratogen effects demonstrate equally clearly that an outcome may be highly atypical for the species.

Evolutionary and Ontogenetic Time Scales Qualitative and quantitative estimates for every sensitive-period structural characteristic can be expected to vary naturally; in addition, many parameters will be modifiable based on experience or experiment. Thus, in describing all of the foregoing structural characteristics of the sensitive period, it is necessary to consider two generic properties, variability and modifiability.

Variability The range and causes of variation within and among species in developmental dating, onset, offset, duration, asymptote, experience, and pathway of a system during the tenure of the sensitive period, and the range and causes of variation in outcome, manner, outcome conditions, and duration of the outcome for a given system's sensitive period subsequently. What are the population's central tendency and distribution values for the several characteristics of a sensitive period? In the absence of systematic study, it is impossible to say with assurance what the nature of the distribution of values for sensitive-period parameters may be, although it is generally reasonable to expect population values to adhere to a normal distribution. Two central issues arise with regard to variability; they concern, first, variation in each sensitive-period characteristic within a species and, second, variation across species for each sensitive-period characteristic. Research documents different kinds of variation in sensitiveperiod parameters within a species. Scott (1958), for example, estimated the variability in onsets of sensitive periods for social behavior of puppies at 3-4 days on either side of the species mean value. Variation can also be found at different levels of system function. Wiesel (1982), for example, observed that temporal parameters of sensitive periods for cortical vulnerability vary among types of deficit, among brain regions, and among layers within individual cortical areas. Taking note of such variation is nontrivial, since sensitive-period effects may interact with one another. For example, the sensitive period for orientation selectivity in kittens ends at about 1.5 months, and the sensitive period for binocular selectivity ends at about 2

187

months; therefore, cortical cells could be simultaneously affected in complex spatial and temporal ways by the interaction of two processes (Daw, Berman, & Ariel, 1978). Separate sensitive-period characteristics also vary depending on known species differences. Immelmann and Suomi (1981) reported that the duration of the sensitive period for sexual imprinting is longer in the Bengalese finch than it is in the zebra finch. Henry (1984) studied the susceptibility of auditory function to environmental noise in two genetically distinct strains of mice and found that 133 C57BL/6 mice have a sensitive period for acoustic priming only, whereas 183 CBA mice have a sensitive period for acoustic priming for audiogenic seizures. Belyaev et al. (1985) reported that the offset of the sensitive period for primary socialization in farm-bred silver foxes takes place at 40-45 days, but for comparable foxes selected for domesticated behavior the offset of the sensitive period takes place at 60-65 days. Since it is to be expected that individuals normally give different but discrete values for a sensitive-period characteristic, it is important to note that the group function for a characteristic may be abrupt and representative of individuals or it may appear continuous and fail to represent individual data. Different species may differ in quantitative details of an analogous sensitive period; such differences are often quite informative. The following studies exemplify interspecific variation in developmental timing, in triggering stimulus, and in the existence of a sensitive period. The Japanese quail develops rapidly and is already adult-like by about Week 3, and its sensitive period for sexual imprinting takes place within that time (Gallagher, 1977); by contrast, the domestic fowl chick develops over a much longer period, and its sensitive period for sexual imprinting takes place at a correspondingly later time (Vidal, 1975). Redhead ducks imprint on visual and auditory stimulation provided by the mother, by contrast, canvasback ducks, their close cousins, imprint more strongly on auditory stimulation alone (Mattson & Evans, 1974). Most dramatically, rhesus monkeys that do not experience interaction with peers in the first 9 months of life later have difficulty establishing peer relationships (Harlow & Harlow, 1969); by contrast, pigtail monkeys that are raised under similar circumstances of social deprivation exhibit many fewer interpersonal difficulties later (Sackett, Holm, & Landesman-Dwyer, 1975). Are these dispersions meaningful? Within species, variability seems to reflect expected genetic dispersion or functional differences; across species, the distribution of sensitive-period characteristics seems to be adjusted to species-appropriate developmental timetables and conditions of rearing. Beyond their naturally occurring variability, sensitive-period characteristics may be experientially or experimentally modifiable. Modifiability The degree to which developmental dating, onset, offset, duration, asymptote, experience, and pathway of a system may be altered during the tenure of the sensitive period, and the degree to which outcome, manner, outcome conditions, and duration of the outcome may be altered subsequently, as well as the agegraded nature of their flexibility. Sensitive periods are custom-

188

MARC H. BORNSTEIN

arily thought of as rigid and refractory to change, but research shows that many sensitive-period characteristics are modifiable. For example, cells in the cortex of the very young kitten are subject to sensitive periods for orientation selectivity and for binocularity, but the duration of those periods can be extended by dark rearing (e.g., Cynader, Berman, & Hein, 1976; Mower & Christen, 1985), as it can be reduced (or eliminated) by depletion of relevant neurotransmitters (e.g., Daw, Rader, Robertson, & Ariel, 1983); the sensitive period can also be reinstated in older cats by the infusion of noradrenalin (Pettigrew, 1982). Likewise, the sensitive period for filial imprinting in chicks and in ducks can be extended by chemical means (Hess, 1957), by selective perceptual experience (Moltz & Stettner, 1961), or by social rearing conditions (Sluckin, 1972), and it can be prevented altogether by stress (Landsberg & Weiss, 1975). Modifiability is not limited to sensitive periods specific to early life; the sensitive period for maternal interest in neonate lambs can be altered when parturition in ewes is chemically induced (Poindron et al., 1979), it can be prolonged by maternal social isolation (Moore, 1960), and it can be reopened by genital stimulation (Keverneetal., 1983). Fixity of outcome is likewise widely thought to be one of the sensitive period's most salient characteristics (e.g., Lorenz, 1937). To many minds, the sensitive period's lasting effect is what stamps this phenomenon with its special character. Among the many aspects of the sensitive period, the permanence of effects has been intensively investigated (and challenged; see Clarke & Clarke, 1976), presumably on account of the important role this eventuality would play in verifying the nature and significance of the sensitive period in the first place. Some outcomes do indeed seem to endure permanently. For example, exposure to one range of ambient temperatures early in development leads to masculinization of turtles, whereas exposure to another range leads unalterably to feminization (Bull, 1980). Research has shown, however, that many sensitive-period outcomes are modifiable or even reversible. Isolation in the first 12 months of the monkey's life eventuates in severely maladjusted social behaviors (Harlow & Harlow, 1969); however, "therapy" at a later time (including adaptation, self-paced visual input, and exposure to younger normal monkeys) can ameliorate adverse effects to a degree and encourage more appropriate species-normal behaviors (Novak & Harlow, 1975). Outcomes may be differentially mutable depending on the level of the system affected. Both anatomy (e.g., Olson & Freeman, 1978) and behavior (e.g., Timney, Mitchell, & Griffin, 1978) in the cat can be altered by visual deprivation in the sensitive period, but both may "recover" in the wake of stimulation in the post-sensitive-period period (see, too, Blasdel, Mitchell, Muir, & Pettigrew, 1977). By contrast, psychological systems of behavior in rats appear to recover from stress by handling or by shock more quickly and completely than do physiological ones (Ader, 1968; Levine & Lewis, 1959). Structural characterisitics of the sensitive period seem not to be fixed, but generically vary and are modifiable. With respect to the refractory nature of sensitive-period characteristics, many additional questions linger, such as How modifiable are individual sensitive-period parameters? Does modifiability vary at different developmental points in the life course of the

system? Is modifiability naturally occurring, must it be induced, or is it an epiphenomenon of non-natural testing procedures? What factors influence the modifiability of sensitive-period parameters? A particularly thorny issue with respect to modifiability turns on the question of how incompleteness or impermanence of outcome impact on credibility and investment in the idea of the sensitive period. Normally, either can be expected to be very damaging. Lenneberg (1967), for example, who postulated a sensitive period for learning language, argued that normal language acquisition depends on normal language experience between approximately 2 and 14 years of age; yet startling contradictory examples, such as Kaspar Hauser in Nuremberg, Germany, who acquired language after the age of 17 years (Simon, 1979) and Genie in California who began to acquire language after the age of approximately 14 years (Curtiss, 1977), cast genuine doubt on the strong form of Lenneberg's hypothesis (see Snow, 1987). That effects of a sensitive period are not fixed need not, however, undermine the notion of the sensitive period, so long as something special for development from that period remains. Thorpe's (1961) classic work on song learning in the chaffinch is illustrative: The fledgling male chaffinch that hears adults singing, or as a juvenile sings itself, will later produce the characteristic adult song, even if competition for territory is later required for adumbration of the fine details of its song. Utopian modifiability opens a door otherwise shut tight by relentlessly strict interpretation of the sensitive period. Overview However diverse and specialized their elected subjects of investigation, researchers of sensitive periods implicitly seem to recognize, first, the existence, function, and prominence of several characteristics of the sensitive period and, second, the requisiteness of specifying their quantitative and qualitative makeup. Essentially, therefore, a comprehensive statement about a sensitive period ought ideally to include information about its temporal and intensive contours: (a) when and how often in the life cycle the sensitive period occurs, (b) the rise of the sensitive period, (c) the decay of the sensitive period, (d) the window of the sensitive period, and (e) how sensitivity changes and the stability of sensitivity; about its mechanisms of change: (f) the nature and origins of the effective experience, (g) what structure or function changes, and (h) the channels by which experience affects the structure or function that changes; and about its consequences: (i) the outcome in later development, (j) how the outcome is effected, (k) when and under what circumstances the outcome occurs, and (1) how long the outcome lasts. Beyond these defining parameters of the sensitive period and its consequences, information about two evolutionary and ontogenetic considerations of sensitive periods is also requisite: (m) individual and species variation in sensitive-period characteristics and (n) the modifiability of the individual parameters of the sensitive period. Thus, 14 conceptually distinct structural characteristics arrayed in four groups can be identified as inhering in the sensitive period. All researchers confront the task of describing each of these characteristics; indeed, many view the quantitative delin-

SENSITIVE PERIODS IN DEVELOPMENT cation of some or all of these parameters to be a primary goal of studying sensitive periods. Indeed, descriptive specification of characteristics constitutes a first-order achievement of sensitive-period study; additionally, it is critical to illuminating causes underlying sensitive-period phenomena. A Sensitive Period Framework: II. Causal Interpretations Next to designating characteristics that define its structure, the aspect of the sensitive period that challenges, entices, and perplexes developmental investigators most is the task of identifying causes underlying its emergence. Questions of cause divide into two classes: The first asks why the sensitive period arises, and the second asks how specifically the sensitive period is regulated. The sensitive period confronts researchers and theoreticians working in different systems with a conceptually and evolutionally enigmatic phenomenon to explain on each of these two levels, and one in which the significance of intricate and complex factors such as nature, nurture, and their transaction through time are clearly at issue. Interpretations of the evolutionary meaning of sensitive periods vary (the Why? question), as do attributions of sources by which sensitive periods are regulated (the How? question). A comprehensive treatment of the sensitive period at the descriptive level (the What? question) is requisite to addressing these two deeper concerns about cause. In this section, I discuss these two foremost questions about cause of the sensitive period in seriatim.

Ultimate Cause: Why Evolutionally do Sensitive Periods Arise? The sensitive period reflects a developmental phase of builtin competence for specific exchange between organism and environment whose consequences presumably endure for the organism. Thus, the sensitive period qua evolutionary mechanism represents a psychological phenomenon of singular consequence. There are two perspectives from which it is necessary to view the evolutionary meaning of the sensitive period. The first is constructive and positive: The sensitive period uniquely prepares the system for its developmental future, forecasting mature structure or function relative to early experience. The sensitive period thereby ensures that the system is responsive to its developmental environment, and it buffers the system against later change in the environment or in the self. Such a powerful and pervasive ontogenetic mechanism presumably reflects a beneficial adaptation within the evolutionary program of the organism. For this reason, perhaps, most sensitive periods are to be found in the infancy of the organism, since their appearance at this time safeguards that the organism will quickly and completely acquire information that is both critical to and typical of its physical and social locale. This is practical information that is commonly thought to enhance the probability of survival and the achievement of adaptive competence. In this construal, the sensitive is an "optimal" period. The sensitive is also a "vulnerable" period, however, since in the sensitive period brief experiences can exert profound and sometimes permanent detrimental effects from very early in the

189

life of an organism—certainly before the conscious, self-aware selection of experiences. That is, the plasticity that inheres in the concept of the sensitive period has a distinctly negative construal, too. The second evolutionary ramification of the sensitive period is, then, the jeopardy to the organism introduced by the sensitive period. Conditions that are necessary to or typical of development but that are not met, or atypical conditions that are encountered haphazardly, place the organism at permanent risk. The sensitive period is, therefore, poised to facilitate adaptation or to handicap the unsuspecting organism perhaps for the remainder of its existence. Indeed, the original discoveries of sensitive periods were based on the permanent damage caused by noxious stimulation during early phases in development. For these reasons, sensitive periods offer especially absorbing subjects of study. In many ways, sensitive periods improve in comprehensibility when contemplated as a form of adaptation. Like all adaptations, even analogous sensitive periods can be expected to vary in their particular manifestations among species on account of the different histories of natural selection of those species. However, a given sensitive period can be expected to be universal within a species principally because of shared evolutionary (genetic and environmental) history. Ethologists, particularly, have concerned themselves with addressing the question of why sensitive periods arise. In doing so, ethologists speculate on the twin foundations of survival and of reproductive success of organisms that inhabit different niches (e.g., Baker, 1938; Thomson, 1950), and sensitive periods have figured prominently among their speculations in this regard. It may be that ultimate causes of sensitive periods are selected for in the natural history of the organism and eventuate as specific adaptations. Consider the case of ungulates; "herds of grazing mammals characteristically produce precocial offspring in synchrony, and it is therefore important for the mother to form a rapid recognition of her own offspring to distinguish them from others" (Pissonnier et al., 1985, p. 361). Immelmann and Suomi (1981) have viewed the sensitive period out of such an evolutionary perspective, accounting thereby for species variation in parameter values of many structural characteristics of sensitive periods. Consider, for example, the duration parameter, \bung animals must acquire critical social knowledge from conspecifics before departing the family, and therefore sensitive periods for learning social behaviors must fall within those chronological bounds when the young of a given species are still with their parents. Avian species as a whole spend considerably less of their developmental history after birth with their parents than do mammals, and therefore it is predictable that avian sensitive periods for the acquisition of social knowledge are shorter lived than are mammalian ones and that the population variation for sensitive-period duration among birds is more restricted than it is among mammals. As Immelmann and Soumi (1981) point out, In many species and for different functional systems, strong selection pressures that favor great sensitivity to certain environmental stimuli and a maximum of learning during early stages of life may well exist. They may also favor great stability of the results of such early experience, providing some kind of degree of protection against some later possible influences on the individual. In addi-

190

MARC H. BORNSTEIN

tion, it becomes apparent that differences in such specific needs and demands between species are almost certainly responsible in large part for the enormous diversity in the degree of stability and in the duration and characteristics of sensitive phases, (p. 399, emphasis added) On this account, it behooves researchers always to consider the sensitive period they study out of an evolutionary perspective, since adopting such an outlook promises to enrich investigators' understanding of individual structural characteristics through the larger meaning of the sensitive period for the species in question. In short, the evolutionary stance helps to unravel some mysteries about why sensitive periods might arise as well as why they possess some of the particular characteristics they do. Proximate Cause: How are Sensitive Periods Regulated? Adaptive consequence aids in attributing ultimate cause of a sensitive period in a species. However, there also must exist corresponding "proximate cause" that regulates sensitivity to meet the aim of natural selection for the sensitive period. Examination of proximate cause addresses the question of how sensitive periods are regulated. To confirm attributions of proximate cause is exacting. Understandably, pinpointing the mechanism of a natural interaction, such as the sensitive period, challenges investigators conceptually. Moreover, signal confounds of empirical designs typical of much research into sensitive periods have thrown up special obstacles against fitting causal inference. Only systematic experimentation can clarify the nature of cause and bring explicitness to deductions so common and seemingly necessary to theorizing about the sensitive period. Happily, Bateson and Hinde (Bateson, 1979, 1983; Bateson & Hinde, 1987; Hinde & Bateson, 1984) have outlined three classes of proximate cause explanation for the sensitive period and delineated for each permissible limits of causal attribution. I recount each in turn, modifying the original terminology somewhat to accord with the vocabulary established in this article. The alteration in individual responsiveness that betokens the sensitive period can logically reflect regulation by physical, physiological, or psychological mechanisms. These three classes of mechanism may operate in concert, and presumably different mechanisms account for different categories of sensitive periods in different proportions. Physical Regulation of characteristics of the sensitive period through physical mechanisms, including variation in genetics and biomolecular process. One way in which researchers have attempted to identify genetic mechanisms of cause, for instance, is by contrasting differences in sensitive periods across related species. Freedman (1967) showed that distinct strains of dogs react qualitatively differently following a sensitive period for handling, and Henry (1984) showed that distinct strains of mice experience qualitatively different sensitive periods to audiogenie seizures. A second way researchers have attempted to assay physical mechanisms that regulate sensitive periods is

through close examination of biomolecular phenomena. Reinisch (1974) showed that hormones (specifically androgens) function critically during the sensitive period of brain organization to influence a wide variety of sexually dimorphic behaviors, and Landsberg and Weiss (1976) showed that hormones (specifically blood corticosterones) play a determinative role in the sensitive period for filial imprinting among young precocial birds. Indeed, Lauder and Krebs (1986) have inferred that neurotransmitters and neurohumors act as general mechanisms underlying all sensitive periods. Physiological Regulation of characteristics of the sensitive period through physiological mechanisms, including variation in state of arousal, anatomy, or perceptual capacity. Different investigators have argued for the pivotal contribution of each of these physiological mechanisms, even with respect to the same class of sensitive period. For example, in imprinting among young precocial birds, Martin and Schutz (1974) determined that general arousal played a significant role; Strobel, Baker, and MacDonald (1967) and Bradley (1985) that change in central nervous system structures played a significant role; and Immelmann (1972) that advances in visual perceptual ability played a significant role. Psychological Regulation of characteristics of the sensitive period through psychological mechanisms, including prominently variation in behavior. Activities of the organism itself may influence the sensitive period. Hess (1973) established that locomotion plays a determinative role during the sensitive period for filial imprinting among precocial birds; Freeman and Bonds (1979) demonstrated that eye movements are decisive in the sensitive period for cortical binocularity in the kitten; Gubernick (1981) discovered that mother goats' licking their young is necessary during the sensitive period for olfactory recognition of offspring; and Gottlieb (1985) confirmed that embryonic ducks must produce their own contact-contentment call during a sensitive period in order to exhibit a normal preference for the species-specific maternal call. Elucidating mechanism in proximate cause is vital as it reveals how and why, among other things, sensitive periods arise, endure, and decay. However, at least three interrelated caveats must be heeded in attributing proximate cause in a sensitive period to one or more mechanisms; this triad concerns exclusivity, reductionism, and temporal priority among mechanisms. First, although physical, physiological, and psychological levels of explanation may be distinguished operationally, the attribution of causality to one or another mechanism can be confounding, since these levels of cause are by no means mutually exclusive. Molecular status of the system can underpin behavioral change that initiates a sensitive period, but behavioral change can equally well induce molecular change that instigates the behavioral change. For example, naturally maturing levels of corticosterones have been implicated in the development of fear, which in turn influences imprinting; perhaps, however,

SENSITIVE PERIODS IN DEVELOPMENT

naturally occurring stressors provoke corticosterone production, or experience with novelty produces stress, and so on. In short, the attribution of causality in proximate cause is potentially subject to a critique of circularity. It may be safest to conceive of causes of a sensitive period as layered. A second correlated caveat concerns the seductive, but rather simplistic, reductionism that often tempts hypothesizing about proximate cause in sensitive-period research. In this orientation, psychological attributions immediately invoke explanations in terms of physiological substrates—perceptual function is reducible to cortical status—just as physiological attributions immediately invoke explanations in terms of physical substrates—cortical plasticity is reducible to neurohormonal flow. Researchers must be ever vigilant to acknowledge and to respect the integrity of each class of mechanism. A third associated caution in attributing proximate cause is that a given mechanism may apply at the time the sensitive period actually arises, or it may apply from an earlier time. For example, physiological mechanisms could engender a sensitive period in a system immediately, or physiological mechanisms could prepare a system for later instigation of the sensitive period. Inevitably, ambiguity must be expected in distinguishing immediate cause from cause displaced in the past. As a consequence of these several considerations, assigning monistic proximate cause in the domain of the sensitive period is extremely hazardous. It ought to be possible to do so theoretically, however, even if the various categories of proximate cause are not restricted. To achieve greater insight into the pertinence of one or another explanation of the sensitive period, it appears to be both desirable and feasible to contrast explanations of different mechanisms experimentally. If, for example, programmed physical factors excite a sensitive period, the system ought to display altered responsiveness independent of associated psychological expression. Perhaps the most pervasive, penetrating, and elusive question regarding proximate cause of the sensitive period recalls the Nature-Nurture debate (and has equally much meaning): Is proximate cause endogenously, exogenously, or interactively influenced? Opinions on this question vary dramatically. Consider the variety of interpretations that have been offered to account for the offset of the sensitive period in imprinting: Explanations include perceptual maturation (Moltz, 1968); the emergence of fear, itself brought about by inherent age-related forces (Hoffman, 1987; Lorenz, 1935; Scott, 1962; Weiss, Kohler, & Landsberg, 1977) or by social experience (Alley & Boyd, 1950; Salzen, 1963); and new objects in the environment (Sluckin, 1964) or other nonspecific photic stimulation (Millikan, 1972); as well as the process of imprinting itself (Guiton & Sluckin, 1969). By turns, too, the phenomenon of the sensitive period itself has received a similarly diverse theoretical treatment. Gottlieb (1961) argued that sensitive periods arise out of endogenous motives; Hess (1973) argued that sensitive periods arise on account of exogenous factors; and Bateson and Hinde (1987) argued that sensitive periods arise ineluctably from the interaction of endogenous with exogenous forces (see, too, Bronson, 1962, 1965; Over, 1979; Schneirla & Rosenblatt, 1963). The explication of cause of a sensitive period calls for exami-

191

nation of both immediate mechanism and adaptive meaning. Clearly, insights into proximate cause along with indications of ultimate cause enlarge the scope of understanding of sensitive periods. However, achieving unabridged intuition about causes of sensitive periods poses a true challenge.

Implications of the Sensitive Period Framework for Research and Theory How can attending to descriptive features and explanatory meaning lead to advances in theoretical and empirical research in individual sensitive periods or in the sensitive period generically? Systemization of research efforts around these characteristics of structure and interpretations of cause presents certain clear advantages. In this section, I discuss how the foregoing organization contributes to understanding sensitive periods, and in doing so I provide both concrete and theoretical examples that show how new insights might be secured from the framework described. Research designs that dominate published investigations of sensitive periods—independent of content area—nearly uniformly are devoted to specifying one or another of the several structural characteristics of the sensitive period. (Proximate and ultimate cause have generally provoked less attention.) Relevant literatures abound with specific examples of both observational and experimental investigations of nearly every characteristic. In the typical experimental design, investigators manipulate values of one parameter, holding (some) other parameters constant (to the degree possible), in order to determine the meaningful range of values for the single parameter of interest. So, for example, in an experimental series a neuropsychologist interested in the sensitive period for visual sensitivity might first expose groups of animals to the same visual experience at different times in their ontogeny and later assess the influence of developmental dating on anatomy, physiology, or behavior in order to evaluate the parameter of developmental dating of the sensitive period. In subsequent experiments, the neuropsychologist might expose groups of animals during that newly defined sensitive period to different visual experiences, holding duration of exposure constant, or for different durations of exposure, holding the experience itself constant, in order to evaluate the critical nature of the experience and its duration in the sensitive period. Other permutations are selfevident. Insofar as examining each of the 14 parameters of the sensitive period (to say nothing of cause) may require numerous separate groups or treatments, exhaustive study is often not possible or might not be deemed worthwhile. In practice, research investigators rarely meet the considerable challenge of specifying parameter values even for a plurality of the 14 characteristics of the sensitive period they study. However, failure to explore certain parameters comprehensively, or to recognize the need to explore them in depth, has led to consequential differences of opinion or of fact about sensitive periods, and thereby limited understanding of this phenomenon. Consider the following examples. First, on account of the primacy, speed, rapidity, and permanence of the social attachments precocial ducklings were observed to form, Lorenz

192

MARC H. BORNSTEIN

(1937) identified and described imprinting as a "critical period."2 Systematic investigation of structural characteristics associated with imprinting subsequent to Lorenz led to quantitative modification of his qualitative attributions about several aspects of the nature of imprinting—and, not insignificantly, to reversals of Lorenz's pronouncements on the abruptness, the uniqueness, and the permanence of these time-bound attachments. Second, it is widely assumed that the sensitive period reflects an individually unique phenomenon normally circumscribed to an early phase of the life-cycle. Yet, phylogenetic study discloses multiple sensitive periods for maternal bonding in a substantial number of adult mammals (Rosenblatt & Siegel, 1981). Not only are these findings revolutionary with respect to the common view of the sensitive period, but such a perspective in turn opens a range of rather novel questions about the sensitive period. For example, if sensitive periods are multiple, rather than singular, in the individual, has the sensitive period a refractory period? Is there one, or are there several separate, maternal sensitive periods for twin births? Third, parameter values of sensitive periods must be appreciated to fluctuate, reflecting different experimental approaches and different empirical indexes. Thus, for example, the finegrainedness of analysis places limits on specifying the developmental timing of visual system sensitivity; behavioral data suggest one parameter value for offset of the sensitive period in language deprivation, whereas physiological data suggest another; and, the experiences of isolated versus communal rearing considerably alter variability in imprinting. Fourth, based on earlier demonstrations that showed handling in a critical period affected rats' later behavior (Denenberg & Zarrow, 1971), Wyly, Denenberg, de Santis, Burns, and Zarrow (1975) examined the rabbit for the existence of an analogous critical period. Unlike the rat, the rabbit does not display especial sensitivity in open-field, exploratory, or social behavior in response to handling early in its infancy, but rather exhibits selective effects that have proved cumulative rather than timelocked. Clearly, cross-species evaluations inform both specific and general comprehension of the sensitive period. Last, the obligatory reliance on correlational techniques and natural experiments in human sensitive-period research has sometimes misled conclusions. For example, Lenneberg (1967) speculated that the effects of language deprivation would reverberate immutably in the child's inability to acquire and to use language with facility; however, Lenneberg failed to examine examples that would test his speculations about the duration of deprivation. Extant case histories, as of Kaspar Hauser, should have given pause to the strong claims for a sensitive period in human language acquisition. On their face, sensitive periods vary bewilderingly both in locus and in time: hours for ducklings to imprint (e.g., Hess, 1973), days for birds to learn a song (e.g., Marler, 1970), weeks for canine socialization (e.g., Scott, 1962), months to ensure sexual normalcy in monkeys (e.g., Harlow & Harlow, 1969), and years in the case of human binocular sensitivity (e.g., Banks et al., 1975). This variety seems understandably, if unhappily, to have daunted researchers and to have masked intelligence

into a uniform view of the phenomenon of the sensitive period per se. Despite this striking diversity among individual instances of sensitive periods, the proposed research framework lends itself to the profitable establishment of common quantitative values for the several parameters of sensitive periods and of sound qualitative intuition about cause. In this regard, what specific advantages does the taxonomy afford? Dividing the sensitive period into constituent processes enables their individualized study and imparts a higher order of understanding both to constituents themselves and to the sensitive period as a whole. Consider, for example, developmental dating, a prime sensitive-period structural characteristic. The sensitive period for binocularity has been found to keep between postnatal Weeks 4 and 12 in the cat (e.g., Hubel & Wiesel, 1970), Months 1 and 2 in the monkey (e.g., Wiesel, 1982), and Years 1 and 7 in the human being (e.g., Banks et al., 1975). By themselves, each of these values provides meaningful species information, but a deeper explication of sensitive-period mechanisms is gained through more comprehensive species comparison, since comparative developmental dating reveals that in each species the sensitive period for binocularity appears to initiate approximately when cells driven by each eye are matured sufficiently to compete for cortical synapses, and the sensitive period in each appears to cease approximately when competitive synaptogenesis stabilizes (see, e.g., Kasamatsu & Pettigrew, 1979; Wiesel & Hubel, 1963). Additionally, having a unified framework within which to study sensitive periods enables knowledge garnered in one subject area to be applied to others. For example, sensitive periods have been observed normally to occur very early in the ontogenetic history of many systems in many species, that is during prenatal development or infancy. The organizational perspective imported here has two notable implications for thinking about this aspect of the sensitive period. First, the standard result suggests that, in searching for sensitive periods in any new system, it might be strategically efficient initially to focus on the early life course of that system. Such an approach can be expected to reduce the risk of missing and to enhance the likelihood of detecting a sensitive period. The literature in infantile stimulation provides an illustration. Bernstein (1952) observed that animals that experienced more handling in adolescence gained more weight in full maturity and performed in a superior manner on discrimination tasks compared with adolescent animals handled minimally or not at all. Later, on the basis of general experience in sensitive-period research, Hunt and Otis (1963) and Schaefer (1963) reasoned—it turned out correctly—that such outcomes might be enhanced if animals were first handled in their infancy rather than their adolescence. The second implication of a general finding is that anomalous or contradictory results may flag close scrutiny. The question of a sensitive period for maternal bonding in human beings illustrates the point (Klaus & Kennell, 1976). Certainly, confirma2

Not insignificantly, Lorenz's actually calling the phenomenon he observed a "critical" period imbued it with a host of unnecessary connotations that many of his heirs, in the exact same and in conceptually related fields, have had to struggle (thanklessly) to rectify and overturn.

SENSITIVE PERIODS IN DEVELOPMENT tion of a sensitive period for maternal bonding in infrahuman species intimates that it exists in human beings. In spite of how extraordinary and commanding an experience the birth of a child may be, claims for a systemic sensitive period for maternal bonding at parturition among human beings ought still to have been greeted with more reserve, simply on the argument that the timing of sensitive periods is almost always localized to infancy rather than to maturity. In this example, the framework instigates search cum skepticism. A sensitive period for adult anything constitutes a low probability event a priori, and— claims and desires to the contrary—contemporary research has largely failed to substantiate the existence of a sensitive period for bonding in humans (see reviews by Goldberg, 1983; Herbert, Sluckin, & Sluckin, 1982;Leiderman, 1981; Myers, 1984). Arriving at this conclusion does not mean that parameter values outside general norms are unobtainable or nonexistent; a sensitive period for maternal bonding can be found in some infrahuman species (Rosenblatt & Siegel, 1981; Pissonnier et al., 1985). Rather, it suggests that investigators can and ought to avail themselves of common parameter values growing out of the research framework to predict and to evaluate new as well as existing findings. Reciprocally, however, researchers are wise to refrain from exclusive reliance on such strategizing, as it may prove overly conservative. To continue the example at hand, wholly depending on general experience risks the potential reward to be reaped, say, from discovering a novel or unique sensitive period specific to mature phases of the life-cycle. Lucid and extensive description are practically imperative with phenomena as disparate and intricate as sensitive periods. The more articulate and learned the characterization, the more likely is the researcher to gain a deeper comprehension of the sensitive period. Sensitive periods constitute unique windows of organism-environment interaction, positive or negative as their consequences may be. To the extent that the sensitive period concept has relevance to biology, psychology, and sociology, not enough information can ever be gleaned about its structural characteristics and causal interpretations. Clear and comprehensive description is vital to accessible research. The purpose of the framework delimited here is to bring order to the complexity and diversity inherent in sensitive period study.

Conclusion Despite its contemporary heterogeneity and high degree of specialization, psychology embraces many general phenomena that are of broad interest across its subdisciplines. This is true, in part, because of the generality and ubiquity of those phenomena. The sensitive period in development qualifies as one. Many examples have emerged in widely different structural or functional systems that the presence or absence of certain experiences at particular times in the life-cycle may influence structure or function well beyond the time of the experience. Examples of sensitive periods may be found widely dispersed in animal and in human neurobiology and behavior. Proof of the extraordinary breadth and rewarding character of the concept of the sensitive period is that it has been applied profitably to explicating development of cells, of human beings, and of institutions.

193

Whatever their subfield, investigators of sensitive periods confront similar challenges, because sensitive periods sui generis are described by a common set of characteristics and require similar kinds of explanation. On these bases, a framework of description and explanation in sensitive-period research has crystallized. A central purpose of articulating this framework is to foster broader and deeper quantitative and qualitative treatments of sensitive periods in future. Clearly, too, the principles of sensitive periods clarified by students of one system can be serviceable to students of other systems. Yet, many investigators seem genuinely unaware of the relevance to their own work of sensitive periods subjected to investigation in allied fields; indeed, some research and write as though oblivious to other manifestations of the phenomenon. This strategy needlessly jeopardizes advances in comprehending individual sensitive periods and the sensitive period in general. No sensitive period in one system must resemble that in any other, and no one sensitive period need serve as a model for others. There is no all-encompassing theory of the sensitive period, although many investigators and theoreticians have searched for one. It is possible, nonetheless, to advance a framework of structural characteristics and causal interpretations out of which to consider sensitive periods generally. In other words, one "sensitive period" does not mediate cortical specificity and filial imprinting; but a unified framework might still apply across the full spectrum of remarkably variegated instantiations of sensitive periods. In turn, that framework will prove to be a rewarding heuristic tool toward redefining both philosophy and experiment in sensitive-period research. Elucidating such a universal framework for the sensitive period has been the task of this commentary. Conceptual analysis and research review suggest that investigators of sensitive periods who wish to grasp the essence of this ubiquitous developmental phenomenon thoroughly are obliged, first, to acknowledge parameter values of 14 different structural characteristics of the sensitive period, as they are embedded in four distinctive conceptual sets, and, second, to appreciate both ultimate cause of the sensitive period in terms of its adaptive significance and proximate cause of the sensitive period in terms of mechanisms that might regulate the sensitive period. Perhaps no existing report of a sensitive period—the undeniable popularity of this phenomenon in many areas of biological, psychological, and sociological science notwithstanding—meets the task of specifying all structural characteristics and causal interpretations. As demonstrated, however, the movement toward a comprehensive treatment promises invaluable direction to all future inquiry. In the end, unfortunately, even this detailed and extensive framework will fail to provide information that is exhaustive and determinative. Even where there may be widespread empirical agreement about values to be attached to sensitive-period characteristics, and some indications of ultimate and proximate cause may offer themselves, theoretical convergence about the conceptual sources of sensitive periods may still be lacking. For example, different theoretical traditions propose different reasons why sensitive periods tend to occur early in ontogeny. Maturational theory argues that the period of greatest susceptibility, vulnerability, or responsiveness to stimulation in an organizational process coincides with the phase of its most rapid growth,

194

MARC H. BORNSTEIN

usually early in the life-cycle; information theory posits that systems in a state of developmental flux, as is usually true of early ontogeny, are more open to external influences than are organized and fixed systems; learning theory suggests that acquisition proceeds most efficaciously when there are no extant competing responses, the situation that necessarily characterizes the earliest life stage of a system; and, cognitive developmental theory proposes that mature and complex capacities are constructed of simple elements that themselves develop first in infancy. Thus, even when the framework of the sensitive period is adequately detailed, inmost facets of sensitive periods will remain enigmatic and unattainable. Sensitive periods appeal to developmental investigators for many reasons—from the superficial attraction of what has been called the "pleasing dactylic rhythm" of the term itself to the "exciting connotation of developmental brinksmanship" the phenomenon connotes. Despite this power, sensitive periods should not enjoy a license to rivet the developmentalist's attention singlemindedly on one point in the life span. The notion of the sensitive period implies that a certain experience at a certain time in the life-cycle of a system may exert a dramatic effect on the future developmental course of that system. Research shows that events subsequent to the sensitive period may intervene to modify or even nullify effects thus established, or it may be in the character of the system itself to modulate at later points in its life history. As is widely recognized, every phase in the life course is somehow critical. Still, to paraphrase the poet, theory and data signify that some periods in life may be more critical than others. References Ader, R. (1968). Effects of early experience on emotional and physiological reactivity of the rat. Journal of Comparative and Physiological Psychology, 66, 264-268. Alley, R., & Boyd, H. (1950). Parent-young recognition in the coot. Ibis, 92, 46-51. Almli, C. R., & Finger, S. (1987). Neural insult and critical period concepts. In M. H. Bornstein (Ed.), Sensitive periods in development: Interdisciplinary perspectives (pp. 123-143). Hillsdale, NJ: Erlbaum. Aslin, R. N. (1981). Experiential influences and sensitive periods in perceptual development: A unified model. In R. N. Aslin, J. R. Alberts, & M. R. Peterson (Eds.), Development of perception: Psychobiological perspectives (Vol. 2, pp. 45-93). New \brk: Academic. Baker, J. R. (1938). The evolution of breeding seasons. In G. R. de Beer (Ed.), Essays on aspects of evolutionary biology (pp. 161-177). New York: Oxford University Press. Banks, M. S., Aslin, R. N., & Letson, R. D. (1975). Sensitive period for the development of human binocular vision. Science, 190, 675-677. Bateson, P. (1979). How do sensitive periods arise and what are they fort Animal Behaviour, 27, 470-486. Bateson, P. (1983). The interpretation of sensitive periods. In A. Oliverio & M. Zappella (Eds.), The behavior of human infants (pp. 5770). New York: Plenum. Bateson, P., & Hinde, R. A. (1987). Developmental changes in sensitivity to experience. In M. H. Bornstein (Ed.), Sensitive periods in development: Interdisciplinary perspectives (pp. 19-34). Hillsdale, NJ: Erlbaum. Beatty, W. W., Dodge, A. M., Traylor, K. L., & Meaney, M. J. (1981). Temporal boundary of the sensitive period for hormonal organization of social play in juvenile rats. Physiology & Behavior, 26, 241-243.

Belyaev, D. K.., Plyusnina, I. Z., & Trut, L. N. (1985). Domestication in the silver fox (Vulpes fulvus Desm): Changes in physiological boundaries of the sensitive period of primary socialization. Applied Animal Science, 13, 359- 370. Bernstein, L. (1952). A note on Christie's "Experimental naivete and experiential naivete." Psychological Bulletin, 49, 38- 40. Blakemore, C., & Mitchell, D. E. (1973). Environmental modification of the visual cortex and the neural basis of learning and memory. Nature, 241, 467-468. Blasdel, G. G., Mitchell, D. E., Muir, D. W., & Pettigrew, J. D. (1977). A physiological and behavioral study in cats of the effect of early visual experience with contours of a single orientation. Journal of Physiology, 265,615-636. Blass, E. M. (1987). Critical events during sensitive periods of social development in rats. In M. H. Bornstein (Ed.), Sensitive periods in development: Interdisciplinary perspectives (pp. 81-98). Hillsdale, NJ: Erlbaum. Boothe, R. G., Vassdal, E. & Schneck, M. (1986). Experience and development in the visual system: Anatomical studies. In W. T. Greenough & J. M. Juraska (Eds.), Developmental neuropsychobiology (pp. 295315). Orlando, FL: Academic Press. Bornstein, M. H. (Ed.). (1987). Sensitive periods in development: Interdisciplinary perspectives. Hillsdale, NJ: Erlbaum. Bradley, P. (1985). A light and electron microscopic study of the development of two regions of the chick forebrain. Developmental Brain Research, 20, 83-88. Bronson, G. (1962). Critical periods in human development. British Journal of Medical Psychology, 35, 127-133. Bronson, G. (1965). The hierarchical organization of the central nervous system: Implications for learning processes and critical periods in early development. Behavioral Science, 10, 7-25. Brook, C. G. D. (1972, September 23). Evidence for a sensitive period in adipose-cell replication in man. Lancet, 2, 624-627. Buisseret, P., Gary-Bobo, E., & Imbert, M. (1982). Plasticity in the kitten's visual cortex: Effects of the suppression of visual experience upon the orientational properties of visual cortical cells. Developmental Brain Research, 4, 417-426. Bull, J. J. (1980). Sex determination in reptiles. Quarterly Review of Biology, 55,3-21. Byrne, E. A., & Smart, J. L. (1980). Delimitation of a sensitive period for the effects of early life undernutrition on social behaviour of adult male rats. Physiology & Behavior, 24, 131-133. Cairns, R. B., Hood, K. E., & Midlam, J. (1985). On fighting in mice: Is there a sensitive period for isolation effects. Animal Behaviour, 33, 166-180. Caldwell, B. M. (1962). The usefulness of the critical period hypothesis in the study of filiative behavior. Merrill-Palmer Quarterly, 8, 229242. Child, C. M. (1921). The origin and development of the nervous system. Chicago: University of Chicago Press. Clark, M. M., & Galef, B. G. (1979). A sensitive period for the maintenance of emotionality in Mongolian gerbils. Journal of Comparative and Physiological Psychology, 93, 200-210. Clarke, A. M., & Clarke, A. D. B. '(1976). Early experience: Myth and evidence. London: Open Books. Collias, N. E. (1956). The analysis of socialization in sheep and goats. Ecology, 37, 228-239. Colombo, J. (1982). The critical period hypothesis: Research, methodology, and theoretical issues. Psychological Bulletin, 91, 260-275. Cronly-Dillon, J. R., & Perry, G. W. (1976). Tubulin synthesis in developing rat visual cortex. Nature, 261, 581 -583. Curtiss, S. (1977). Genie: A psycholinguistic study of a modern-day "wildchild."New York: Academic Press.

SENSITIVE PERIODS IN DEVELOPMENT Cynader, M., Herman, N., & Hein, A. (1976). Recovery of function in cat visual cortex following prolonged deprivation. Experimental Brain Research, 25, 139-156. Daw, N. W., Herman, N. E. J., & Ariel, M. (1978). Interaction of critical periods in the visual cortex of kittens. Science, 199, 565-567. Daw, N. W., Rader, R. K., Robertson, T. W., & Ariel, M. (1983). Effect of 6-hydroxydopamine on visual deprivation in the kitten striate cortex. Journal of Neuroscience, 3, 907-914. Denenberg, V. H. (1963). Early experience and emotional development. Scientific American, 208, 138-146. Denenberg, V. H. (1968). A consideration of the usefulness of the critical period hypothesis as applied to the stimulation of rodents in infancy. In G. Newton & S. Levine (Eds.), Early experience and behavior (pp. 142-167). Springfield, IL: Charles C Thomas. Denenberg, V. H., & Bell, R. W. (1960). Critical periods for the effects of infantile experience on adult learning. Science, 131, 227-228. Denenberg, V. H., & Zarrow, M. X. (1971). Effects of handling in infancy upon adult behavior and adrenal activity: Suggestions for a neuroendocrine mechanism. In D. N. Walcher & D. L. Peters (Eds.), Early childhood: The development of self-regulatory mechanisms (pp. 39-71). New York: Academic. Dobbing, J., & Sands, J. (1973). Quantitative growth and development of human brain. Archives of Diseases of Childhood, 48, 757-767. Einon, D. E, & Morgan, M. J. (1977). A critical period for social isolation in the rat. Developmental Psychobiology, 10, 123-132. Elberger, A. J. (1984). The existence of a separate, brief critical period for the corpus callosum to affect visual development. Behavioural Brain Research, 11, 223-231. Evans, R. M. (1970). Imprinting and mobility in young ring-billed gulls, Larus delawarensis. Animal Behaviour Monographs, 3, 193-248. Fabricius, E. (1964). Crucial periods in the development of the following response in young nidifugous birds. Zeitschrift fur Tierpsychologie, 21, 326-337. Forgays, D. G., & Read, J. M. (1962). Crucial periods for free-environmental experience in the rat. Journal of Comparative and Physiological Psychology, 55, 816-818. Freedman, D. G. (1967). The origins of social behaviour. Science Journal, 3, 69-73. Freedman, D. G., King, J. A., & Elliot, O. (1961). Critical period in the social development of dogs. Science, 133, 1016-1017. Freeman, R. D. (1978). Some neural and non-neural factors in visual development of the kitten. Archives of Italian Biology, 116, 338-351. Freeman, R. D., & Bonds, A. B. (1979). Cortical plasticity in monocularly deprived immobilized kittens depends on eye movement. Science, 206, 1093-1095. Gallagher, J. E. (1977). Sexual imprinting: A sensitive period in Japanese quail (Coturnix coturnix japonica). Journal of Comparative and Physiological Psychology, 91, 72-78. Ganz, L., & Fitch, M. (1968). The effect of visual deprivation on perceptual behavior. Experimental Neurology, 22, 638-660. Goldberg, S. (1983). Parent-infant bonding: Another look. Child Development, 54, 1355-1382. Gottlieb, G. (1961). Developmental age as a baseline for determination of the critical period in imprinting. Journal ofComparative and Physiological Psychology, 54, 422-427. Gottlieb, G. (1983). The psychobiological approach to developmental issues. In M. M. Haith & J. J. Campos (Vol. Eds.), Infancy and developmental psychobiology (Vol. 2). P. H. Mussen (Ed.), Handbook of child psychology (pp. 1-26). New York: Wiley. Gottlieb, G., & Klopfer, P. H. (1962). The relation of developmental age to auditory and visual imprinting. Journal ofComparative and Physiological Psychology, 55, 821-826. Gubernick, D. J. (1981). Parent and infant attachment in mammals. In

195

D. J. Gubernick & P. H. Klopfer (Eds.), Parental care in mammals (pp. 243-305). New York: Plenum. Guillery, R. W., & Stelzner, D. J. (1970). The differential effects of unilateral lid closure upon the monocular and binocular segments of the dorsal lateral geniculate nucleus in the cat. Journal ofComparative Neurology. 139, 413-422. Guiton, P., & Sluckin, W. (1969). The effects of visual experience on behavioural development in neonatal domestic chicks. British Journal of Psychology, 60, 495-507. Harlow, H. F., & Harlow, M. K. (1969). Effects of various mother-infant relationships on rhesus monkey behaviors. In B. M. Foss (Ed.), Determinants of infant behaviour (Vol. 4, pp. 15-36). London: Methuen. Hasler, A. D., Scholz, A. T, & Horrall, R. M. (1978). Olfactory imprinting and homing in salmon. American Scientist, 66, 347-355. Henry, K. R. (1983). Lifelong susceptibility to acoustic trauma: Changing patterns of cochlear damage over the life span of the mouse. Audiology, 22, 372-383. Henry, K. R. (1984). Noise and the young mouse: Genotype modifies the sensitive period for effects on cochlear physiology and audiogenic seizures. Behavioral Neuroscience, 98, 1073-1082. Henry, K. R., & Bowman, R. E. (1970). Behavior-genetic analysis of the ontogeny of acoustically primed audiogenic seizures in mice. Journal of Physiological Psychology, 70, 235-241. Herbert, M., Sluckin, W., & Sluckin, A. (1982). Mother-to-infant 'bonding.' Journal of Child Psychology and Psychiatry, 23, 205-221. Hess, E. H. (1957). Effects of meprobamate on imprinting of waterfowl. Annals of the New York Academy of Sciences, 67, 724-732. Hess, E. H. (1959). Imprinting. Science, 130, 133-141. Hess, E. H. (1973). Imprinting. New York: Van Nostrand Reinhold. Hill, D. L., & Przekop, P. R. (1988). Influences of dietary sodium on functional taste receptor development: A sensitive period. Science, 241, 1826-1828. Hinde, R. A. (1962). Sensitive periods and the development of behaviour. Little Club Clinic in Developmental Medicine, 7, 25-36. Hinde, R. A. (1970). Animal behaviour: A synthesis of ethology and comparative psychology. New \fork: McGraw-Hill. Hinde, R. A., & Bateson, P. (1984). Discontinuities versus continuities in behavioral development and the neglect of process. International Journal of Behavioral Development, 7, 129-143. Hirsch, H. V. B., & Tieman, S. B. (1987). Perceptual development and experience-dependent changes in cat visual cortex. In Ml H. Bornstein (Ed.), Sensitive periods in development: Interdisciplinary perspectives (pp. 39-79). Hillsdale, NJ: Erlbaum. Hoffman, H. S. (1987). Imprinting and the critical period for social attachments: Some laboratory investigations. In M. H. Bornstein (Ed.), Sensitive periods in development: Interdisciplinary perspectives (pp. 99-121). Hillsdale, NJ: Erlbaum. Hubel, D. H. & Wiesel, T. N. (1970). The period of susceptibility to the physiological effects of unilateral eye closure in kittens. Journal of Physiology, 206, 419-436. Hudson, S. J., & Mullord, M. M. (1977). Investigations of maternal bonding in dairy cattle. Applied Animal Ethology, 3, 271-276. Hunt, C. E., & Brouillette, R. T. (1987). Sudden infant death syndrome: 1987 perspective. Journal of Pediatrics, 110, 678-699. Hunt, H. F, & Otis, L. S. (1963). Early "experience" and its effects on later behavioral processes in rats: I. Initial experiments. Transactions oftheNew YorkAcademy of Sciences, 25, 858-870. Immelmann, K. (1972). Sexual and other long-term aspects of imprinting in birds and other species. Advances in the Study of Behavior, 4, 147-174. Immelmann, K., & Suomi, S. J. (1981). Sensitive phases in development. In K. Immelmann, G. Barlow, L. Petrinovich, & M. Main

196

MARC H. BORNSTEIN

(Eds.), Behavioral development (pp. 395-431). New York: Cambridge University Press. Jaisson, P., & Fresneau, D. (1978). The sensitivity and responsiveness of ants to their cocoons in relation to age and methods of measurement. Animal Behaviour, 26, 1064-1071. Kasamatsu, T., & Pettigrew, J. D. (1979). Preservation of binocularity after monocular deprivation in the striate cortex of kittens treated with 6-hydroxydopamine. Journal of Comparative Neurology, J85, 139-162. Keverne, E. B., Levy, E, Poindron, P., & Lindsay, D. R. (1983). Vaginal stimulation: An important determinant of maternal bonding in sheep. Science, 219, 81-83. Klaus, M. H., & Kennell, J. H. (1976). Maternal-infant bonding. St. Louis: Mosby. Klopfer, P. H., Adams, D. K., & Klopfer, M. S. (1964). Maternal imprinting in goats. Proceedings of the National Acadamy of Science, 52,911-914. Landsberg, J.-W. (1975). Posthatch age and developmental age as a baseline for determination of the sensitive period for imprinting. Journal of Comparative and Physiological Psychology, 90, 47-52. Landsberg, J.-W., & Weiss, J. (1976). Stress and increase of the corticosterone level prevent imprinting in ducklings. Behaviour, 57, 173189. Lauder, J. M., & Krebs, H. (1986). Do neurotransmitters, neurohumors, and hormones specify critical periods? In W. T. Greenough & J. M. Juraska (Eds.), Developmental neuropsychobiology (pp. 119175). Orlando, FL: Academic Press. Leiderman, P. H. (1981). Human mother to infant social bonding: Is there a sensitive phase? In K. Immelmann, G. Barlow, L. Petrinovich, & M. Main (Eds.), Behavioral development (pp. 454-468). New \brk: Cambridge University Press. Lenneberg, E. H. (1967). Biological foundations of language. New \brk: Wiley. Leventhal, A. G., & Hirsch, H. V. B. (1975). Cortical effect of early selective exposure to diagonal lines. Science, 190, 902-904. Levine, S., & Lewis, G. W. (1959). Critical periods and the effects of infantile experience on the maturation of the stress response. Science, 129, 42-43. Levitt, F. B., & Van Sluyters, R. C. (1982). The sensitive period for strabismus in the kitten. Developmental Brain Research, 3, 323-327. Lorenz, K. L. (1935). Der Kumpan in der Umwelt des Vogels. Journal fur Ornithologie, 83, 137-213,289-413. Lorenz, K. L. (1937). The companion in the bird's world. Auk, 54,245273. Marler, P. (1970). A comparative approach to vocal learning: Song development in white-crowned sparrows. Journal of Comparative and Physiological Psychology, 71, 1-25. Marler, P., & Peters, S. (1980). Birdsong and speech: Evidence for special processing. In P. D. Eimas & J. L. Miller (Eds.), Perspectives on the study of speech (pp. 75-112). Hillsdale, NJ: Erlbaum. Martin, J. T, & Schutz, F. (1974). Arousal and temporal factors in imprinting in Mallards. Developmental Psychobiology, 7, 69-78. Mattson, M. E., & Evans, R. M. (1974). Visual imprinting and auditory discrimination learning in young of the canvasback and semi-parasitic redhead. Canadian Journal of Zoology, 52, 421-427. Millikan, G. C. (1972). The development of filial behavior in ducklings. Behaviour, 43, 13-47. Mitchell, D. E., & Timney, B. (1986). Postnatal development of function in the mammalian visual system. In J. M. Brookhart, V. B. Mountcastle, & H. W. Magoun (Eds.), Handbook of physiology: The nervous system (pp. 507-547). Bethesda, MD: American Physiological Society. Moltz, H. (1968). An epigenetic interpretation of the imprinting phe-

nomenon. In G. Newton & S. Levine (Eds.), Early experience and behavior (pp. 3-41). Springfield, MA: Charles C Thomas. Moltz, H. (1973). Some implications in the critical period hypothesis. Annals of the New York Academy of Sciences, 223, 144-146. Moltz, H., & Stettner, L. J. (1961). The influence of patterned light deprivation on the critical period for imprinting. Journal of Comparative and Physiological Psychology, 54, 279-283. Money, J., & Annecillo, C. (1987). Crucial period effect in psychoendocrinology: Two syndromes, abuse dwarfism and female (CVAH) hermaphroditism. In M. H. Bornstein (Ed.), Sensitive periods in development: Interdisciplinary perspectives (pp. 145-158). Hillsdale, NJ: Erlbaum. Montarolo, P. G., Goelet, P., Castellucci, V. F, Morgan, J., Kandel, E. R., & Schacher, S. (1986). A critical period for macromolecular synthesis in long-term heterosynaptic facilitation in Aplysia. Science, 234, 1249-1254. Moore, A. U., & Moore, M. (1960). Studies on the formation of the mother-neonate bond in sheep and goats. American Psychologist, 15, 413. Mower, G. D., & Christen, W. G. (1985). Role of visual experience in activating critical period in cat visual cortex. Journal ofNeurophysiology, 53, 572-589. Myers, B. J. (1984). Mother-infant bonding: The status of this criticalperiod hypothesis. Developmental Review, 4, 240-274. Nash, J. (1978). Developmental psychology: A psychobiological approach. Englewood Cliffs, NJ: Prentice-Hall. Novak, M. A., & Harlow, H. F. (1975). Social recovery of monkeys isolated for the first year of life: I. Rehabilitation and therapy. Developmental Psychology, 11, 453-465. Olson, C. R., & Freeman, R. D. (1978). Monocular deprivation and recovery during sensitive period in kittens. Journal of Neurophysiology, 41, 65-74. Over, R. (1979). Early experience and human development. Ada Psychologica Sinica, 11, 360-364. Oyama, S. (1976). A sensitive period for the acquisition of a nonnative phonological system. Journal of Psycholinguistic Research, 5, 261283. Oyama, S. (1979). The concept of the sensitive period in developmental studies. Merrill-Palmer Quarterly of Behavior and Development, 25, 83-103. Paulson, G. W. (1965). Maturation of evoked responses in the duckling. Experimental Neurology, 11, 324-333. Pettigrew, J. D. (1978). The paradox of the critical period for striate cortex. In C. W. Cotman (Ed.), Neuronal plasticity (pp. 311-330). New \brk: Raven Press. Pettigrew, J. D. (1982). Pharmacologic control of cortical plasticity. Retina, 2, 360-372. Pettigrew, J., Olson, C., & Barlow, H. B. (1973). Kitten visual cortex: Short-term, stimulus-induced changes in connectivity. Science, 180, 1202-1203. Pissonnier, D., Thiery, J. C., Fabre-Nys, C., Poindron, P., & Keverne, E. B. (1985). The importance of olfactory bulb noradrenalin for maternal recognition in sheep. Physiology & Behavior, 35, 361-363. Poindron, P., Levy, F, & Krehbeil, D. (1988). Genital, olfactory, and endocrine interactions in the development of maternal behaviour in the parturient ewe. Psychoneuroendocrinology, 13, 99-125. Poindron, P., Martin, G. B., & Hooley, R. D. (1979). Effects of lambing induction on the sensitive period for the establishment of maternal behaviour in sheep. Physiology & Behavior, 23, 1081-1087. Presson, J., & Gordon, B. (1979). Critical period and minimum exposure required for the effects of alternating monocular occlusion in cat visual cortex. Vision Research, 19, 807-811. Reinisch, J. M. (1974). Fetal hormones, the brain, and human sex

SENSITIVE PERIODS IN DEVELOPMENT differences: A heuristic, integrative review of the recent literature. Archives of Sexual Behavior, 3, 51-90. Reppert, S. M. (1985). Maternal entrainment of the developing circadian system. Annals of the New York Academy of Sciences, 453, 162169. Rosenblatt, J. S., & Siegel, H. I. (1981). Factors governing the onset and maintenance of maternal behavior among nonprimate mammals. In D. J. Gubernick & P. H. Klopfer (Eds.), Parental care in mammals (pp. 13-76). New York: Plenum. Rothblat, L. A., & Schwartz, M. L. (1979). The effect of monocular deprivation on dendritic spines in visual cortex of young and adult albino rats: Evidence for a sensitive period. Brain Research, 161, 151-161. Rothstein, S. I. (1978). Mechanisms of avian egg-recognition: Additional evidence for learned components. Animal Behaviour, 26, 671677. Sackett, G. P., Holm, R., & Landesman-Dwyer, S. (1975). Vulnerability for abnormal development: Pregnancy outcome and sex differences in macaque monkeys. In N. Ellis (Ed.), Aberrant development in infancy: Human and animal studies (pp. 59-88). New %rk: Halstead Press. Salzen, E. A. (1963). Imprinting and the immobility reactions of domestic fowl. Animal Behaviour, 11, 66-71. Schaefer, T. (1963). Early "experience" and its effects on later behavioral processes in rats: II. A critical factor in the early handling phenomenon. Transactions for the New York Academy of Sciences, 25, 871889. Schneirla, T. C., & Rosenblatt, J. S. (1963). "Critical periods" in the development of behavior. Science, 139, 1110-1115. Scott, J. P. (1958). Critical periods in the development of social behavior in puppies. Psychosomatic Medicine, 20, 42-54. Scott, J. P. (1962). Critical periods in behavioral development. Science, 138, 949-958. Scott, J. P. (Ed.). (1978). Critical periods. Stroudsburg, PA: Dowden, Hutchinson & Ross. Scott, J. P. (1986). Critical periods in organizational processes. In F. Falkner & J. M. Tanner (Eds.), Human growth (2nd ed., pp. 181196). New York: Plenum. Scott, J. P., Stewart, J., & DeGhett, V. J. (1974). Critical periods in the organization of systems. Developmental Psychobiology, 7, 489-513. Sillito, A. M. (1983). Plasticity in the visual cortex. Nature, 303, 477478. Simon, N. (1979). Kaspar Mauser's recovery and autopsy: A perspective on neurological and sociological requirements for language development. Annual Progress in Child Psychiatry & Child Development, 215-244. Sluckin, W. (1964). Imprinting and early learning. London: Methuen. Sluckin, W. (1972). Early learning in man and animals. London: Allen & Unwin. Smith, F. V, & Van-Toller, C., & Boyes, T. (1966). The "Critical Period" in the attachment of lambs and ewes. Animal Behaviour, 14, 120125. Snow, C. (1987). Relevance of the notion of a critical period to language acquisition. In M. H. Bornstein (Ed.), Sensitive periods in develop-

197

ment: Interdisciplinary perspectives (pp. 183-209). Hillsdale, NJ: Erlbaum. Spalding, D. A. (1873). Instinct with original observations on young animals. Macmillan'sMagazine, 27, 282-293. Spemann, H. (1938). Embryonic development and induction. New Haven, CT: Yale University Press. Stockard, C. R. (1907). The artificial production of a single median cyclopean eye in the fish embryo by means of seawater solutions of magnesium chloride. Archiv fur Entwicklungsmechanik, 28, 249. Stockard, C. R. (1921). Developmental rate and structural expression: An experimental study of twins, "double monsters," and single deformities and their interaction among embryonic organs during their origins and development. American Journal of Anatomy, 28, 115275. Strobel, M. G., Baker, D. G., MacDonald, G. E. (1967). The effect of embryonic x-irradiation on the approach and following response in newly hatched chicks. Canadian Journal of Psychology, 21, 322-328. Suomi, S. J., & Harlow, H. F. (1975). The role and reason of peer friendships. In M. Lewis & L. A. Rosenblum (Eds.), Friendship and peer relations (pp. 310-334). New \brk: Plenum. Thomson, A. L. (1950). Factors determining the breeding seasons of birds: An introductory review. Ibis, 92, 173-184. Thorpe, W. H. (1961). Sensitive periods in the learning of animals and men: A study of imprinting with special reference to the induction of cyclic behavior. In W. H. Thorpe & O. L. Zangwill (Eds.), Current problems in animal behaviour (pp. 194-224). Cambridge: Cambridge University Press. Timney, B., Mitchell, D. E., & Griffin, F. (1978). The development of vision in cats after extended periods of dark-rearing. Experimental Brain Research, 31, 547-560. Tinbergen, N. (1951). The study of instinct. Oxford: Oxford University Press. Vidal, J. Y. (1975). Influence of social deprivation on self perception during the sexual behavior of the domestic cock. Behaviour, 52, 5783. Wiesel, T. N. (1982). Postnatal development of the visual cortex and the influence of the environment. Nature, 299, 583-591. Wiesel, T. N., & Hubel, D. H. (1963). Single-cell responses in striate cortex of kittens deprived of vision in one eye. Journal ofNeurophysiology,26, 1003-1017. Weiss, J., & Kohler, W, & Landsberg, J.-W. (1977). Increase of the corticosterone level in ducklings during the sensitive period of the following response. Developmental Psychobiology, 10, 59-64. Wyly, M. V, Denenberg, V. H., De Santis, D., Burns, J. K., & Zarrow, M. X. (1975). Handling rabbits in infancy: In search of a critical period. Developmental Psychobiology, 8, 179-186. Zajonc, R. B., Reimer, D. J., & Hausser, D. (1973). Imprinting and the development of object preference in chicks by mere repeated exposure. Journal of Comparative and Physiological Psychology, 83, 434440.

Received August 8,1986 Revision received February 19,1988 Accepted June 20,1988 •