According to their specific goals for children, different cultures provide different child rearing environments and obtain different behavioral outcomes. As a result, the behavior of infants, young children and older individuals varies across cultures. Although we would not argue that these early experiences are completely formative by themselves, there tends to be continuity of socialization over time. The fundamental behavior patterns set into place during early socialization are further elaborated on as the child grows, and they persist into adulthood. Cultures may have markedly different goals and socialization practices. In addition, what is viewed in one culture as normal emotional learning and obviously the correct methods for achieving that learning, may be seen by another culture as strange, deficient or even pathological.

As will also become clear, although there is considerable work on early emotional learning (during the first 6-8 months), there is little work explicitly and directly relating this early learning to later behavior. This paper will conclude with some suggestions for doing so. In particular, it is suggested that early stressful experiences may result in a differential ability to handle stressful experiences later in life. The mechanisms by which this is accomplished are: a) that early child care practices that produce stress in infants, may result in higher levels of cortisol on a long term basis, and b) that certain emotional behaviors may be learned subcortically during the first few months of life, and that these behaviors will persist.

Much emotional and interactive behavior is learned during the first 6- to 7-months of life, and much of this learning takes place subcortically. According to Emde and his colleagues (1976 Emde and his colleagues (1976) as well as others (e.g. Emde and his colleagues (1976 Fischer & Rose, 1995) seven to nine months is the time of a major bio-behavioral shift. During this shift, changes take place in the frontal lobes of the cortex such that the cortex becomes more involved in planning and carrying out deliberate actions. These changes involve both myelination of the frontal cortex, the growth of connections between that area and other brain areas, and the death of some of the extra neurons present in these areas. Therefore, we would infer that before this biobehavioral shift is the period during which subcortical learning might be most prevalent.

Even after this biobehavioral shift, when the cortex is more involved in behavior, there is a variety of different evidence that suggests that subcortical processing may continue to be an important part of learning and experiencing emotions. For example, Todd et al. (1995) Todd et al. (1995) summarize evidence suggesting that myelination continues in some areas of the brain until 6 years of age (for example, for the primary motor and sensory areas of the brain), and in a few areas (for example, some of the associative areas) even beyond that. Other evidence suggests that at least some emotions, such as fear, may be learned subcortically at least into early childhood (for example, Jacobs and Nadel Izard & Harris, 1995 Izard & Harris, 1995; Jacobs and Nadel, 1985; or LeDoux, 1987 LeDoux, 1987).

Today, we focus on experiences that infants have before the 7- to 9- month bio-behavioral shift takes place. A more detailed paper would show either that the continuation of practices into early childhood and beyond could strengthen patterns already established, or that a shift in practices might change the eventual outcome.

Rather than relying on data from one study, this paper integrates results from a variety of published studies. None of these data were originally collected with the purpose of this paper in mind. We feel that this gives us a relatively conservative method of evaluating the hypothesis that a large amount of emotional learning can take place early in life.

In examining how infant behavior develops in different cultures we will use the model proposed by Sigel (1985) Sigel (1985). Sigel proposes a model that relates parental beliefs and goals to parental behavior, which is then related to child behavioral outcomes. Others (e.g. LeVine et al., 1994 LeVine et al., 1994) have proposed similar models.

Parents in different cultural settings have different goals for their children. LeVine and colleagues (LeVine et al., 1994) contrast North American (that is, U.S.) parents with Kenyan Gusii parents. The U.S. parents, he argues have a pedagogic model. The major goals are for children to learn to feel emotionally independent from their parents and to develop interactive and language skills. We will outline a number of features of this model.

In order to accomplish their goals, American parents engage in a number of behaviors. Richman, Miller & Solomon (1988), Richman, Miller & Solomon (1988), Morelli et al. (1992 Morelli et al. (1992) and others have described the North American practice of placing even newborn infants in their own beds and often in their own rooms. In Morelli et al.'s sample of 18 North American families, not a single family reported bringing their newborn into bed with them, although 15 allowed the infant to sleep in a bassinet or crib in the parental room. For the vast majority, this lasted only for the first 3 months or so of life, at which time infants were moved into their own rooms. The Gusii would be quite shocked by these practices.

North American parents are also relatively tolerant of other separations: infants are left with unrelated nonresident babysitters on a routine basis by many parents. More extended separations continue to be tolerated by some. For example, parents of young infants may go away for the weekend and leave their infant with a non-resident grandparent. There are no data on the extent of this practice. One practice that reflects the degree of comfort of the culture as a whole with such separations is the fact that until relatively recently, parents were not allowed to stay with their infants and children in hospitals. Most traditional societies are aghast at this practice and have not followed it when introduced to it by Northern Europeans and Americans. The lack of co-sleeping and the tolerance for separation have been added by us to other practices of the pedagogical model seen as essential by LeVine et al (1994 LeVine et al. (1994).

As detailed by many investigators (e.g. Brazelton, Koslowski & Main, 1974 Brazelton, Koslowski & Main, 1974; LeVine et al., 1994) American mothers behave as if their infants have the capacity to be active participants in interactions with adults: they talk to, look at, smile at and otherwise actively interact with infants from birth. In the absence of actual vocal responses from these very young infants, mothers respond to burps, hand movements, and other behaviors as if these are conversational overtures. Accompanying this strong emphasis on verbal interaction is a deemphasis on physical contact when compared to many other cultures. For example, LeVine et al (1994 Richman, Miller & Solomon (1988), and Miller (1994) Miller (1994) detailed the low rate of touching and holding among U.S. mothers when compared with Kenyan Gusii mothers in one case and with Mexican mothers in another case. Miller (1994) has suggestedthat one could have high rates of touching and high rates of verbal interaction, but at least in the U.S., this does not occur.

In contrast to the U.S. parent's high responsiveness to positive vocal/verbal behavior, LeVine et al., 1994) reported that American mothers are relatively tolerant of infant crying, allowing it to continue for much longer than mothers in other cultures would. For example, Gusii mothers watching videotapes of U.S. mothers were upset by how long it took these mothers to respond to infant crying.

In contrast to the pedagogical model, LeVine (LeVine et al., 1994) described some cultures as having a pediatric model, that is, a model in which the most important goal is protecting the health and survival of the infant. Although he used this term specifically in describing the Gusii of Kenya, aspects of this model apply to other cultures as well (and perhaps particularly to cultures in which infant mortality is still high or has only recently decreased).

In many cultures, including the Gusii studied by LeVine et al. (1994) and the Mayans studied by Morelli et al. (1992), mothers and infants sleep together. This practice generally lasts until the mother's next child is born. Sleeping together allows the mother to more easily breast feed the infant, and breast feeding on demand both at night and during the day typically continues into the second year, and perhaps beyond. LeVine et al. (1994) argue that frequent breast feeding has played an essential role in cultures where infant mortality was very high, by helping to ensure early weight gain and the possible maintenance of hydration in the presence of diarrhea.

Mothers whose behavior can be described as following the pediatric model tend to respond rapidly to crying, and show high rates of holding and touching. By responding rapidly to infant distress, and otherwise keeping the infant as quiet as possible (not encouraging other types of interaction that may excite the infant too much), and by engaging in higher rates of holding and touching (Miller, 1994) mothers in these cultures aim to produce children who are quiet, and when older are respectful and obedient to adults. As shown by experimental studies (Barr, 1990 Barr, 1990) increased holding by mothers tends to result in infants who cry less. LeVine argues that minimizing caloric expenditure due to excessive crying and too much activity may have improved infant survival under difficult conditions.

Mothers in cultures like the Gusii do not typically talk to their infants extensively (e.g. Ochs & Schieffelin Ochs & Schieffelin, 1984; LeVine et al., 1994) nor, in the case of the Gusii of Kenya at least, engage in much or any eye-contact with them. They do not see infants as capable of communicating or of understanding language and so do not engage in the type of stimulating face-to-face interactions with them that are seen among U.S. mothers and infants. Middle class mothers from the U.S. tend to find this bizarre.

What do infants learn in these situations? Much of the existing evidence has focused on the more overt, cortically controlled behaviors. We know, for example, from the work of LeVine and his colleagues (LeVine et al., 1994; Richman, 1983 Richman, 1983) that older U.S. infants and children are more active and talkative than the Kenyan (Gusii) infants and children. What types of subcortically controlled behaviors might they also be learning? There is less actual evidence of these types of behaviors, but knowing something about what systems are involved we can offer some possibilities for future exploration. This discussion will focus on nonverbal behaviors and on emotions.

A first difference might involve an infant or child's general level of arousal. Because interactions with U.S. infants involve a lot of stimulation and reinforcement of vocal/verbal behaviors, and elicitation of smiling andlaughing, these interactions have been observed to involve high peaks of excitement, longer interactions that involve play and other exciting activities, and generally higher levels of arousal in the infants (Dixon, Tronick, Keefer & Brazelton, 1981 Dixon, Tronick, Keefer & Brazelton, 1981). Dixon and colleagues contrasted this with the style of interaction of the Kenyan (Gusii) mothers. These mothers engaged in the more exciting types of interactions for only very brief periods of time, and when their infants began to get too aroused and excited they were observed to avert their eyes and/or turn away. It is inferred that in general, the U.S. infants spend more of their time in highly aroused and excited states and infants in soothing/distress responsiveness cultures (like to Gusii) spend more of their time in less aroused and excited states. Later on, U.S. children may be more likely to be stimulus seeking, although Gusii children may not or may even prefer quietness, sedateness and low key interactions.

Due to relative tolerance of crying among U.S. mothers, and less holding, American infants are expected to be more fussy (Barr, 1990). It may be inferred that the holding and soothing strategies used by Gusii mothers, among others, produce infants that spend less time distressed. U.S. infants and children, therefore, are more likely to spend time in both excited and happy states of arousal and in distressed states of arousal. Infants who are soothed and held will be less likely to experience both peaks of excitement and peaks of distress. The question we will take up below is whether exposure to such peaks of both positive and negative arousal helps infants to cope better with later stresses, or not.

Finally, it is apparent that U.S. infants must learn to cope early with being alone and specifically with being separated from their mothers. Tennes, (1982) Tennes (1982) has shown that in human infants there is a positive linear relationship between amount of separation protest and the amount of cortisol secreted. The information we have about sleep patterns in American infants and children also suggests that these produce stress in them. Although we are not aware of studies that have measured cortisol levels in infants sleeping apart from their parents and those sleeping with their parents, there is some evidence that these sleeping practices are stressful for American infants. For example, bedtime rituals seem to occur in U.S. settings, where infants and children are put to bed at set times and in separate areas, but rarely in other settings (e.g. Morelli et al., 1992). These rituals may last up to an hour in some cases and seem to be a response to the difficulty the infant or child has with going to bed on their own. A majority of U.S. infants in the Morelli et al. study also required transitional objects such as pacifiers, "blankies" or stuffed animals. It is well known from studies of adults (as summarized by Fackelmann, 1998 Fackelmann, 1998) that cortisol is produced during all types of stressful events, and that high levels of cortisol seem to be associated with a number of effects, including low immune system functioning.

We are suggesting two possible early effects of early child rearing practices:

a) different emotional behaviors related to arousal versus lack of arousal, expectations for contact with consoling individuals when one is upset or alone, and differential anticipations for other features of interactions with others (Commons, 1991 Commons, 1991).

b) possibly long term differences in the release of cortisol and possible related changes in brain structure due to the relative stressfulness of certain practices

It is expected that much of the early emotional learning takes place subcortically and results in children who have quite different expectations from their interactions with other people. As discussed above, such subcortically-learned expectations may not be accessible to conscious awareness and therefore control and relearning.

One implication is that cultures that train for independence, and therefore self-responsibility, risk having more PTSD later in life. That is because these cultures do not support infants (nor adults) during stressful and traumatic events to the same extent as do cultures that emphasize dependence and mutual support. Whether distressed or not, Individuals are isolated or separated and do not experience as much physical contact, which is known to be soothing (it has been shown to reduce blood pressure and so on). Early stressful practices may produce lasting effects, as these areas of the brain are still developing. As has been suggested by numerous investigators, the pathways that remain in place after early development are very determined by experience (as summarized in Todd et al., 1995). There may therefore be permanent alterations in stress-related neurotransmitter systems (such as the release of higher levels of cortisol).