Intersections With Attachment

Jacob L. Gewirtz, William M. Kurtines, Jacob L. Lamb

• Publisher: Psychology Press
• Year: 2013
• Topic: Psychology
• eBook ISBN:
  9781134739134
• Subtopic: History & Theory in Psychology
• Index: Psychology

Attachment has long been a key area of social development. Work on attachment processes has involved a variety of species as well as humans in diverse cultures and at various points in the life cycle. This volume presents research devoted to the meaning and implications of the attachment concept, including possible indices of attachment, the role of learning, whether or not attachment is best treated as continuous or discontinuous, and considerations for viewing attachment as a trait across environmental settings or as a process with functions that operate differently in disparate settings. Other psychological-process concepts, such as imprinting, relationships, and identification are also discussed. Because the contributors are active researchers and theorists, this volume may help establish trends and determine directions to shape literature on attachment for years to come.

I

The Environment and Attachment

1

Attachment as Personal Characteristic or a Measure of the Environment

Michael Lewis

Candice Feiring

Abstract

In trying to understand the etiology of human development, three basic models have been utilized: a trait model, an environmental model, and an interactional model. These three models, prototypes of various views of development, make clear how certain assumptions are used to understand the etiology of human behavior. The attachment model is the most accepted and utilized theory on the development of social relationships and has as its central thesis the assumption that one set of social experiences is directly connected to the next. Most studies on the outcome of early attachment behavior have not considered the nature of the child’s environment at the time the outcome status of the child is measured. The failure to do so is evidence of a paradigm which holds attachment to operate as a trait. Longitudinal data are examined to understand how attachment and early and subsequent environment predict behavior problems. The results suggest that attachment in fact may be a measure of environment as much as it is a measure of a child’s characteristic.

In the study of social development, certain beliefs are accepted as facts, because they fit the current Zeitgeist or paradigm (Kuhn, 1962). The paradigm under which most of social development operates is one that states that the child’s earliest social experiences impact on its later life and that these early social experiences are for the most part caused by particular parenting, specifically the mother–child interaction and relationship. Currently, beliefs exist that support this general paradigm. These beliefs are spoken of as facts, although there is considerable reason to doubt the validity of each one.

The first of these facts has to do with childrens’ responsivity to the social environment as an attribute of their humanness. This responsivity supposedly occurs immediately and is demonstrated by children’s preferences for human faces and face-like stimuli to all other visual stimuli. This finding is attributable to a result first reported by Fantz (1965), which was subsequently repudiated by him when the correct analyses were performed on the data. However, this unreplicated result has found its way into texts on development. Fantz’ early work did not control for stimulus complexity, and when controlled by presenting scrambled faces instead of a real face, no preference in the very young organism is found for human faces. Nevertheless, this fact, that children are from the moment of birth socially oriented toward conspecifics, remains a strong belief in spite of the evidence that this is not the case.

The second fact centers on the mother–child relationship and the issue of responsivity. By the late 1960s, the notion of responsive maternal behavior as a cause of appropriate social development had become well fixed (Lewis & Goldberg, 1969). Of concern to some scientists was the idea that responsivity was more appropriately defined as reinforcement. Reinforcement of specific behaviors, such as cry, should lead to more cry behavior. In contrast, the responsivity notion asserted that a responsive mother, rather than reinforcing crying and perpetuating the crying response, should satisfy the child’s basic need that the crying response reflects. Thus, it should not lead to more crying. In a study by Bell and Ainsworth (1972), maternal responsivity toward infant crying in the first three months of life was compared to infants cry behavior during months 10–12. The authors concluded that a mother’s responsivity to her infant’s cry did not lead to more crying behavior but led to more communicative behavior. The importance of this study is considerable, because it confirmed the belief that maternal responsivity, especially to infant cry, leads to positive consequences, not more crying. Although this study is widely referenced, Gewirtz and Boyd’s (1977) critique is not. Gewirtz and Boyd (1977) were able to show that the design and analysis of the study does not allow us to conclude what Bell and Ainsworth (1972) asserted. Nevertheless, the results of this study remain as facts, because they fit the common beliefs that responsive mothering produces socially healthy infants.

Harlow and his now classical work on motherless monkeys stands as the best demonstration of the effect of the lack of mothering experience on children’s subsequent peer and psychosexual development. Recall that Harlow (1958) demonstrated that children raised without their mothers showed severe problems as adults: difficulty engaging in sexual reproductive acts, and once the females had become pregnant, inadequate mothering. Many of these motherless monkeys were abusive and even killed their children. Harlows’ work (Harlow & Harlow, 1965; Harlow, 1969) stands as the best demonstration of the effect of the lack of mothers in one generation on the mothering behavior of the children in the next. More recently, this work has come under question from several sources. To begin with, the motherless monkeys were deprived of all social experiences. They were raised in social isolation. It would be difficult to conclude, therefore, that the effects observed in adulthood were due to the lack of the mother rather than the lack of any social contact. Even more critical, however, is the fact that these motherless monkeys behaved quite differently to their second child than to their first child (Suomi & Harlow, 1978; Suomi, 1978). If the poor mothering experience, or in this case the lack of mothering experience, led to poor mothering behavior in the next generation as a result of a trait, one would have anticipated that this characteristic of the monkeys should have extended to all children they raised. The evidence suggests that their mothering behavior was not fixed by their early lack of mothering experience. Thus, even if the major effect of being raised in isolation was the lack of mothering experience, it can not be said that these motherless monkeys were inherently abusive or bad mothers, because their maternal behavior changed significantly from first to second born infants.

In discussing child abuse, two issues emerge that are difficult to explain in terms of the consequences of the mother–child experience on the child’s subsequent social development. The first is the often repeated finding that human children who are abused by their mothers will become abusing mothers themselves. Unfortunately, as far as can be determined, all the studies that report such findings are retrospective in nature; that is, abused children are observed in clinics and data gathered on their parent who are found themselves to have been abused in their childhood. This retrospective finding creates a logically incorrect notion that abuse in one generation will lead to abuse in the next. Although this may be the case, there is no proper evidence to support this assertion, because a retrospective analysis of this kind does not allow for the observation of those children who are not abused but who have parents who were abused as children. Without prospective study and analyses, it is not possible to determine the effect of abuse in one generation upon abuse in the next.

Finally, some recent work on child abuse that suggests that the supposed link between early negative mothering and subsequent child disturbance is not a causal one. Cicchetti and Braunwald (1984), reported that there are significant numbers of abused and neglected children who are securely attached. This finding and ones like it raise serious questions about attachment as a measure of the parent–child relationship, but even more importantly, on the nature of the parent–child interaction and how it impacts on the child’s attachment status or mental health. If these securely attached children are truly secure and are also abused or neglected, then there would be strong evidence to suggest that unresponsive maternal behavior does not necessarily constitute the only variable that produces different types of attachment. In other words, poor mothering (e.g., abuse and neglect) does not necessarily lead to poor adjustment (e.g., insecure attachment).

These five fallacies of early childhood social behavior raise the serious question about the nature of early social experience as it impacts on subsequent development. The attachment paradigm as delineated by Bowlby (1969) and subsequently by Ainsworth, Blehar, Waters, and Wall (1978) represents the most extensively researched model of the belief that early experience impacts on later adjustment and as such will serve as our focus for the general broader model it represents.

The attachment model is the most accepted and utilized theory on the development of social relationships and has as its central thesis the assumption that one set of social experiences is directly connected to the next. More specifically, the model argues for a linear relationship such that the child adapts to one relationship, and from this one relationship, all subsequent ones follow. This model is characterized by three features: (a) sequence, (b) determinism, and (c) trait or structural quality. This last feature, the trait or structural quality of attachment, will be the focus of this chapter.

The Issue of Trait

The issue and controversy surrounding the notion of a trait or enduring aspect of personality has dominated much contemporary thinking (Pervin, 1983). The issue is raised again when considering the effect of one relationship on another and speaks to one of the differences between the epigenetic attachment model and the social network model (see Lewis, 1987). One example of a difference between the epigenetic and social network systems model can be observed by considering the issue of multiple attachment. That an individual child can have multiple attachments and that some of them may be secure whereas others are not suggests that attachment refers to relationships not to the quality or trait of a particular baby. Thus, by considering multiple attachments, we are forced to move from a consideration of a trait notion, a quality located in the child, to a systems notion in which the various relationships vis-a-vis a system influence the child. Simple attachment relationships give rise to endogenous qualities within the child as explanations for behavior, thus, an epigenetic cause, whereas multiple relationships give rise to exogenous forces such as the nature of the social network. In the attachment view, the mother–child attachment relationship endows the infant with a trait or characteristic-like structure that is located within the organism, thus the expression an “A” baby or a “B” baby. This trait or its absence then determines subsequent relationships. The nature of the trait has not been clarified, however. Sroufe (1979) and Block and Block (1979) for example, have associated it with ego skills. Although such a notion is reasonable, the trait could be self-esteem or self-efficacy or some combination of the two. Whatever its nature, it is the presence or absence of the trait-like structure that influences other relationshíps. An often-used metaphor is that the child is like an empty vessel that needs filling. Once filled, the child can move on to new relationships. If the child is not filled, movement will be inhibited, or the new relationships will differ in their nature or degree from those of the filled child. Thus, the task of the earliest attachment relationship is to fill the vessel, or as is sometimes thought, to give the child those skills necessary to develop normally. In these explanations, the force for the establishment of subsequent relationships rests with an adaptability coping mechanism that resides within the organism.

The notion of a trait provides a mechanism for the deterministic nature found in the attachment model. One relationship can affect another through the creation of a trait within the child. The child then brings this trait to bear in its next relationship. Moreover, this trait or its absence based on the outcome of the first relationship may not easily be affected by experience. Thus, for example, a securely attached infant is competent at a later age due to the possession of that competence attribute or is incompetent due to its lack. The use of the term attribute does not imply a static or unchanging characteristic but merely implies that some aspect of the child is affecting his or her behavior. Such an explanation is less than adequate when applied to attachment given that the attachment classification of an individual child appears stable only as long as the environment in which the child functions remains stable (see Thompson & Lamb, 1983a, 1983b; Thompson, Lamb, & Estes, 1982; Vaughn, Egland, Sroufe, & Waters, 1979; Vaughn, Gove, & Egland, 1980; Waters, 1979). Such findings suggest that environmental attributes must be considered and must play some role in developmental outcomes in psychopathology (Mischel, 1968). Alternatively, one might view children’s social development in terms of environmental attributes. Here the early behavior of the child is dependent on the environment, and, thus, the attachment classification can be seen as a manifestation of the environment rather than the child. Changes in the environment will be reflected in changes in the child’s behavior or attachment classification. The competence of the child at a later point in time reflects a concomitant positive environment. The relationship between the child’s earlier and later behavior is not mediated by the attributes of the child but by the fact that the environments at both earlier and later points in time (i.e., when the original attachment classification was observed, and when the child’s behavior is subsequently observed) are related. Unlike a static trait, the behavior of the child exists only so long as the environment supports or maintains that behavior. The strong environmental or situational view is that the secure attachment and later social competence are no more than measures of a positive environment at two points in time. As we shall see, it is interesting that this model is rarely if ever tested in spite of the fact that there are data to do so. A third model, one in which one looks at both the nature of a social environment as well as the attribute of the child may be necessary to explain children’s subsequent social behavior and adjustment. In such a model, stability and change need to be seen as a function of both factors, the characteristic of the child as well as the environment in which the child lives. An infant, for example, who is securely attached as a function of a positive environment in the first year will show competence at a later age as a function of both the early secure attachment as well as the nature of the environment at a later age. Not addressed is whether the attribute exists as an independent factor and interacts with the current environment to produce a new set of behaviors or whether the attribute itself is transformed by the current environment thereby producing a new set of behaviors. The former view would be favored by an interactional approach (see Lewis, 1972), whereas the latter would be favored by a transactional view (see Sameroff & Chandler, 1975). For example, does a securely attached child at time t, show negative behavior at t₂ because: (a) secure attachment at t₁, has interacted with a negative env...

Table of contents

• Cover Page
• Half Title page
• Title Page
• Copyright Page
• Contents
• Dedication
• Preface
• Contributors' Biographies
• I The Environment and Attachment
• II Imprinting and Attachment
• III Early Communication and Attachment
• IV Stress, Temperament and Attachment
• V Conceptions Impacting Attachment
• VI Identification and Attachment
• VII Pathology and Attachment
• Author Index
• Subject Index