Complicating Causes.

Witkin, Stanley L.

Causality may be considered as a mode of perception by which we reduce our sense impressions to order.

Niels Bohr

Jim, a social worker with a state child welfare agency, is interviewing an adolescent male (Chad) who was arrested recently for robbing a convenience store. Part of the interview went something like this:

Jim: Why did you rob the store?

Chad: I dunno. I needed the money to get my girlfriend something for her birthday.

Jim: You did it for her?

Chad: Yeah. But I wasn't planning on robbing the store. I was in there buying some stuff when the clerk went into the back room. I had just seen him put a bunch of bills in the drawer and when he left I just sort of reached over and opened it. It wasn't like I planned it.

Jim: So you did not intend to rob the store until you saw the money and the clerk left the room.

Chad: Right, it was an impulsive sort of thing.

Jim: Do you consider what you did wrong?

Chad: Yeah, but not as wrong as if I had planned it out.

Jim: Well, could you have not taken the money?

Chad: I guess so, but it all happened so fast.

Although fictitious, conversations like this one between social workers and their clients are not uncommon. We want to understand the causes of peoples' harmful actions and how they understand causes, so we can help them lead more productive lives. Causal discourse and the beliefs that underlie it are so common that the complexity of causality is easily overlooked. Such circumstances may reduce the depth of our analyses and our options for understanding and responding to different situations. Therefore, my aim in this editorial is to complicate our thinking about causal reasoning and by doing so loosen its taken-for-granted place in our practice and research discourse.

In common parlance, the word "cause" has to do with something that leads to a specific change in something else. My dictionary defines it as, "a person or thing that acts, happens, or exists in such a way that some specific thing happens as a result; the producer of an effect" (Random House Webster's Unabridged Dictionary, 1999). Despite this seemingly unambiguous meaning, identifying causes, or more generally the nature of causality, turns out to be a complicated issue that has been debated by philosophers for centuries. Questions such as "Must a specific effect always occur in the presence of a presumed cause?" Or, "Must a cause always precede an effect?" have proven more equivocal than my dictionary implies. Part of this complication is related to the relationship between cause and the broader philosophy of determinism. The latter, which is usually attributed to the French astronomer and mathematician, LaPlace (1749-1827), posited that everything that happens does so because of some prior force (see, for example, Green, 1995). Therefore, theoretically at least, a person with infinite knowledge could predict everything that would happen. This doctrine anchors one side of the debate about free will, another longstanding and complex issue.

Although I am not sure where social workers stand on the free will-determinism continuum, I do know that we are interested in causes, particularly those that can be manipulated to produce desired change. This interest extends beyond the causes of a particular event for a particular client, as in the case of Chad, to the identification of general causes for use with future clients.

The notion that events have causes is an old one, dating to at least the 5th century BC according to Green (1995). As noted in the opening quote by Nobel physicist Niels Bohr, this belief is probably related to the human need to create order in the world. Our Western intellectual traditions, particularly as embodied in science, have largely embraced a view of causality that emphasizes determinism (Rubin & Babbie, 2001) and a relatively narrow view of cause as something that brings about change (sometimes called efficient causality). Indeed, for many Americans, the scientific quest to discover (and eliminate) the causes of those things that plague us represents the greatest hope for the future of our species.

But cause can be considered in other ways. For example, causes can be located in the properties or nature of the thing acted on. Thus, we may say that a glass shattered when hit with a hammer because brittleness is a property of glass (see Secord, 1984), or that a cat pounced when I pulled a string in front of it because it is in the nature of cats to jump on things that might be their prey. And if you buy Bohrs' quote, you might argue that causality is important because it is in the nature of humans to create order.

Aristotle distinguished between four types of causes and thought the more types one could explain about an event the better the explanation (Buss, 1978). For example, one type, called final cause, referred to the aim or purpose for which a change is produced. Thus, I might say the cause of my writing this editorial was my desire to explore and share my thoughts on causality with other social workers. From the perspective of Western science such a view is "teleological"--that is, implying purpose--and therefore not suitable for scientific inquiry.

Today, the idea of efficient cause is widespread and taken for granted. What causes children to be violent? What are the causes of mental illness? Does cognitive-behavioral therapy "cure" depressed people? These kinds of questions are of interest to social workers. We assume that phenomena such as violence, mental illness, and depression have causes and that the identification of those causes will help us ameliorate, eliminate, or prevent them. This is a noble pursuit; however, it might also be helpful to consider other less discussed aspects of causal thinking (at least in the social work literature) and their implications for understanding and practice.

Identification of Causes Is Contextual.

One of the aims of an interview like the one with Chad is to give the social worker a more complete understanding of the situation in which Chad took the money. Practitioners recognize that this additional information is necessary to determine the causes of Chad's action. Actions are always situated actions and these situations impart meaning to clients' causal accounts. Change the context and the causes also might change.

When we try to generalize from an identified cause we remove it from its context. But, as Goergoudi and Rosnow (1985) pointed out, events do not have identities apart from their contexts--the meanings of each depend on the other. Extracting causes from contexts reifies them (makes them into things). However, what a person does in any given situation may or may not be what she or he would do in another situation or in the same situation at another time.

Causes and Effects Are Interdependent.

Causes are implied by effects.

When we engage in causal analysis there is a tendency to assume the reality of the event or the state that is the object of our analysis. Only "real" things are considered to have causes. If I asked, "Why do aliens abduct people?" most people would deny that there is an event to be explained. In some cases they would change the question to what they considered real, such as "why do people believe that aliens abduct people?" Although this example might seem a bit extreme, the same reasoning has been a source of contention between those who do research on the characteristics of a condition or syndrome (for example, borderline personality disorder) and those who question its reality as an independently existing condition.

In the case of Chad, the social worker wants to understand the causes for Chad's theft rather than his impulsiveness or foolish indiscretion, although it is conceivable that either of those terms could be substituted for "theft." Labeling Chad's action as theft directs us to look for certain kinds of causes, those that are associated with criminal behavior. It also leads Chad to generate certain explanations (or, in some cases, to dispute the label). Because concept, like theft can have multiple definitions, how we define them may imply different causes. For example, viewing depression as the expression of irrational cognitions, repressed hostility, or oppression invites us to seek its likely causes in either thought processes and beliefs, childhood events, or social factors.

The examples given suggest that identifying causes requires some understanding of the event or phenomenon to be explained. For example, when we ask about the causes of child abuse, we already have some understanding of what constitutes such abuse. This understanding implies the causes or at least generates what might be reasonable causal candidates. If we believe that corporeal punishment constitutes abuse, then in effect the cause is contained within the definition. Of course, we may wish to go further and ask what causes parents to use corporeal punishment on their children. Now we have made the cause the event to be explained (the effect), and once again, our definition of corporeal punishment points us to its causes. For example, if corporeal punishment is considered the use of physical force with the intent to inflict pain, then we might look for causes in areas like anger and personality disorder, because these make sense given our current understandings. People defining corporeal punishment differentl y will search for different causes. Thus, changing the definition of an effect can change its cause.

Causes can change effects.

Not only do effects imply causes, but changing causes can alter the meaning of an effect. Consider the following example. Mr. Jones has provided a large amount of time to a local community child care program. You ask a friend, who knows Mr. Jones, why he does this-that is, what is the cause of his actions-and are told that he lost his only child to leukemia and wants to ensure other children have a healthy start in life. Now contrast this explanation with one in which your friend tells you that Mr. Jones recently was found guilty of a crime and ordered to perform community service at the program as part of his sentence. Suddenly, the meaning of Mr. Jones' actions change.

Similarly, in the case of Chad, consider how the effect (theft) might change if the cause was believed to be any or some combination of the following: mental illness, ignorance of the law, the influence of a materialistic culture combined with lack of opportunity, the intense desire to win the heart of his girlfriend, unjust laws, the action of the convenience store clerk, or his abuse history.

Causality and Morals

The acceptance of a particular causal explanation is influenced not only by empirical considerations, but also by moral ones. For example, suppose a client tells you that she did not feed her child because she lacked money for food. Before assigning causality to this explanation, we are likely to consider whether she could have fed her child and whether she ought to have. The first consideration, could, refers to the voluntariness of her action, an important criterion in the assessment of responsibility (for example, see Shultz & Schleifer, 1983). If we believe that despite her claim of no money she could have fed her child--for example, by having applied for food stamps--we are likely to reject her proposed cause and hold her responsible for the situation. Other factors also may influence this judgment, such as whether we believe she knew about the availability of food stamps (Shultz & Schleifer, 1983). Even if we accept that she could not have done otherwise (on that day), we might reject the cause because we condemn the behavior. For example, we might believe that she should have budgeted more carefully, in which case poor budgeting rather than no money is likely to be considered the "real" cause.

Thus, causes are not so much discovered as negotiated. Like other concepts, the identification and legitimation of causes and effects are influenced by the dynamics of social interchange. In the example, power differences between the social worker and client may ultimately determine what counts as the "real" causes.

The example of Chad illustrates another aspect of the moral (and social) dimension of causal thinking. People accused of wrongdoing often respond in ways that justify their actions (for instance, the greater good). Or they may accept the cause but claim they were not responsible, that they could not have done otherwise. This response is common in legal settings where, for example, a defendant may claim innocence by reason of insanity. In such situations, language skills and knowledge of causal beliefs (and the law) may be the most important factors in providing convincing justifications.

Why Should We Care?

Causes matter. They matter because what we believe about causes and their relationship to "effects" has consequences for the people with whom we work. Consider the history of causes believed to underlie conditions such as developmental or physical disability, mental illness, depression, or poverty. As the causes changed so did the effects-that is, who was identified as having these conditions-and the way people so labeled were treated.

I have tried to show that causes and effects are largely social phenomena created in social interaction and through language. This view contrasts with the belief that causes exist apart from our experience--for example, as attributes of people or events. Viewing causes as social constructions encourages greater plasticity in our analyses by legitimizing alternative forms of causal or even noncausal explanations. It also may sensitive social workers to the interpersonal dynamics and implications of causal identification and interpretation. The widespread belief and interest in efficient causality reflects, in large part, the dominance of scientific ideology--and its control orientation--in our understanding of the world. Although this orientation may have certain benefits, it has other implications that deserve consideration; for example:

* Research that searches for generalized causes tends to redirect attention away from cause as a "language game" and its role in regulating social life.

* Research, specifically experimental research, that focuses on "manipulable" causes and operationalized effects produces, at best, partial analysis, reducing complex situations to elemental components (Goergoudi & Rosnow, 1985). In addition, it may draw attention away from larger, less manipulable factors that form the context of, and give meaning to, the phenomena we wish to address.

* The pervasive search for underlying causes encourages belief in a deterministic universe and directs attention away from alternative explanations of events, conditions, and situations.

I am not suggesting that we do away with the notion of cause, but that we broaden our understanding of this concept and lessen its privileged status. Along with the "why" of causal analysis, we need to make room for the "how" of interpretive analysis (Denzin, 1989). In the best tradition of social work we need to maintain a critical stance toward dominant cause and effect relationships (such as psychotropic drugs and psychological conditions) and always consider the larger social and political context. Finally, we must always ask how claims about the "real world," be they causal or otherwise, fit with our professional values about human dignity and social justice.

References

Buss. A. R. (1978). Causes and reasons in attribution theory: A conceptual critique. Journal of Personality and Social Psychology, 36, 1311-1321.

Denzin, N. K. (1989). The research act (3rd ed.). Englewood Cliffs, NJ: Prentice-Hall.

Goergoudi, M., & Rosnow, R. L. (1985). Notes toward a contextualist understanding of social psychology. Personality and Social Psychology Bulletin, 11, 5-22.

Green, R. (1995). The thwarting of LaPlace's demon. New York: St. Martin's Press.

Rubin, A., & Babbie, E. (2001). Research methods for social work. Belmont, CA: Wadsworth.

Secord, P. (1984). Determinism, freewill and self-intervention: A psychological perspective. New Ideas in Psychology, 2, 25-33.

Shultz, T. R., & Schleifer, M. (1983). Towards a refinement of attribution concepts. In J. Jaspars, F. D. Fincham, & M. Hewstone (Eds.), Attribution theory and research: Conceptual, developmental and social dimensions (pp. 37-62). New York: Academic Press.