What Is Meant, and Measured, by "Education"?
Anthony T. Carter
In recent years, as fertility declines have been observed to follow or accompany large expansions of education in societies that "have not experienced marked economic growth or industrialization" (Singh and Casterline, 1985:201), education, particularly the education of women, increasingly has been singled out as a prime determinant of fertility decline. That education should have such consequences comes as no surprise to those of us who read and write academic papers. Products of a powerful education system and now mostly educators ourselves, we "know" that education is a wonderful thing. At our worst, we may do little more than pour or hammer a few facts into our students, but, like Dickens' Thomas Gradgrind, we are prone to think that fact is always preferable to fancy. At our best, we like to think, we draw our students into the practice of critical thinking, helping them to replace unreflecting acceptance of traditional beliefs with proven knowledge and the capacity to make reasoned choices. Nevertheless, the source of education's efficacy remains mysterious. As the LeVines have noted, there is "scant information" on the precise mechanisms through which "the formal education of women affects their reproductive and health behaviors'' (LeVine et al., 1991:459).
This chapter is a critical discussion of selected literature in demography, anthropology, and cognate disciplines on the nature and consequences of education. It begins by outlining two sharply contrasting views of education. In one, education is seen as a single autonomous process of internalization. In the other,
education is regarded as a diverse collection of socially situated practices. In the second and third sections of the paper, I argue, first, that the autonomous concept of education pervades the literature on education and fertility change and, second, that it is unsatisfactory. In the fourth section of the paper, I sketch some of the implications of the alternative concept of education as socially situated practices for further research on fertility change. The final section presents some final reflections on the relationship between education and fertility from the anthropological point of view.
Though this volume is concerned with the demographic consequences of education in the developing world, I do not confine myself to examples from developing societies. To ignore the many critical discussions of education in the United States and other industrial societies is to leave unexamined what appears to be a highly optimistic if not ideologically loaded view of the universal effects of education.
Two Concepts Of Education
Education as an Autonomous Process of Internalization
According to a view that is widespread in the social sciences, formal education or schooling may be conceived of as an autonomous process of intellectual internalization (see Table 3-1). Knowledge, cognitive skills, and values that are originally external to students are made part of their internal mental apparatus. The agents of education are teachers; schools; the already-educated middle class; and, especially when education is being introduced into the developing world, "the West." Students, the objects of education, are passive vessels to be filled. Formal education is accomplished through the decontextualized use of language in schools that are removed from the contexts and values of everyday life. In this it is radically unlike apprenticeship and other forms of informal education in which what is learned is conditioned by and cannot readily be carried outside of specific settings. The longer one spends in school, the more information, ideas, and values are transferred or the more firmly they are inculcated. Education ends when schooling is completed. After one leaves school, passively internalized knowledge is put to use. The products of education are modern individuals. Education operates on and, indeed, makes individuals. This process is universal. Autonomous with respect to its social and cultural environment, education makes individuals and thus transforms society in the same way without regard to the historical situation in which it occurs.
Education as Socially Situated Practices
In the past two decades, a sharply contrasting view of education has been
TABLE 3-1 Two Contrasting Concepts of Education
Education as an Autonomous
Process of Internalization
Education as Socially
• The transfer or internalization of given information, ideas, and values
• ". . social process[es] involved in instructing, acquiring and transforming knowledge" (Pelissier, 1991:75)
• Learners are passive recipients
• Learners are coparticipants in pedagogical practices
• Accomplished by means of decontextualized language in the value-neutral, context-free setting of the school
• Accomplished by means of activities that are embedded in a local context and vary from one locale to another
• Formal education is radically unlike apprenticeship and other kinds of informal education
• Formal and informal education are differently situated, but not more or less situated
• After school, what is learned is put to use
• The trajectories of learners from peripheral to full participation in communities of practice extend beyond the school
• Education operates on and, indeed, makes individuals
• Education operates on relationships
• The effects of education are universal
• The effects of education are contingent on the historical and sociocultural situation in which it occurs
developed (again, see Table 3-1).1 According to this view, formal education may be conceived of as situated practices of interaction among teachers and students in school settings. Attention must be given to the agency of learners as well as to
teachers, school systems, and bodies of knowledge. Students are not mere passive recipients of education, but are active participants in the construction of pedagogical relations and of knowledge. Education is accomplished by means of communicative practices and other activities that are embedded in local contexts and vary from one locale to another. Learning is not confined to the classroom. It continues after schooling is completed. Education operates on relationships among persons and may or may not produce individuals. Education is not an autonomous process producing uniform results. Instead, its effects are contingent on the historical and sociocultural situations in which it occurs.
The origins of the conception of education as socially situated practices lie in a series of movements, in anthropological and psychological studies of teaching, learning, and language, across a major theoretical divide.2 On one side of the divide are positions that share a deterministic view of cognitive skills, knowledge, and teaching. On the other side, there is a convergence on a position that emphasizes practice, activity, and agency. In cross-cultural studies of cognition, the shift is from "a concern with cognitive properties as static phenomena that people do or do not have in their heads" to an interest in "the embeddedness of cognitive skills in particular interactive contexts rather than in isolated minds" (Pelissier, 1991:80). In studies of socialization, the parallel shift is from concern with the passive internalization of "given" cultural norms, values, beliefs, and behaviors to interest in how the "recipient[s] of sociocultural knowledge" are ''active contributors] to the meaning and outcome of interactions with other members of a social group" (Pelissier, 1991:83-84, quoting Schieffelin and Ochs, 1986:165).
Work along these lines has two foci. One focus is the interactions of students and teachers inside classrooms. The other is the ways in which schools are embedded in and permeated by the wider society in which they are located. Both foci figure prominently in the work of the anthropologist Jean Lave, which, because it takes us out of the classroom, along the life course, and into the worlds of persons who are old enough to form families, I describe in some detail.
Lave's first major work on education and learning, Cognition in Practice (1988), is a study of arithmetic problem solving among grocery shoppers and participants in the Weight Watchers diet program in California. It begins with, and is organized around, a critique of transfer theory in psychology. Developed in the early years of this century, transfer theory is at the core of the concept of education as an autonomous process of internalization. It holds that knowledge may be conceived of as a set of tools for thinking learned in the value-neutral, context-free setting of the school and then transferred to the activities of everyday life (Lave, 1988:37).
Some of Lave's findings are consistent with transfer theory. The participants
in her Adult Math Project averaged just 59 percent on standardized tests of arithmetic skills. Since "years of schooling is a good predictor of [test] performance," these research subjects might be regarded as poorly schooled and numerically incompetent. However, other findings point to a very different conclusion. The same research subjects performed similar calculations nearly flawlessly during the course of supermarket shopping (98 percent) or in simulated "best-buy" problems at home (93 percent) (Lave, 1988:52-56). Such results, Lave observes, confound the presuppositions of transfer theory.
Math is the central ongoing activity in the test situation and should command resources of attention and memory greater than those available in the supermarket where math competes for attention with a number of other concerns. School algorithms should be more powerful and accurate than quick, informal procedures (that's why they are taught in school). Finally, 98% accuracy in the supermarket is practically error-free arithmetic, and belies the image of hapless [just plain folks] failing cognitive challenges in an everyday world. (1988:5859)
To explain these paradoxical observations, Lave (1988:97) offers the concept of "activity-in-setting." Against the standard view of cognition, decision making, and agency as intramental processes, this concept holds that agency is socially distributed. "'Cognition' observed in everyday practice is distributed—stretched over, not divided among—mind, body, activity and culturally organized settings (which include other actors)" (Lave, 1988:1).
Lave's next major contribution is the concept of legitimate peripheral participation (see Lave, 1989, 1991; Lave and Wenger, 1991). This concept continues her concern with practice and adds a concern with the life course or what she calls "trajectories" along which persons move from legitimate peripheral to full participation in communities of practice. Legitimate peripheral participation directs attention to the changing practices through which newcomers and old-timers engage with one another. No longer entirely intramental, "learning, thinking, and knowing are" conceived of as located in "relations among people engaged in activity in, with, and arising from the socially and culturally constructed world" (Lave, 1991:67, emphasis removed).
The concept of learning as legitimate peripheral participation was constructed around the literature on apprenticeship, but it is intended to apply to formal schooling as well. The special character of formal schooling rests on the claim that knowledge can be decontextualized and that learning opportunities can be separated from legitimate peripheral participation. However, Lave and Wenger argue that schooling always involves a "learning curriculum" as well as "teaching curriculum."
A learning curriculum consists of situated opportunities... for the improvisational development of new practice (Lave, 1989). A learning curriculum is a field of learning resources in everyday practice viewed from the perspective of learners. A teaching curriculum, by contrast, is constructed for the instruction
of newcomers. When a teaching curriculum supplies—and thereby limits—structuring resources for learning, the meaning of what is learned (and control of access to it ...) is mediated through an instructor's participation, by an external view of what knowing is about. The learning curriculum in didactic situations, then, evolves out of participation in a specific community of practice engendered by pedagogical relations and by a prescriptive view of the target practice as a subject matter, as well as out of the many and various relations that tie participants to their own and other communities. (Lave and Wenger, 1991:97)
The learning curriculum in schools thus remains a form of legitimate peripheral participation.
Education In Analyses Of Fertility Change
Analyses of the relationship between education and fertility appear to revolve around a series of disagreements. The areas of disagreement include the salience of microeconomic and macrosociological perspectives, the generation through which education affects fertility,3 the character of the socioeconomic conditions that facilitate or impede the effects of education, and the question of whether education alters fertility by changing family-size preferences or the implementation of preferences within marriage.4 The great majority of such studies rely on statistical analyses of census and survey data. At least one is based on ethnographic research. For all its surface disagreement, however, this literature is permeated by the view of education as an autonomous process of internalization.
Micro and Macro Perspectives
Consider, for example, the difference between, on the one hand, Easterlin's (1978) "The Economics and Sociology of Fertility: A Synthesis," Cochrane's (1979) Fertility and Education: What Do We Really Know?, and Kasarda et al.'s (1986) monograph on Status Enhancement and Fertility and, on the other hand, Caldwell's (1980) essay on "Mass Education as a Determinant of the Timing of Fertility Decline."5 The microeconomic approaches of Easterlin and Cochrane
and the microsociological approach of Status Enhancement all emphasize the ways individuals, or couples treated as if they were individuals, respond to the changing parameters of economic calculation. In response to the microeconomic theories of fertility change that came to prominence in the 1960s and 1970s, Caldwell restates the classic macrosociological version of demographic transition theory, now emphasizing the role of formal education. In Caldwell's view, an account of "the onset of fertility transition" is logically prior to and different in kind from microeconomic and other explanations of fertility differentials that occur after the demographic transition is under way or when it has been completed. "The onset of fertility transition" per se is a consequence of mass involvement in formal schooling. The effects of mass education on the onset of fertility decline are macrosociological in that they involve changes in values or conventions rather than in the parameters of economic calculation. Mass schooling does not simply produce changes in the costs and benefits of children at the margins. Rather, it causes traditional "family moralities" to be replaced by "a new, community morality that is eventually necessary for fully developed non-family production (whether described as capitalist or socialist), and that is taught, explicitly and implicitly, by national education systems" (Caldwell, 1980:226-229).
Despite their theoretical differences, the authors of all of these studies are content to define education by the ways in which it is conventionally measured. These measures include "literacy status, years of school attended, years of school completed, [and] the possession of certain levels of certification" (Cochrane, 1979:29; see also Kasarda et al., 1986:105-106, and Caldwell, 1980:233).6
The micro theorists, Cochrane, Easterlin, and Kasarda et al., agree that education is one of several features of the process of modernization that reduce fertility by increasing the supply of children, decreasing the demand for them, and decreasing the costs of fertility regulation. Together with the related growth of the mass media, the expansion of formal education increases the supply of children by raising natural fertility and lowering infant mortality. It accomplishes the former by "break[ing] down traditional beliefs and customs and thus undermin[ing] cultural practices, such as an intercourse taboo, which have had the latent function of limiting reproduction." It "improve[s] health conditions"
and thus lowers infant mortality "by diffusing improved knowledge with regard to personal hygiene, food care, environmental dangers, and so on." Formal education and mass media reduce the demand for children through both a "taste" and a cost effect. The taste effect comes about in two ways. Educated persons increasingly value "a 'liberated' life style for women, involving greater market work and less family activity." "Children, and the life style associated with them," come to be seen as an '''old' good." Education and exposure to mass media also may shift tastes from numbers of children to children of higher "'quality.'" The cost effect is a function of the fact that education improves women's income-earning opportunities and thus increases the cost of the time they spend in childrearing. Education and mass media lower "the subjective costs of fertility regulation by challenging traditional beliefs and encouraging a problem-solving approach to life." They decrease the costs of contraception in both money and time by providing better information.7 (Easterlin, 1978:1 10-12; Cochrane, 1979:143-44; see also Kasarda et al., 1986:98-103).
To return to Caldwell, his macrosociological analysis is constructed around five mechanisms through which mass education affects fertility:
First, it reduces the child's potential for work inside and outside the home.... Second, education increases the cost of children far beyond the fees, uniforms, and stationery demanded by the school.... Third, schooling creates dependency, both within the family and within the society.... Fourth, schooling speeds up cultural change and creates new cultures .... Fifth, in the contemporary developing world, the school serves as a major instrument—probably the major instrument—for propagating the values, not of the local middle class, but of the Western middle class (Caldwell, 1980:227-28).
Of these mechanisms, Caldwell regards the last three as the most important. As he notes, this emphasis distinguishes his work from that of the microeconomists. It permits him to shift the generational impact of education, stressing the effects on parents of the education of their children rather than the effects on childbearing of the education of parents. It also permits him to downplay the economic costs and benefits of education. Nevertheless, his emphasis on new family moralities remains perfectly consistent with the microeconomists' inculcation of new social values and consequent abandonment of tradition.
An Anthropological Contribution
A critical reading of the LeVine et al. (1991) paper on "Women's Schooling and Child Care in the Demographic Transition: A Mexican Case Study" confirms the preceding analysis. This important paper might be expected to offer a novel approach to the relation between education and fertility. It represents an excursion into demography by anthropologists whose major work has been concerned with childrearing and education. It is based on extensive direct observations as well as surveys designed explicitly to identify the missing links through which education influences fertility. In an earlier paper on the relation between culture and fertility, LeVine and Scrimshaw (1983) outline an approach that is sharply critical of conventional microeconomic accounts of fertility change.8 Nevertheless, in the end, LeVine et al. fall back on the conventional view.
Much of the LeVines' work is familiar. As in other demographic studies, education is measured and, in effect, defined by years of schooling. The effects of schooling on fertility conform to the classic pattern. In rural Tilzapotla, the expected number of children ever born falls sharply with increasing maternal school attainment. Women who never attended school are expected to have an average of 3.83 children. Those who completed 9 years at school are expected to have an average of 2.62 children, nearly a third fewer. In urban Cuernavaca, however, increased maternal schooling is not associated with large declines in the expected number of children ever born. Women with 2 years of schooling are expected to have an average of 2.47 children. This number falls to 2.24 for women with 9 years of schooling. These results are consistent with the observation that current practice of contraception is strongly correlated with mother's education in Tilzapotla, rising from 17.5 percent of women with no schooling to 57.4 percent of women with 7-9 years of schooling, but not in Cuernavaca, where the range is only from 88 to 92 percent. Curiously, years of schooling is strongly associated with length of engagement in the urban sample. Women with 1-5 years of education are engaged for an average of 1 1.6 months, while those with 79 years of schooling are engaged for an average of 21.6 months.
Following Caldwell (1980), Easterlin (1983), and Kasarda et al. (1986), the LeVine study hypothesizes that schooling disseminates a modern ideology of Western individualism, "marital egalitarianism and domestic independence" (LeVine et al., 1991:473). Such an ideology breaks down the dominance of the elders in the domestic group, promotes the autonomy of the nuclear family, and increases women's motivation to enhance the status of themselves and their children. These changes tend to reduce fertility in several ways. Women who are no longer under the domestic control of elder relatives and who have a more
egalitarian relationship with their husbands are more likely to be able to pursue their own goals and to seek and use information on child survival and family planning. Women who value education for their children find that the costs of rearing children are increased, while the economic returns are decreased. The authors suggest that these hypotheses are confirmed by their survey results. In the more nuanced Cuernavaca survey, women with more schooling reported more joint marital decisions, increased discussion of family planning with their husbands, higher aspirations for their children, and reduced expectations of help from and coresidence with children. In Tilzapotla, women with more years of schooling are more likely to read and to watch television on a regular basis, thus exposing themselves to information about family planning (LeVine et al., 1991:482-84, 489-490).
The novel aspect of the LeVine study derives from Robert LeVine's own work "on the classroom as a source of new models of adult-child interpersonal relationships and communication for schoolgirls from agrarian communities" (LeVine et al., 1991:486). The authors propose that "schooling leads women to reconceptualize child care as a labor-intensive task requiring a great deal of her own attention throughout the preschool years and that this concept ultimately reduces her willingness to bear more than a few children." This hypothesis rests on a broad contrast between agrarian and modern industrial societies. Agrarian societies are characterized by a ''protective style" of childrearing "which emphasizes physical nurturance and comfort" and by systems of apprenticeship in which learning takes place through "graduated participation." Modern industrial societies, on the other hand, are characterized by systems of formal education in which learning takes place through decontextualized verbal instruction in the classroom. In such settings, children achieve widely varying levels of competence. As a consequence, the school "introduces considerations of long-term competitive advantage into childhood." The core of the LeVines' hypothesis runs as follows:
As schooling becomes institutionalized, mothers who have acquired this model in the classroom increasingly prepare their children for school, engaging them in pedagogical interaction at younger and younger ages. This means verbal responsiveness to the child during infancy, which has the effect of producing a verbally active toddler who frequently initiates demands for maternal attention during the post-infancy years. Such children are on the average less compliant and more 'difficult' and 'exhausting' to raise, reinforcing the mother's assumption that child care is a labor-intensive task—requiring more of her time and energy than it did for her own mother (prior to female schooling in an agrarian community) and inducing her to bear fewer children (LeVine et al., 1991:48688).9
To test this hypothesis, 72 Cuernavaca women were observed interacting with their infants at home. Observers noted (1) the proportion of infant vocalizations followed by maternal speech, (2) the proportion of infant looks followed by maternal speech, (3) the proportion of infant looks followed by maternal looks, (4) the proportion of infant motor acts followed by maternal speech, and (5) the proportion of time mother held infant. Significant positive correlations with maternal education were found for (1) and (4) at all infant ages, for (2) at 10 and 15 months and for (3) at 15 months. Significant negative correlations with maternal education were found for (5) at 15 months. The authors conclude "that women who attend school longer acquire a conception of child care as a labor-intensive task requiring more attention for a longer period of time—a conception that may contribute to child survival and impede fertility" (LeVine et al., 1991:488).
The LeVines did not do home observations in rural Tilzapotla, but in their view several components of their Tilzapotla survey provide insights into the same processes. One of the survey questions asked about arrangements made for childcare when the mother left the home. As mother's education increases, the percentage of children cared for by the oldest child or left alone in these circumstances declines, while the percentage of mothers taking their child with them increases. The percentage of mothers reporting that they leave their child with another adult does not vary significantly with mother's education. This last is also the preferred arrangement, ranging from 57 to 60 percent. LeVine et al. argue that this finding "supports the hypothesis that schooling enhances a woman's concept of child care as a labor-intensive task to be carried out by the mother herself or by another adult" (pp. 488-89).
Another part of the survey tested mastery of abstract or decontextualized language by asking respondents to define nouns and complete simple syllogisms. Following Snow (1990), LeVine et al. (1991:489-90) argue that exposure to formal schooling increases mastery of abstract or decontextualized uses of language in which the speaker/hearer must be able "to assume the perspective of someone who does not know the contextual background taken for granted in conversation" in order to communicate effectively. "This skill . . . is notably absent from normal conversation in small face-to-face communities, where people can safely assume that hearers share their contextual perspective." "Transmitted in the classroom," decontextualized language becomes a "pathway to the increased use of medical and contraceptive services. Specifically, women with more schooling are likely to be more conversant with the decontextualized language that is standard in health bureaucracies... ."10
LeVine et al. (1991:492) regard these as the primary effects of education:
Formal education everywhere, regardless of its quality, entails the presence of an adult whose role is entirely instructional, talking to children, often in a formal language they have to learn to understand. For girls in rural areas of countries where mass schooling is still a relatively recent innovation, this model of social interaction between an adult and children stands in contrast to their previous experience, and over time it reshapes their skills and preferences in social communication. They acquire in school and retain in adulthood skills of literacy and decontextualized language providing access to distant sources of information and institutionalized health care. Identifying with the role of pupil, they continue to seek useful knowledge wherever they can find it; identifying with the role of teacher, they are verbally responsive to their children during infancy and after. A new kind of mother-child relationship is built around reciprocal verbal interaction, one which helps mothers monitor the needs of their preschool children but which also demands so much of their attention that fertility control becomes imperative.
This conclusion carries Levine et al. some distance from an economic analysis of the consequences of education. However, its emphasis on the ways in which education promotes new values and enhances cognitive skills places it squarely within the conventional approach. The agency of young women in participating in the construction of classroom talk or in creatively applying what is learned in the classroom to the home is given short shrift. Like knowledge of abstract noun definitions, knowledge of the pedagogical style of childrearing is produced at school in a process in which the objects of education are passive recipients.
Formal Schooling And Decontextualized Language: An Ethnographic Test
With its pioneering attempt to trace connections between the experience of formal schooling and subsequent routine activities involved in fertility, the work of LeVine et al. (1991) provides an opportunity to test some of the central ideas
more likely to establish nuclear family households immediately following their weddings. If this is true generally, the observations of LeVine et al. on mother-child interactions and childcare arrangements must be reconsidered. A number of other studies, most carried out in North America, report that mothers often expand "the child's vocalization into a sentence which is in turn responded to by the mother in a conversational frame of alternating turns at talking," but careful reviews of this literature point out that most are "of first-born children reared in homes in which mother and child have only each other as a communicative partner for a large portion of the day" (Heath, 1983:374, note 7; see also Snow, 1977). Subsequent children or children born into extended family households may be reared differently. Similarly, the fact that more-educated Tilzapotlan women are more likely to take their children with them when they leave the house may simply be a function of the fact that there are no other adults or older children in their households or that the work is agricultural. On methods of observation, see note 1 1 below.
of the education-fertility literature against the ethnographic record. As noted above, LeVine et al. put particular emphasis on the peculiar features of classroom talk. They adopt the widely held view that classroom teaching universally includes the use of formal or decontextualized language. Mothers who have been to school are more likely to use a version of such language in caring for their own children. This transforms the care of children into a time-consuming activity and shifts more of the burden to mothers.
The Situated Character of Pedagogical Discourse
The work of Heath (1983) and Minick (1993) is prominent among the ethnographic studies relevant to this view of education and its consequences. Both Heath and Minick describe instructional speech practices in which routine contexts are suspended, and participants attempt to "construct close relationships between what is meant and what is said, between what is made known through an utterance and what is explicitly represented in language" (Minick, 1993:346). Such speech practices are what the LeVines and others call "formal" or "decontextualized" speech; Minick writes of "representational speech." To this extent, the positions of Heath and Minick are similar to the one adopted in the LeVine study, but the agreement ends there.
In the primary classrooms observed by Minick, representational speech often took the form of directives. Minick argues that the seemingly decontextualized character of representational directives is more accurately seen as a contextually defined accomplishment. The evidence for this argument consists of two sets of observations. The first concerns the fact that "formal training" in the representational directives genre "is a recognized part of the school curriculum, beginning with the introduction of what are commonly referred to as 'listening exercises' at the kindergarten level" (Minick, 1993:358). In these exercises, the students' task is limited to following the teacher's directions. Any "'situational sense' that might create an interpretation of intended meaning that would go beyond what is represented in a particular sentence or phrase" is excluded by the invocation of the listening exercise context. The invocation of this context is precarious: other, more routine contexts continually threaten to intrude. This is a consequence of the fact that the teacher utterances of which such exercises are composed are ambiguous. Teachers attempt to render their directives unambiguous and to protect the listening exercise context by increasing "the amount of 'information' actually represented in language. Rather than say 'point to' the teacher says 'put your finger on"' (Minick, 1993:360).
The second set of observations turns on the fact that in other kinds of classroom activities, it is often unclear that teachers intend their utterances to be understood as representational directives at the time they are made. In these instances, a representational interpretation of an utterance may be constructed
after the fact in response to the actions of students. In one episode discussed by Minick, a second-grade teacher working with four children in a "reading group"
.. . attempts to shift from a discussion of library books that are on the table in front of the children to work on a story in the basal readers that are under the children's chairs. The teacher initiates this shift by clearing away several notebooks that lie atop her copy of the reader while saying, "Now. We are going to read a story. Please put your books under your chair. And, we are going to read a story which you are going to enjoy."
One boy, Todd, promptly responds by putting his library book under his chair and putting his reader on the table. When the teacher begins to page through her reader, Todd looks through his.
Framed by the suggestion that "we are going to read a story" and the teacher's subsequent actions, Todd has apparently taken the teacher's directive as a [nonrepresentational directive] indicating that they are to prepare to do that reading. This task begins with putting the library books away, but also includes taking the reader out and locating the new story.
Another boy puts his library book away under his chair and sits back up, but when he sees "Todd looking through his reader, he begins to look nervously back and forth from Todd to the teacher." Two girls in the reading group put their library books away but then remain bent over, also looking from Todd to the teacher. When the teacher pays no attention to what Todd is doing, they, too, take "their readers out from under their chairs and begin to look through them." At this point the teacher takes note of the children's actions, saying in an irritated tone, "Todd, did Mrs. W. say, 'Open your book to...' Did she?" [Todd shakes his head, "No."] "No, she did not.'' In response to this reinterpretation of the teacher's initial utterance as a representational directive, all of the children put their readers back under their seats. The teacher begins "to review new vocabulary before beginning to read," but several of the children again looked "furtively" at their readers. Now the teacher lowers the representational boom, emphatically invoking the listening exercise context:
"Excuse me. I have not good listeners today. Now, put your hands on your books." [The "guilty" three immediately place their hands on their books.] "Put your books under your chairs." [The three immediately put their books under their chairs.] Here, the teacher makes several nested moves that make it clear that she is demanding a representational interpretation of directives. First, she begins with a directive that has a clear representational meaning but a rather opaque situational significance (i.e., "put your hands on your books"), encouraging a shift to representational interpretation. She follows this with a directive that is identical to that which initiated the episode, returning them to the position they would have been in had they followed a representational interpretation of her first directive (i.e., "put your books under your chairs"). (pp. 355-56)
In addition to representational directives, Heath (1983:279-83) describes a distinct genre of classroom talk connected with reading aloud and discussing the content of books. This genre is distinct from talk about "contextualized firsthand experiences." Like play, reading aloud and talking about reading "suspends reality, and is so framed, either through verbal or prop-type cues, that everyone knows immediately that it is not normal conversation." Reading involves "decontextualized representations of experience" insofar as it suspends the conventions in which first-hand experiences are discussed, but, as in the case of representational directives, this is a matter of reframing or recontextualization, rather than the removal of all contexts. The reframing is achieved in part through the use of "a particular kind of prosody which is different from regular conversational prosody.''11
Both Minick and Heath link the use of representational directives to classroom control. Minick (1993:361) found that teachers most often use representational directives to negotiate "comparatively mechanical task[s] that [have] little immediate pedagogical significance," tasks such as putting one book away and getting another out. Implicit in the use of the representational frame are differences in power. "The ubiquity of representational directives in the classroom," Minick suggests, "stems from the fact that the activities that are to be carried out there are defined by social realities such as curriculum, standardized testing, and teaching materials that have their roots in social systems that extend substantially beyond the classroom walls" (pp. 370-71).
That the exercise of power in the teacher/student relation cannot be reduced to the teaching curriculum or to differences in age is suggested by the fact that not all teachers manage their classrooms in the same way. A teacher who used representational directives less often also differed from others in Minick's sample in that
... she delegate[d] to her pupils much of the responsibility for defining 'situational sense.' Because she [did] not assume 'ownership of meaning' in organizing classroom activities, this teacher [was] not constantly faced with the task of conveying her definition of situational sense to her class, a task that demands either a nonrepresentational communication of this definition or a resort to the use of representational directives. (p. 370)
In the primary classrooms observed by Heath, issues of classroom control are linked to cultural diversity as well as power. Her study (1983) was designed to explain how children from two culturally distinctive groups in the Carolina
Piedmont responded to "mainstream" schools. Roadville was inhabited by whites, fairly recent immigrants from the Appalachians, most of whom worked in the textile mills. Trackton was inhabited by African-Americans. The "mainstream" was composed of white townspeople, many of whose "norms of conduct and bases for forming judgments about their own and others' behavior [had] much in common with the national mainstream middle class generally presented in the public media as the American client or customer" (Heath, 1983:236). In the decade 1969 to 1978, when Heath worked in the area, the residents of Roadville and Trackton were new to mainstream Piedmont schools, the first as a result of immigration and the second as a result of integration.
Heath (1983:279-83) notes that mainstream teachers expected their students to follow classroom rules. Lists of the rules were posted on bulletin boards, but in the first few months of school they were seldom stated explicitly or explained.
Instead teachers used indirect instruction and modeling for the children. They used familiar verbal formulae from their own home experiences: Is this where the scissors belong? . . Someone else is talking now; we'll all have to wait. We have visitors coming to the school this afternoon; we want our school to look nice.
That is to say, they used contextualized language. This had the desired effect with children from the same mainstream background as the teacher, children who had learned these forms of contextualized language at home. But "Roadville and Trackton children had difficulty interpreting these indirect requests for adherence to an unstated set of rules." The teachers perceived these children as impolite.
Many teachers learned to deal with this by expressing directives as directly as possible. Instead of 'Can we get ready on time?' as an indirect directive to tell children to put their toys away and begin lining up for snack time, teachers tried to say: 'put your toys back where you took them from. We have to line up for lunch. Table three will wash hands first.'
The Acquisition of Pedagogical Discourse
These observations cast doubt on the universal validity of the distinction between informal, contextualized language and formal, decontextualized language on which the distinctiveness of schooling rests. Heath's work also undermines the LeVines' argument that women who have attended school model their interactions with young children on the interactions between teachers and children in school. Heath was concerned with the effects on school success of prior participation in speech practices that differed from those used in the school, but it appears that her material can be read the other way around as well.12 The young
mothers Heath studied could read and write and certainly had attended school. Most likely, all remained in school until they were at least 16 years old. Most likely, too, given that Heath's research was carried out from 1969 to 1978, most attended mainstream schools for at least part of their educational careers. Nevertheless, it would appear that the principal influence on their childrearing practices was their position as legitimate peripheral participants in their residential communities and family networks.
Of the two kinds of distinctively contextualized speech practices observed by Heath, those involved in reading aloud and talking about reading were found in "mainstream" homes as well as in the newly integrated "mainstream" Piedmont schools. Representational directives appear to have occurred mostly in the school setting. Heath suggests that this is a consequence of the fact that they served as a technology of control where the participants in education did not have a shared sense of the rules. Though some "mainstream" childrearing practices were consonant with the practices of formal schooling, there is no evidence that schooling was the source of these practices.
Trackton and Roadville parents had culturally variable ways of interacting with and rearing children that were little affected by formal schooling. A brief discussion of one case will suggest the magnitude of the variation. African-Americans in Trackton speak of "children 'coming up"' (Heath, 1983:144). A new baby need not be taken home to a nuclear family household consisting of its parents and siblings. For example,
Annie Mae's daughter, Marcy, became pregnant at fifteen after a one-time liaison with Miner Baine, a boy of sixteen .... When the child—a boy—was born, Miner came around to visit infrequently, and brought milk and diapers occasionally in the first few months. Soon after he went into the service and left the area. Miner's family, however, enjoyed the baby, Larry Lee, and they took the child every weekend, and Miner's sister and parents lavished attention on the baby. Annie Mae became the baby's 'mamma,' taking major responsibility for the child. Larry Lee learned to call his biological mother by her given name, and Annie Mae was 'mamma.' Marcy, after the birth of the child, went back to school at night, then on to a technical college, working steadily at various jobs and helping her mother pay the expenses for Larry Lee. Mr. and Mrs. Smith, the old couple near the railroad tracks, became co-parents with Annie Mae of Larry Lee, keeping him a good part of the time. (Heath, 1983:69-70)
Regardless of its domestic arrangements, much of a child's time is spent outdoors, where it interacts with the wider community.
During the first year of life, Trackton babies are constantly in the midst of family and community life, often "carried astride the hip or nestled in the cradle of an arm" so that "they can see the face of their caregiver or the person the caregiver is talking to."
[They] are in the midst of nearly constant human communication, verbal and nonverbal. They literally feel the body signals of shifts in emotion of those who
hold them; they are never excluded from verbal interactions. They are listeners and observers in a stream of communication which flows about them, but is not especially channeled or modified for them. Everyone talks about the baby, but rarely to the baby." (Heath, 1983:75)
Adults regard the early vocalizations of children as noise.
[They] respond incredulously to queries such as 'Did you hear him say milk?' They believe they should not have to depend on their babies to tell them what they need or when they are uncomfortable. Adults are the knowing participants; children only 'come to know.' Thus, if asked, community members explain away their lack of responses to children's early utterances; they do not repeat the utterance, announce it as a label for an item or event, or place the 'word' in an expanded phrase or sentence. To them, the response carries no meaning which can be directly linked to an object or event; it is just 'noise'. (Heath, 1983:76)
Varieties Of Educational Experience: New Directions For Research
As Kasarda et al. (1986:106) observe, "cross-national comparisons of [quantitative] research findings" on the association of increased education with reduced fertility rest on the assumption "that educational experience is equivalent from country to country (in terms of economic value and opportunities availed) or that the standardized measures we use to summarize educational experience have the same meaning in all cultural settings." The core of this assumption is the idea, integral to the concept of education as internalization, that education is autonomous with respect to its social and cultural environment and that its effects are therefore universal.
The alternative view of education as situated processes of interaction abandons the search for universals. It directs attention instead to the ways in which schooling and its consequences vary with the practices of which it is composed and the social and cultural context in which it is embedded. This view opens up new ways of investigating the links between education and fertility.
Ideologies of Education
It is helpful to keep in mind that many of our ideas about the nature and value of education are themselves situated and contested. According to the conventional view of education employed in the literature on its demographic effects, education is liberating and empowering. In Easterlin's (1978:111) formulation, it "encourag[es] a problem-solving approach to life" by breaking down tradition, giving people access to information and "improved knowledge," and inculcating reasoning skills. This view is by no means universal. Advocates of formal schooling for working-class communities in nineteenth-century England (Vincent,
1989) and France (Furet and Ozouf, 1982) and twentieth-century South Carolina (Heath, 1983) all saw it as providing new channels through which the emerging class of industrial workers could be disciplined. Both views of education—that which equates it with liberation and empowerment and that which equates it with discipline—reflect the concerns of middle- and upper-class interests. They fail to recognize that schools
exist as a site of struggle between and among interests of the state, capital, labor, educators, community representatives, advocates, students, and parents. Schools distribute skills and opportunities in ways textured by class, race, and gender asymmetries. Privileged are notions of individualism, competition, mobility, meritocracy, and marriage, and silenced are discussions of social class, race, gender, and sexual arrangements. Finally, the presentation of public schools as 'public' obscures the vastly differential educations and outcomes made available to students by virtue of their social class, race/ethnicity, gender, sexual orientation, disability, and geography, and undercuts the creation of spaces which nurture democracy, difference, critique, and movements for social change. (Fine, 1991:199-200)
Conventional accounts of the relation between education and fertility erase educational histories, replacing them with years of schooling completed as a measure of education. No doubt this is in part simply a matter of research convenience, but it also is consistent with the concept of education as an autonomous process of internalization. Such a measure assumes that formal education is additive. The more years are spent in school, the more the individual is transformed. It is recognized that educational achievement may be affected by economic factors and the events in a woman's reproductive history (see, e.g., Kasarda et al., 1986:106), but schools also are assumed to provide the same sorts of opportunities to all categories of students. Within broad limits, therefore, the number of years of schooling completed can also be treated as an indication of the student's inherent capacity to benefit from education. In contrast, the concept of education as situated practices, with its emphasis on the active participation of students in educational processes and the trajectories of interaction described by learners as they move from peripheral to full participation in communities of practice, places educational histories in the foreground.
The available data on educational histories are fragmentary and incomplete, the byproducts of studies directed to other concerns. They are sufficient, nevertheless, to suggest that this would be a fruitful area for further research.
Harbison and Hanushek (1992) provide a rare glimpse of the educational histories of primary school children in rural northeast Brazil in the 1980s. The data are derived from a series of surveys designed to evaluate the effects of the Northeast Brazil Basic Education Project (EDURURAL) on access to primary schooling, progress through the primary grades, and educational achievement.
This study was "the most comprehensive survey of rural education ever attempted in a developing country" (Harbison and Hanushek, 1992:37). Nevertheless, as regards educational histories per se, the EDURURAL data have major limitations. The surveys were designed to sample schools, rather than to track individual students. It was expected that many of the second graders who were studied during one round of the survey would reappear as fourth graders 2 years later. However, this expectation was not met. Schools themselves led a precarious existence, often disappearing from one survey to the next. Students in schools that did survive had high rates of retention in grade and dropout (p. 41). The final, 1987, round of data collection was changed to compensate for these difficulties. Abandoning any attempt to produce a representative sample, this round concentrated on a local subset of schools and incorporated special efforts to locate second graders surveyed in 1985. Because the surveys were collected at 2 year intervals, the educational histories derived from them are discontinuous. The survey data themselves are limited in at least two respects. The investigators were unable to obtain "reliable direct measures of attendance" (pp. 97, 42). They measured student work with a crude "dummy variable reflecting employment status," thus setting aside "wide variations in time commitments, in strenuousness of the activity, or in effects on attendance or homework" (p. 326, note 110). No observations were made in classrooms.
Figure 3-1 is Harbison and Hanushek's (1992:60) representation of the "possible paths for a student initially observed in second grade." If they followed the standard course, the educational histories of all primary students would carry them along the path on the extreme left of Figure 3-1. That a great many students in rural northeastern Brazil deviated from this standard path is only partly a consequence of high rates of school demise, the inability of many schools to provide a full set of primary grades, and the geographical mobility of students' families. Progress through this fragile school system also was affected by gender; age; performance on initial standardized tests; and characteristics of the student's family, school, and region.
Additional complexity is introduced by the interaction between the student's experience in school and his or her role in rural agrarian households. The families of students in counties with higher agricultural productivity were less likely to migrate. Students who belonged to families in which the father was a farmer were more likely to drop out. Boys and older students were more likely to drop out than girls and younger students. Harbison and Hanushek (1992:70) suggest that the difference between boys and girls "reflects a lower opportunity cost of school attendance for girls; their value on the farms is less, so they are less likely to quit school to work." Since the work done "by second graders typically involves lesser time commitments and therefore is less intrusive on schooling" (p. 98), opportunity costs also may have been involved in the differences between older and younger students. Among students who remained in school and were promoted from second to fourth grade on time, work had a negative effect on
achievement as measured by changes in scores on standardized tests of Portuguese and mathematics. Underlying these findings is the undoubted but also unobserved fact that the educational histories of "many students [consist of] on-and-off attendance over extended periods of time" (p. 97).
Harbison and Hanushek (1992:100-103) also explore the factors contributing to achievement among students who remained in school and were promoted from one grade to another on time. Their findings can be read to suggest that students' experiences of classroom interaction also are a significant feature of educational histories. The EDURURAL data provide no support for the idea that the achievement of individual students is affected by the gender composition of their classes. Though the statistical results are not robust, data from the 1985 survey do suggest that a student's achievement is affected by the gender of his or her teacher. Achievement also is affected by the socioeconomic status of students in the classroom, especially the proportion of families not farming.
For additional material on educational histories one must turn to the United States and Europe. Fine's (1991) study of dropping out in a public high school in New York City is a particularly valuable study of a much-discussed problem (see also Cicourel and Kitsue, 1963; Willis, 1977). As Fine observes, dropping out is a majority phenomenon. Something like 80 percent of ninth graders drop out before graduating. Fine (1991:14) argues that this cannot be accounted for by the poor quality of the students. "When students drop out of a high school in majority proportions, their exit must be read as a structural, if not self-conscious critique."
A key element of Fine's argument is the claim that educational achievement is not well correlated with ability.
[It is true that] students who read and compute at or above the ninth-grade level are more likely to be promoted than retained, and that students who read and compute below the ninth-grade level are more likely to be retained than promoted. However, it is also clear that almost half of those students who have been retained in ninth grade are reading and computing at or above grade level, (p. 240)
Retaining students in grade level in turn increases the chances that they will drop out. Conversely, enrollment in bilingual education programs results in a small increase in the probability of persisting to graduation, even though "students in bilingual education are allegedly most at risk" and were more likely to be retained in ninth grade (Fine, 1991:236).
The EDURURAL research and Fine's study of dropping out in effect look backward. They show that students can arrive at a given number of years of education completed by very diverse routes. Furstenberg et al.'s (1987) study of Adolescent Mothers in Later Life looks forward, cautioning that the apparent end of formal education may not mark a real completion.
Furstenberg et al. report on a follow-up study, 16 to 17 years after delivery, of an initial cohort of 403 adolescent mothers and their firstborn children, mostly black, who participated in a program for pregnant adolescents at Sinai Hospital in Baltimore, Maryland. The researchers expected that "premature" childbearing would have a negative effect on educational attainment. These expectations were confirmed by the first phase of the study. Most of the adolescent mothers in the Baltimore study left school "before or soon after their child was born." Five years later, slightly fewer than half had returned to school and finished high school. ". . . [O]nly 8% of the women who had not graduated were currently enrolled (along with 9% of the high school graduates)" (Furstenberg et al., 1987:25).
Reinterviewing these women 16 to 17 years after the birth of their first child, Furstenberg and his colleagues were surprised to find that many of them had continued their educations beyond this apparent end.
Of all the educational attainment that occurred following the birth of the first child, more than half took place in the second segment of the study. High school graduates, pursuing higher education, accounted for a significant proportion of the further schooling. But one woman in six completed high school in the second segment of the study or a third of the young mothers who had not graduated by 1972.... Clearly, many women returned to school to obtain a diploma, or more often a GED, when they were well into their twenties. This often occurred when their youngest child entered school. (Furstenberg et al., 1987:25-26)
Beyond the Classroom
More ambitiously, anthropologists (and others) can pursue the LeVines' lead, attempting to produce detailed ethnographic accounts of the culturally and historically specific connections between forms of education and literacy and subsequent practices involved in fertility. One study that allows us to imagine what such ethnographic accounts might look like is Brodie's (1994) Contraception and Abortion in 19th-Century America. McLaren (1978) and Gordon (1977) have written extensively about the lively public debates concerning population and family planning that have taken place in England and the United States. Nevertheless, Brodie's work is unusual in the degree to which it traces the ways in which ordinary persons consumed or made something of the representations with which the debates are filled, putting them to use in their own lives (see also Seccombe, 1992). Her rich material allows us to explore the links between literacy practices and fertility in the lives of ordinary men and women, well after they have left school.
In brief, Brodie argues that during the course of the nineteenth century, more American women used contraception. They also used new forms of contraception, shifting from coitus interruptus, abortion, and breastfeeding to postcoital douching, a variety of rhythm methods, condoms, pessaries, and sponges. The evidence for this comes from Clelia Duel Mosher's small turn-of-the-century survey of reproductive control and sexual practices among her women patients, an extensive body of personal letters, and a few private diaries. Brodie also traces some of the connections between changing contraceptive practices and a wide range of features of nineteenth-century American popular culture and economic organization. Among these are overlapping groups of participants in the debate over birth control, authors of self-help manuals, physicians, providers of watercure therapy, concepts of personal hygiene, lecture circuits, organizations devoted to free thought, and publishers of newspapers and other periodical literature. The developing postal system and network of railroads also were important.
As regards the relationship between education and fertility, Brodie's work demonstrates that it is not enough to say that new discoveries were circulated in printed mass media and passively absorbed by women and their partners. The
spread of new contraceptive practices cannot be accounted for simply by arguing that the members of an increasingly educated population gained access to already existing scientific knowledge because they could now read and write.
The changing nineteenth-century American "culture of contraception" (Gillis et al., 1992:5) depended on at least three kinds of reading and writing. There was an enormous genre of self-help or advice manuals. These included Robert Owen's Moral Physiology, a staple work in histories of the birth control debate; Charles Knowlton's Fruits of Philosophy; or, The Private Companion of Young Married People; Russell Thacher Trall's The Hydropathic Encyclopedia . . . with an Appendix on Conception; and A. M. Mauriceau's (pseudonym for Ann Trow Lohman, Charles Lohman, and/or Joseph Trow) The Married Woman's Private Medical Companion (Brodie, 1994:357-65). Many of the advice manuals were privately published. Their distribution depended on a network of commercial agents, the postal system, the railroads, halls of science and other organizations of freethinkers, organizations of medical sectarians, and so on. They also were offered for sale by their authors at public lectures. In some cases, written advice manuals could convey information in ways that oral communication could not.
...[The first three editions of [Knowlton's] Fruits of Philosophy were almost miniature books, only about three inches by two and half inches in size. It was designed for private perusal and passing on to friends, not for public display. (Brodie, 1994:105)
More speculatively, Brodie (1994:162) suggests that "it may have [been] easier for Knowlton to leave copies of his pamphlet with his patients than for him or them to circle the topic in conversation with embarrassed euphemisms."
Advertising also was critical. Advertisements were included in many advice manuals.
The water-cure publication the Herald of Health, with one of the largest circulations of any nineteenth-century medical journal, publicized douching syringes, promoted "voluntary motherhood," and argued the need for controlled reproduction.
Racy tabloids "aimed at an urban class of stable boys, maids, day laborers, upwardly mobile young clerks, and salespeople" carried advertisements for condoms and diaphragms. "In the 1830s and 1840s abortion drugs, condoms, cures for venereal disease, aphrodisiacs, and abortions were advertised in major urban newspapers in New England and the Middle Atlantic states." Opponents of birth control noted that advertisers could obtain the names of prospective customers from newspaper announcements of marriages. John Todd, a Congregational minister, complained:
No young lady in New England (and probably in the United States) can have a marriage announced in the papers without receiving in the mail within a week a printed circular offering information and instrumentalities and all needed facilities. (Brodie, 1994:190-93)
The whole circulation of contraceptive information depended, finally, on the personal letter. Until the passage of the Comstock Law in 1873, this was a mail order business.
The advertising pamphlet put out by the Beach Company in the 1860s, Habits of a Well-Organized Life, explained how to cope with the intricate process of a mail transaction. The circular gave detailed directions about how to order and pay for the products, how to contact the nearest express company office or post office to place an order through the mail, how to use a C.O.D. express, how to get a money order, how to endorse it so that no one would know to whom it was sent, how to involve a third person in the transaction. (Brodie, 1994:232)
A great many purveyors of advice conducted all or part of their business by mail. Following the publication of his popular manual The Marriage Guide; or, Natural History of Generation, Frederick Hollick was able to give up lecturing and ''devote himself to writing popular medical books and to his growing private practice, conducted almost entirely by correspondence. About fifty letters arrived daily at his post office box in Manhattan" (Brodie, 1994:117). In at least some social classes, people used personal letters to share contraceptive information with others in their personal networks of acquaintances.
In 1885, Rose Williams wrote from the Dakota Territory to answer a question put to her by an Ohio friend, Allettie Mosher: how to prevent pregnancy. "You want to know of a sure preventative [sic]. Well plague take it. The best way is for you to sleep in one bed and your Man in another & bet you will laugh and say 'You goose you think I am going to do that' no and I bet you would for I don't see any one that does. Well now the thing we [use] (when I say we I mean us girls) is a thing: but it hasn't always been sure as you know but that was our own carelessness for it is we have been sure. I do not know whether you can get them out there. They are called Pessairre [sic] or female preventative. They cost one dollar when Sis got hers it was before any of us went to Dak. She paid five dollars for it. The Directions are with it." (Brodie, 1994:212)
As Lave and Wenger ( 1991 ) suggest, it is communities of practice that learn. As a result of their use of culturally specific literacy practices, women were legitimate participants in these communities. Rather than being passed from the knowledgeable to the ignorant, information was generated out of the interactions among participants. In the 1852 edition of his popular The Hydropathic Encyclopedia, Russell Thacher Trall wrote:
[The fertile period generally] occurred from the 'commencement of menstruation through twelve days' after. By 1867, however, he added the warning that the mid-menstrual period, though generally safe, was not infallibly infertile. He therefore advised waiting ten to twelve days after menstruation. In 1867 he cited statistics collected from 'several hundred' patients to prove that for none of the women had an ovum passed into the uterus before the third day after menstruation or after the fourteenth. The satisfaction Trall's patients appear to have found with the sterile period probably did not come from his timetables as
much as from his careful instructions to women about how to discern their own ovulation. He recommended: 'By noticing the time for two or three successive periods at which the egg or clot passes off she will ascertain her menstrual habit.' (Brodie, 1994:84)
One imagines that the basis of Trall's instructions was the prior experience of women. Certainly the results of women's noticings were passed back to medical "professionals" of one stripe or another, sometimes in writing. Brodie ( 1994:161 ) reports that
Knowlton asked the female attendants at a birth for their opinions about treatments and about their own histories of conception and pregnancy. These discussions of women's earlier reproductive histories were sometimes carried on in writing—an important point, for the fact that a brash and direct doctor like Knowlton communicated in written notes with his female patients suggests a shift to greater reliance on the written word.
Kasarda et al. (1986) are correct. Research on classroom language and its very circumscribed acquisition suggests the assumption that the experience of formal education in one country is broadly equivalent to the experience of formal education in another is untenable. It seems likely that education always is socially situated. There is no satisfactory culture-free definition of education as an autonomous process of internalization.
Where are we, then, if our notions of the universal cognitive consequences of education are ill-founded while the statistical support for the idea that education is a powerful indirect determinant of mortality and fertility change remains? In her 1997 presidential address to the Population Association of America, Mason (1997:446,449) observes that explanations of fertility transition have been "set . . .up for failure" by "the tendency to assume that there is only one 'master' cause of all fertility transitions." To the contrary, she argues:
Fertility transitions occur under a variety of institutional, cultural, and environmental conditions; they occur when combinations of conditions are sufficient to motivate or enable a substantial portion of the population to adopt birth prevention measures on a parity-specific basis.
Similarly, it seems likely that education influences fertility (and mortality) through a variety of combinations of links, not just one.
The consequences of education, including its links with fertility, may be expected to vary from one sort of student to another within national populations, as well as from one national population to another. Most research on the consequences of education appears to assume that those consequences are consistent and additive. A person who leaves school after three completed years gets three doses of something, a person who leaves school after seven years gets seven
doses, and a person who leaves school as a college graduate after seventeen years gets seventeen doses, but the something is always the same. Not surprisingly, however, the classroom experiences of children in particular grades appear to be conditioned by aspects of their ascribed identities and their experiences prior to entering that grade. In the Carolina Piedmont, Trackton and Roadville children experience teachers who feel they, the children, are rude. Harbison and Hanushek's (1992) data from Northeast Brazil suggest that teachers respond differently to male and female students and that classes composed largely of children from farming families operate differently from classes composed largely of children from nonfarming families. Rather more sharply, Fine's New York research suggests that students from poor urban neighborhoods, largely racial and ethnic minorities, also receive differential treatment and leave school in disproportionate numbers, in part precisely because they reject that treatment. This and a great deal more evidence concerning what Fine (1991:200) calls "the vastly differential educations and outcomes made available to students by virtue of their social class, race/ethnicity, gender, sexual orientation, disability, and geography" suggests that schooling has different kinds of consequences for different kinds of people.
At a minimum, in large-scale quantitative studies it would be helpful to qualify measures of educational attainment such as years of education completed with some sense of the variety of individual educational histories and of the ways these weave through reproductive histories. Relatively simple measures of this variety might include years behind standard grade for individual students or the distribution by age of students in various grades. Because education is "a process with components rather than a cumulative trait" that interacts with fertility over the life course in complex ways, it may be necessary "to obtain educational histories with the same rigor as we obtain pregnancy histories" (Kasarda et al., 1986:106). This would follow the lead, especially, of Furstenberg et al. (1987).
At a maximum, it would be helpful to carry out much more intensive studies of single societies or groups of connected societies. This would follow the paths opened by the LeVines (1991) and Brodie (1994). Research along these lines would bring developments in the anthropology of education and literacy to bear on the recent upsurge of demographic interest in social interaction and "diffusion." Where they are produced by schooling, literacy and other educated practices are likely to be an important element in at least some of the "networks and other structures of social relationships" in which individuals "are embedded'' and through which they exercise mutual influence (Montgomery and Casterline, 1996:152; see also Bongaarts and Watkins, 1997). Because they are unlikely to be the simple reflex of school practices and years of exposure to them, the educated practices of adults must be studied independently, in all their cultural specificity and historical contingency. Searching for multiple links between education and fertility may require new research strategies.
Barton, D., and R. Ivanic 1991 Writing in the Community. Newbury Park: Sage Publications.
Bongaarts, J., and S. Watkins 1997 Social interactions and contemporary fertility transitions. Population and Development Review 22(4):639-682.
Bornstein, M.H., ed. 1989 Maternal Responsiveness: Characteristics and Consequences. San Francisco, Calif.: Jossey-Bass.
Brodie, J.F. 1994 Contraception and Abortion in 19th-Century America. Ithaca, N.Y.: Cornell University Press.
Bulatao, R., and R. Lee, eds. 1983 Determinants of Fertility in Developing Countries. New York: Academic Press.
Caldwell, J. 1980 Mass education as a determinant of the timing of fertility decline. Population and Development Review 6(2):225-255.
Carter, A.T. 1998 Cultural models and reproductive behavior. In A. Basu and P. Aaby, eds., The Methods and Uses of Anthropological Demography. Oxford: Oxford University Press.
Cicourel, A.V., and J.I. Kitsue 1963 The Educational Decision-Makers. Indianapolis, Ind.: Bobbs-Merrill.
Cleland, J., and G. Kaufmann 1998 Education, fertility and child survival: Unraveling the links. In A. Basu and P. Aaby, eds., The Methods and Uses of Anthropological Demography. Oxford: Oxford University Press.
Cochrane, S.H. 1979 Fertility and Education: What Do We Really Know? Baltimore, Md.: The Johns Hopkins University Press.
Cook-Gumperz, J. 1977 Situated instructions: Language socialization of school-age children. In S. Ervin-Tripp and C. Mitchell-Kernana, eds., Child Discourse. New York: Academic Press.
DeCastell, S., and T. Walker 1991 Identity, metamorphosis, and ethnographic research: What kind of story is ways with words? Anthropology and Education Quarterly 22:3-20.
Easterlin, R.A. 1978 The economics and sociology of fertility: A synthesis. In C. Tilly, ed., Historical Studies of Changing Fertility. Princeton, N.J.: Princeton University Press.
1983 Modernization and fertility: A critical essay. Pp. 562-586 in R. Bulatao and R. Lee, eds., Determinants of Fertility in Developing Countries. Volume 2. New York: Academic Press.
Erickson, F.D. 1987 Transformation and school success: The politics and culture of educational achievement. Anthropology and Education Quarterly 18:335-355.
Fine, M. 1991 Framing Dropouts: Notes on the Politics of an Urban Public High School. Albany, N.Y.: State University of New York Press.
Furet, F., and J. Ozouf 1982 Reading and Writing: Literacy in France from Calvin to Jules Ferry. Cambridge: Cambridge University Press.
Furstenberg, F.F., Jr., J. Brooks-Gunn, and S.P. Morgan 1987 Adolescent Mothers in Later Life. Cambridge: Cambridge University Press.
Gillis, J.R., L.A. Tilly, and D. Levine 1992 The quiet revolution. In J.R. Gillis, L.A. Tilly, and D. Levine, eds., The European Experience of Declining Fertility. Cambridge, Mass.: Blackwell Publishers.
Goody, J., and L. Watt 1962 The consequences of literacy. Comparative Studies in Society and History 5:304-345.
Gordon, L. 1977 Woman's Body, Woman's Right: Birth Control in America. New York: Penguin Books.
Greenfield, P.M. 1972 Oral or written language: The consequences for cognitive development in Africa, U.S. and England. Language and Speech 15(2):169-178.
Gumperz, J.J. 1986 Interactional sociolinguistics in the study of schooling. In J. Cook-Gumperz, ed., The Social Construction of Literacy. Cambridge: Cambridge University Press.
Hanks, W.F. 1996 Language and Communicative Practices. Boulder, Colo.: Westview Press.
Harbison, R.W., and E.A. Hanushek 1992 Educational Performance of the Poor: Lessons from Rural Northeast Brazil. New York: Oxford University Press.
Heath, S.B. 1983 Ways with Words: Language, Life, and Work in Communities and Classrooms. Cambridge: Cambridge University Press.
1993 The madness(es) of reading and writing ethnography. Anthropology and Education Quarterly 24:256-268.
Hildyard, A., and D. Olson 1978 Literacy and the Specialisation of Language. Toronto: Ontario Institute for Studies in Education (mimeographed).
Kasarda, J.D., J. Billy, and K. West 1986 Status Enhancement and Fertility: Reproductive Responses to Social Mobility and Educational Opportunity. Orlando, Fla.: Academic Press.
Lave, J. 1988 Cognition in Practice. Cambridge: Cambridge University Press.
1989 The acquisition of culture and the practice of understanding. In J. Stigler, R. Shweder, and G. Herdt, eds., The Chicago Symposia on Human Development. Cambridge: Cambridge University Press.
1991 Situated learning in communities of practice. In L.B. Resnick, J.M. Levine. and S.D. Teasley, eds., Perspectives on Socially Shared Cognition. Washington, D.C.: American Psychological Association.
Lave, J., and E. Wenger 1991 Situated Learning: Legitimate Peripheral Participation. Cambridge: Cambridge University Press.
LeVine, R.A. 1983 Fertility and child development: An anthropological view. In D. Wagner, ed., Child Development and International Development: Research-Policy Interfaces. San Francisco, Calif.: Jossey-Bass.
1987 Women's schooling, patterns of fertility and child survival. Educational Researcher 16:21-27.
LeVine, R.A., S.E. LeVine, A. Richman, F.M. Tapia Uribe, C. Sunderland Correa, and P.M. Miller 1991 Women's schooling and child care in the demographic transition: A Mexican case study. Population and Development Review 17(3):459-496.
LeVine, R.A., and S.C.M. Scrimshaw 1983 Effects of culture on fertility: Anthropological contributions. Pp. 666-695 in R. Bulatao and R. Lee, eds., Determinants of Fertility in Developing Countries. Volume 2. New York: Academic Press.
LeVine, R.A., and M.I. White 1986 Human Conditions: The Cognitive Basis of Educational Development. New York: Routledge and Kegan Paul.
Mason, K.O. 1997 Explaining fertility transitions. Demography 34(4):443-454.
McLaren, A. 1978 Birth Control in Nineteenth-Century England. London: Croom Helm.
Michaels, S. 1981 'Sharing time': Children's narrative styles and differential access to literacy. Language in Society 10(3):423-442.
Michaels, S., and J. Cook-Gumperz 1979 A study of sharing time with first-grade students: Discourse narratives in the classroom. Proceedings of the Fifth Annual Meetings of the Berkeley Linguistics Society. Berkeley, Calif.: University of California.
Minick, N. 1993 Teacher's directives: The social construction of 'literal meanings' and 'real worlds' in classroom discourse. In S. Chaiklin and J. Lave, eds., Understanding Practice: Perspectives on Activity and Context. Cambridge: Cambridge University Press.
Montgomery, M., and J. Casterline 1996 Social learning, social influence, and new models of fertility. Population and Development Review 22(Suppl.): 151-175.
Pelissier, C. 1991 The anthropology of teaching and learning. Annual Review of Anthropology 20:75-95.
Rogoff, B. 1990 Apprenticeship in Thinking: Cognitive Development in Social Context. New York: Oxford University Press.
Rosen, H. 1985 The voices of communities and language in classrooms. Harvard Educational Review 55:448-456.
Schieffelin, B.B., and E. Ochs 1986 Language socialization. Annual Review of Anthropology 15:163-191.
Scollon, R., and S. Scollon 1979 Linguistic Convergence: An Ethnography of Speaking at Fort Chipawyan. Alberta. New York: Academic Press.
1981 Narrative, Literacy, and Face in Interethnic Communication. Norwood, N.J.: Ablex.
Seccombe, W. 1992 Men's 'marital rights' and women's 'wifely duties': Changing conjugal relations in the fertility decline. In J.R. Gillis, L.A. Tilly, and D. Levine, eds., The European Experience of Declining Fertility. Cambridge, Mass.: Blackwell Publishers.
Singh, S., and J. Casterline 1985 The socioeconomic determinants of fertility. Pp. 199-222 in J. Cleland and J. Hobcraft. eds. Reproductive Change in Developing Countries. Oxford: Oxford University Press.
Snow, C.E. 1977 Mother's speech research: From input to interaction. In C.E. Snow and C.A. Ferguson. eds., Talking to Children. Cambridge: Cambridge University Press.
1990 The development of definitional skill. Journal of Child Language 17:697-710.
Street, B.V. 1984 Literacy in Theory and Practice. Cambridge: Cambridge University Press.
Street, B.V., ed. 1990 Cross-cultural Approaches to Literacy. Cambridge: Cambridge University Press.
Tilly, C., ed. 1978 Historical Studies of Changing Fertility. Princeton, N.J.: Princeton University Press.
Vincent, D. 1989 Literacy and Popular Culture: England 1750-1914. Cambridge: Cambridge University Press.
Willis, P.E. 1977 Learning to Labour. Westmead: Saxon House.