Keywords

Introduction

The past 15 years have heralded increased attention among some researchers of language and education to the temporal aspects of their questions of interest. A subset of linguistic anthropologists of education, ethnographers of literacy, and investigators of second language development have begun to investigate how time figures in social and academic identification and language learning. How does a single class session or a brief conversational exchange contribute to longer-term linguistic, academic, or identity development? How do phenomena developed over decades or centuries influence short-term exchanges in the present? How can attention to these temporal relationships better illuminate questions of learning? The literature reviewed below speaks to these and related questions.

This chapter provides an overview of language and education research that incorporates explicit attention to timescales into its theoretical and methodological approach. A timescale is “the characteristic spatiotemporal envelope within which a process happens” (Wortham 2012, p. 133). Attention to timescales within the study of education can be traced back principally to Jay Lemke’s discussions of ecosocial systems (2000, 2001, 2002). Because of the centrality of his work, I begin by outlining the most widely applied components of Lemke’s theory of timescales for educational research. In the subsequent section, I discuss major insights from linguistic anthropological studies of academic and social identification and provide select details on a few major studies that elucidate these points. A brief discussion of how timescales figure in recent studies of second language development follows. The chapter ends with a discussion of the challenges to researching timescales and some promising approaches to refining empirical approaches.

Early Developments: Theorizing Timescales in Educational Research

Lemke (2000, 2001) draws attention to the fact that there are various temporal scales during and across which all processes occur, from chemical processes that take milliseconds to cosmological processes that occur over billions of years. Educational processes occupy a narrower, though still quite wide, range of timescales: from the seconds occupied by an individual student or teacher’s utterance, the months it takes to complete a semester, the many decades that constitute an individual’s lifetime educational development, to the hundreds of years it can take for educational institutions to change, and so forth. Lemke proposes that to understand any kind of human activity, including those related to education, researchers must identify relevant processes at multiple timescales and investigate how those processes interact with others at different scales.

Lemke proposes a few important ways in which processes at various scales relate to one another. The first relates to emergence and constraint. He explains that what can happen in a process at any given timescale is made possible by the processes going on at the shorter, faster timescale immediately below it and, at the same time, is constrained by those processes at the longer, slower scale above it. Take, for example, a given class session in one teacher’s classroom. What is likely to happen in that one, say 40 min, lesson is influenced by patterns of interaction that have developed over longer timescales – over that week, month, and school year, by ways of organizing classroom instruction at that particular school over its decades-long history, by cultural patterns of schooling that have developed over hundreds of years. Each of these scales constrains the one below it, shaping what is likely to happen at the lower, shorter scale. At the same time, what happens in that one class session is also made possible by practices at shorter timescales than the focal 40-min class period, by minutes-long interactions and seconds-long utterances that may either contribute to longer-standing patterns or introduce unexpected ones. Likewise, new patterns that emerge in that class session make possible changes in the way that classroom will operate during that particular week, and so forth, up the scalar hierarchy.

The second way in which various scales relate is when activities at distant timescales produce effects on one another, when a short-timescale process has a rippling effect into longer timescales or vice versa. Lemke (2000) claims that heterochrony, “in which a long timescale phenomenon produces an effect in a much shorter timescale activity” is “the basis for human social organization across timescales” (p. 280). Heterochrony is most obvious in “semiotic objects” which “carr[y] significant information across time and space” in that they recall meanings developed on a different timescale, while they are also used in the present moment (Lemke 2001, p. 21). The notebook a student refers to when her teacher asks a question to the class provides an example, since the notes were written days or weeks before but are being read and interpreted in response to the teacher’s question in that moment (Lemke 2000, p. 281). Human beings themselves also function in this way, as Lemke (2002) explains, “[A 12 year-old child] may be 12 by the calibrations of calendars, but as a member of the community he is dynamically heterochronous, some mix of every age he’s already been, and every age he’s learned to cope with, and many ages he’s begun to understand and imagine and model” (p. 81). To consider any phenomenon at only one timescale, therefore, is to miss a good deal of what is relevant to it. Because human activity, including educational processes, are so complex, an adequate account of them must attend to multiple timescales and to the interrelationships between processes at different scales, including emergence, constraint, and heterochrony.

Proposed Topics and Methods

While Lemke claims that attention to timescales would benefit the study of any social process, he notes the particular potential of this framework in understanding identity formation, classroom learning (2000, 2001), and language development (2002). While we know a good deal more about short-term social processes, Lemke argues that we know much less about longer-scale processes that occur over multiple days, months, years, or more (2000, p. 287). Access to these longer-timescale activities is key because human processes such as identity formation, changes in habits of reasoning, and language development all occur over longer timescales (Lemke 2000, 2002).

Lemke proposes that gaining access to these longer timescales would require borrowing methods from historians, archivists, and biographers, who regularly engage with timescales of decades or longer (2000). This suggestion would entail the wider use of oral histories, life history interviews, archival research, and deep engagement with secondary historical source documents. In addition, in order to understand societal processes and changes, teams of researchers would be necessary, as well as a “self-sustaining institution that would last long enough to observe major historical change in the system” (2000, p. 288), such that multiple generations of researchers could observe and analyze longer-timescale processes. Lemke suggests that modern technologies and networks may hold some promise for the kinds of collaboration and long-term corpus building that he envisions here.

Major Contributions: Timescales and Social Identification

Much of the scholarship that employs the concept of timescales, as Lemke outlined it, focuses on social identification, or “the process through which individuals and groups become identified as publically recognizable categories of people” (Wortham 2004, p. 716). For example, studies of social identification may examine how a child becomes identified as a “nerd,” a “jock,” a “good reader,” or any multitude of widely circulating and/or locally relevant identities. This literature focuses on how social identification happens over time in classrooms, through the language that others use to talk about and to particular individuals or groups. In many cases, researchers in this area are also interested in the relationship between social identification and academic learning (Bell et al. 2012; Polman and Miller 2010; Wortham 2006, 2008) and, in particular, literacy (Bartlett 2007; Burgess and Ivanic 2010; Compton-Lilly 2011; Roswell and Pahl 2007). In this section, I highlight key findings that have come out of this work, followed by empirical examples that illustrate these major insights.

By attending to timescales in the study of social and academic identification, researchers have come to three principal conclusions and related methodological implications. First, many of these studies illustrate empirically that social identification occurs at multiple timescales, from speech events to prolonged social and academic interaction over months or years. Therefore, it must be studied at these various scales, with attention to the ways in which shorter-timescale events can develop into more durable patterns at longer scales (Wortham 2006, 2008). Second, and relatedly, this scholarship has shown that social identification, including students’ academic and literate identifications, can change (sometimes drastically) over time and that it is therefore crucial to study such identification over the long term, rather than relying exclusively on shorter-term analyses (Bartlett 2007; Compton-Lilly 2011; Wortham and Rhodes 2012). Third, this body of work has demonstrated how, in making sense of themselves and others, people draw on identity categories, discourses, and other resources from multiple timescales, both within and far beyond the scale of their own lifetimes (Compton-Lilly 2011; Hall et al. 2010; Pahl 2007; Wortham 2006). These multiple timescales are simultaneously invoked in talk and texts (Bell et al. 2012; Brown et al. 2005; Wortham 2006). Together, these studies show that attention to multiple timescales can lead to more accurate and nuanced analyses than attention to only one scale (Wortham and Rhodes 2012; Zhang and Sun 2011) and, in particular, can provide a more precise alternative to “micro” and “macro” explanations of emergence and constraint common in social theory (Hult 2010; Wortham 2006, 2012; Wortham and Rhodes 2012). The following examples illustrate these key findings.

Pioneering work by Wortham (e.g., 2006, 2008) illustrates how social identification occurs on multiple timescales, in single interactions as well through a trajectory of interactions over time that “sediment” social identification, or result in the formation of longer-lasting identities. Working with video data from a middle-school science classroom (2008), for example, he analyzed 2 months of recorded interaction to show how one student, Philip, becomes understood as both a good student and a low-status peer. Wortham illustrates how one particular speech event, in which Philip grabs a female lab partner’s wrist, marks a pivotal moment in Philip’s identification as an academic leader as well as a social undesirable. Through his words and action, Philip asserts his academic authority by taking control of the experiment. At the same time, however, his lab mate’s strongly negative reaction to his touch positions Phillip as an undesirable peer. Through a close analysis of short segments of recorded interaction between Philip and his lab partners, Wortham goes on to show how subsequent events over the following months make these particular identifications of Philip enduring, even when other isolated interactions offer divergent possibilities.

In his study of a high school Paideia classroom (2006), Wortham combined discourse analysis of 50 recorded class sessions over an academic year with ethnographic observations and interviews. His analyses revealed that students and teachers drew upon identity categories, intellectual positions, and cognitive models from various timescales in their classroom discussions. For example, in minutes-long classroom interactions, teachers and students invoked sociohistorical categories, developed over decades and centuries of “unpromising boys” and “promising girls,” identity models from the curriculum (such as Aristotle’s “beast,” conceived of over 2000 years before the study took place), and local versions of these models that developed over weeks and months in the classroom. Wortham illustrates how two students, Maurice and Tyisha, become identified as outsiders from the group over time, in part through the heterochronous “participant examples” teachers and students used to discuss curricular themes. His account of this interdependent process of social identification and academic learning illustrates the relevance of a wide range of timescales in social identification as well as the ways in which short, “microgenetic” timescale events (such as individual classroom discussions) contribute to longer-timescale social identification.

Following Wortham, Bartlett (2007) describes a case in which a student’s social identification can change dramatically over time and space, thus underscoring the importance of analyzing social identification over longer intervals. Bartlett shows how over the course of 4 years, the social identification of one bilingual high school student, María, shifted drastically. Maria went from being described as a stigmatized SIFE (student with interrupted formal education) to being understood by her teachers as a hard-working, successful student. As she made choices that influenced the way her teachers viewed her, Maria drew upon a local version of school success that reflected the more widely circulating, longer-timescale, sociohistorical narrative of immigrant opportunity in the USA. As María’s identification shifted, so too did the possibilities she had for learning English and academic content, as she was able to move through the beginner classes into those that afforded richer opportunities for language and content learning. During the 4-year study, Bartlett’s research team collected field notes from multiple classrooms and “recorded and reported speech” via audio-recorded interviews and participant observation. Bartlett notes that her decision not to audio record classroom interaction meant that she was unable to conduct discourse analyses like Wortham (2006); however, her team’s longitudinal, multiple-classroom approach allowed her to see how this particular student’s identity changed across both time and classroom contexts.

Compton-Lilly (2011), in an even longer-scale project than Bartlett’s, tracked the discourses around literacy and schooling that several children and their family members drew upon, adopted, and adapted over an 8-year period. Her data include parent and student interviews, field notes, classroom assessments, and students’ writing samples. She employed a method of “periodic restudy” in which she collected data at regular intervals (in 1st, 4th/5th, and 7th/8th grades). Compton-Lilly examines the discourses employed by one student, Alicia, and her mother and siblings during and after recorded interviews, with a particular focus on how multiple timescales are invoked in their talk about literacy and schooling. She shows that family members accessed discourses from what she calls “historical,” “family,” and “ongoing” timescales to make sense of their experiences, as when Alicia’s mother invoked historical events, childhood memories, and ongoing teacher comments in describing why she advocated for her children’s education. In this way, Compton-Lilly emphasizes the interrelations between events and processes at various timescales (see also Hall et al. 2010; Wortham 2006). Pahl (2007) and Roswell and Pahl (2007) demonstrate similar findings about layered timescales, though through a focus on written texts as semiotic objects, which they argue can be analyzed to reveal the identities of the text-makers at one moment in time (Roswell and Pahl 2007).

Through their attention to time as part of their theoretical and methodological approaches, these studies bring us closer to understanding the mechanisms through which students become identified and come to inhabit social and academic identities. Unsurprisingly, the methods used to arrive at these central findings were ethnographic, discourse analytic, and, in most cases, longitudinal. As Pahl (2007) notes, ethnographers are uniquely positioned to attend to various timescales because of their methodologies and the length of their studies. Ethnographic methods typically afford access to some recorded speech (e.g., classroom discourse recordings, interviews), produced at short timescales (minutes to hours), as well as to developments over longer timescales (months to years). Furthermore, ethnographers develop deep knowledge of their contexts that can attune them to relevant timescale processes beyond the span of their actual studies. Researchers attuned to timescales in understanding educational phenomena also look for dynamic interactions across data produced at disparate scales. Such analyses go beyond reliance on the broad categories of “macro” and “micro” to explain their phenomena of study and instead identify more specific ranges during which ideas or discourses circulate, including intermediate-level processes (such as the classroom category of “beast” that developed over weeks and months in Wortham’s Paideia classroom or the “family”-level timescale in Compton-Lilly’s work) that are often left out of micro/macro accounts (Hult 2010; Wortham 2006, 2012; Wortham and Rhodes 2012). These studies also highlight the ways in which emergence and constraint happen at many different scales, rather than only at the individual or the societal levels (see Wortham 2012; Wortham and Rhodes 2012 for extensive discussion of this issue).

Attending to timescales in the investigation of educational processes, therefore, calls for the integrated use of such methods as ethnography and discourse analysis as well as for more protracted studies, since all of these methods provide access to different timescale processes. In addition to systematic data collection at multiple scales, it requires analysis of data for relationships across disparate scales and for influences of longer-timescale processes to which the analyst does not have direct access. Doing so entails a step away from the all-too-common siloing of the historical from the ethnographic and the discourse analytic in educational research. Instead, researching timescales in education requires researchers to attend carefully to change over time and to dynamic interactions across scales that help to explain their questions of interest.

Work in Progress: Timescales and Language Development

In addition to the attention timescales have received in studies of social identification, another noteworthy strand of research uses timescales as part of a theoretical framework of second language development and use (Kramsch 2008; Larsen-Freeman and Cameron 2008; Steffensen and Fill 2014). A call to research multiple timescales appears in several newer approaches to studying language and language development, including dynamic systems (e.g., vanGeert 2008), complexity theory (e.g., Larsen-Freeman and Cameron 2008), and ecolinguistics (e.g., Kramsch 2008; Steffensen and Fill 2014). All of these approaches view language and language development as complex, dynamic, and emergent systems.

According to complexity theory and dynamic systems theory , such systems are composed of many different component parts, change over time, are nonlinear in their development, and are emergent, in that components on lower levels of the system interact to “produce patterned behavior at various levels over time” (Churchill 2007, p. 342). Theorists approaching language and language development from this perspective eschew reductionism in their data collection and analysis. Rather than isolating variables and researching their effects, they instead seek to understand “how the interaction of the parts gives rise to new patterns of behavior” (Larsen- Freeman and Cameron 2008, p. 201). Timescales are relevant to these approaches because the nested subsystems within any complex system operate at varying timescales. Steffensen and Fill (2014) explain:

For instance, when we engage in verbal activity, we integrate the fast timescales of synaptic activity and interbodily dynamics (bodily and vocal gestures) with the longer, slower timescales of sociocultural dynamics (e.g. the logics of the arena in which we engage in social interactions) and the historical resources of social and symbolic patterns and norms (Thibault 2011; Steffensen 2011). Thus, saying that language is caused by brains, or microsocial norms, or human interactions, or phenomenological experiences, is an unwarranted reduction of a multifaceted reality. (pp. 14–15)

From this perspective, it is necessary to consider as many aspects of the system as possible, at the multiple timescales at which these component processes occur, in order to understand a complex process like language development. However, researchers using these approaches also recognize the “unknowableness” of complex systems, explaining that it is impossible to fully describe the changing, emergent, and nonlinear behavior and interactions between component processes within the system (Larsen-Freeman and Cameron 2008).

Drawing on complexity theory, emergentist theories of language acquisition, and postmodern theories of language use, Kramsch (2008) outlines the components of an ecological theory of language development and use. She explains that timescales are relevant to that approach because “the meanings expressed through language occur on multiple timescales” (p. 391) and “utterances…recreate environments from other scales of space and time, produce fractals of patterns from one timescale to another” (p. 401). That is, language is both uttered in the present moment and contains multiple layers of meaning, developed at different temporal (and spatial) scales, only some of which are invoked by the participants, though others are also present (Bakhtin 1981; Blackledge and Creese 2014; Blommaert 2005, as cited in Kramsch 2008, p. 393). Language can also reproduce patterns that recur at multiple scales (see also vanGeert 2008). Kramsch (2008) provides a fascinating empirical example in her analysis of spoken interaction between a Yucatecan immigrant, his literacy teacher, and two Asian merchants. Using a discourse-analytic approach, she points to moments in the conversation where participations invoke multiple timescales and shows how examining connected timescales in language use allows the analyst to capture aspects of linguistic interaction that would be otherwise invisible.

In addition to discourse-analytic and ethnographic approaches, other methodologies have been proposed to understand processes occurring at multiple timescales in language development. Larsen-Freeman and Cameron (2008) describe several newer methodologies that are especially useful for researching complex systems, two of which specifically target temporal aspects in the study of complexity. The first is the longitudinal, case-study, time series approach, which they state, “enables connections to be made across levels and timescales” (p. 208). In this methodology, it is necessary to identify both the timescale and the rate at which some particular change of interest occurs in order to determine how long a study to conduct as well as how often to sample. At the other end of the spectrum, microdevelopment can illuminate complex processes at shorter timescales. Studies using a microdevelopment approach produce dense corpora of data collected over short periods of time, while study participants are engaged in some activity (Granott and Parziale 2002). For example, Gelman and colleagues (2002) analyzed the real-time development of scientific knowledge and English as a second language (ESL) in a high school “science into ESL” program. By coding and conducting quantitative analyses of students’ science journals, the authors were able to see evidence of learning in process. A microdevelopment approach promises the possibility to directly observe moments of change that can illuminate longer-timescale development and can provide insights into “the how of [language] development” (Thelen and Corbetta 2002, as cited in Larsen-Freeman and Cameron 2008, p. 208).

Broadly, incorporating attention to timescales in research on second language development calls for an expansion of focus, away from studying individual sounds or structures, to investigating the ways in which any particular aspect of language develops within the broader system, not only of the language itself, but also of the environment in which that development happens. The timescales at which these component processes occur are but one aspect to attend to within these complex systems. More specifically, adopting a timescales lens in the study of language development might also entail identifying the timescales under which particular linguistic structures of interest develop, designing longitudinal studies that document change over time, collecting data on the same phenomena at multiple scales, or attending to relevant scales that are not available for direct data collection, but which nevertheless shed light on language use and are accessible to the analyst via discourse-analytic and ethnographic analyses (Kramsch 2008).

Problems and Difficulties

As reviewed above, researching timescales in language and education offers unique insights into complex processes such as social identification and language development. However, attending to multiple timescales also poses real methodological challenges. Lemke (2000) noted that the kinds of fine-grained analyses that are possible at the interactional or activity level become impossible as the scale lengthens, because of the logistical difficulties of collecting and analyzing that quantity of data over extended periods of time. Furthermore, as the temporal scale expands, so usually does the spatial scale, producing an untenable range of research sites. Longer-timescale work is also difficult because of the limits imposed by the length of the typical research project and the career of a single analyst.

In addition, if there are an infinite number of timescales (Lemke 2000), and complex processes and systems are ultimately “unknowable” (Larsen-Freeman and Cameron 2008), then clearly no analyst can hope to attend to all possible timescales acting on a focal phenomenon. Furthermore, the theory that only certain scales are “within the grasp of” interactional participants even when other “layers of historicity” are also present (Bakhtin, 1981; Blommaert 2005, as cited in Kramsch 2008, p. 392) presents another question. Should researchers focus primarily on those scales that have been directly invoked by participants or attempt to also elucidate those that are invisible to participants? In either case, a methodology for determining which scales are relevant is necessary (Wortham 2012; Wortham and Rhodes 2012). Researchers have always had to make choices about how to focus their research and about what balance to strike between emic and etic perspectives. However, as the range of possible foci expands, as it must do when attending to multiple timescales, a proposal for guiding these analytic choices becomes increasingly important.

Future Directions

Partial solutions to these difficulties have been proposed. With regard to the challenges of collecting data at long timescales, the method of periodic restudy employed by Compton-Lilly (2011) and the longitudinal, case-study, time series approach, for example, both offer systematic approaches to collecting manageable amounts of data over longer timescales. Incorporating such methodologies over timescales longer than a decade, however, would very likely require a multi-researcher, multi-sited, corpus-building approach.

In cases where the issue is not data collection but a question of what timescales are relevant to analysis, recent attention to spatial and temporal scales in sociolinguistics and anthropology offer promising directions. Blommaert (2010) proposes attending to moments of “scale jumping ” in discourse, when norms, expectations, ideas, and other phenomena from higher-level temporal and spatial scales are referred to by participants. He draws attention to the fact that these moves are often performed as power tactics, since some actors have more access to higher-level scales than others and provides the example of the doctor who scale jumps when he switches from vernacular to medical jargon in discussion with a patient, a jump which many patients would be unable to make (p. 36). Collins (2012) similarly argues that scales do not have a priori relevance but instead are produced in real-time interaction by participants. Like Blommaert, he notes that “scaling” can be rife with power dynamics, as when certain scales carry greater authority than others. In one example from a school context, he shows how a tutor’s use of Spanish with an immigrant child “licenses a bilingual local scale” by legitimating Spanish as appropriate in this learning context. However, when her teacher later declares that she doesn’t want the tutor speaking to this child in Spanish, she references a wider-scale English-only prohibition that has greater authority at school (p. 204). Lempert (2012) goes further, in stressing a need to account for how scale jumping happens. He exhorts analysts to investigate, rather than presume, the scale of any given interaction, not only those scales that are invoked but even the scale of the focal phenomenon. He states, “despite the inexhaustible scalar complexity that can (nb. a potential) be discovered and exploited, there is no guarantee that scale matters – in the sense of being registered by interactants in some respect or capacity and pragmatically consequential in discursive interaction” (p. 153, emphasis in original). Future work in this area should continue to articulate an empirical approach to determining such registering and consequence, specifically articulating how analysts can identify timescales that are “within the grasp” of the participants, as well as those that are not, but which may still be important in understanding the focal phenomenon.

While the authors reviewed in this chapter have taken up Lemke’s (2000, 2001, 2002) call to attend to multiple timescales in researching complex educational and social processes, his methodological recommendations have not been widely applied. In particular, his exhortation to adopt methodologies from historians and biographers, such as archival research, oral histories, and others, to conduct research in teams comprised of analysts with a wide range of perspectives and positionalities and to develop longitudinal corpora of interactional data that can be analyzed over the long term by multiple researchers (2000) all still hold much promise for researchers interested in understanding the temporal aspects of complex systems, including the many of interest to researchers of language and education.

Cross-References

Related Articles in the Encyclopedia of Language and Education

Sune Vork Steffensen and Claire Kramsch: The Ecology of Second Language Acquisition and Socialization. In Volume: Language Socialization