This post is a consolidated form of the 8-part series I posted from April 16-May 30, 2013 as part of a class seminar on literacy. I have done a rough edit to attempt to make the posts more cohesive; if some issues are unclear, please refer to the original posts. Thanks again to all who commented and encouraged me during this endeavor. A special thanks to Dr. Carol Hopkins, who supported me in this project.
The Reading Wars of the 1990s were documented by the exchanges between Edelsky and McKenna, Robinson, & Miller in Educational Researcher. This theoretical conflict between Whole Language and Traditional Literacy represents what I call a “traditional academic feud.” A traditional academic feud is a social phenomenon in an academic field of study where two or more groups theorize about or describe a single phenomenon in categorically different ways. That a traditional academic feud eventually achieved “war” status is a curious phenomenon–but perhaps no more curious than the fact that academic feuds occur in the first place. In this series, I plan to explain 1) why the academic feud of the Reading Wars occurred, 2) why the feud escalated to a war, and 3) why a similar feud developed recently in the reading comprehension field (this element of the thesis was abandoned due to time/space constraints), and (if I am brave enough) 4) how we might avoid wars and have more productive feuds in Literacy Studies.
The traditional academic feud often occurs in what Bernstein called “horizontal knowledge structures” (as cited in Maton & Muller, 2007). A knowledge structure is a description of the way an academic field organizes its theories and the rules that govern the production of such theories. Bernstein’s horizontal knowledge structure is a “series of specialized languages with specialized modes of interrogation and criteria for the construction and circulation of texts” (as cited in Maton & Muller, 2007). These “specialized languages” are ways of writing and talking about a phenomenon. They are used to communicate the propositions of various theories–of which there are many. In horizontal knowledge structures, multiple theories compete for adherents and dominance in the field–often competing to explain a similar phenomenon, but talking about that phenomenon in different specialized languages. Following the analysis from Christie & Macken-Horarik (2007) that Literacy Studies is a horizontal knowledge structure, I argue that the Reading Wars are characteristic of academic feuds in horizontal knowledge structures, namely because they are an instance of two specialized languages explaining the same phenomenon and competing for dominance in the field.
Theoretical conflict itself is not symptomatic of horizontal knowledge structures alone. Their counterpart, vertical knowledge structures, also experience theoretical conflict. A vertical knowledge structure is “‘a coherent, explicit and systematically principled structure, hierarchically organised’ which ‘attempts to create very general propositions and theories, which integrate knowledge at lower levels, and in this way shows underlying uniformities across an expanding range of apparently different phenomena‘” (as cited in Maton and Muller, 2007). Rather than multiple theories competing for dominance to explain common phenomena, vertical knowledge structures seek integration and uniformity of theory. In an academic feud, vertical knowledge structures maintain a single specialized language during the conflict and aim to integrate the two theories, discarding the error, rather than achieving dominance in the field at the expense of the other theory.
Vertical knowledge structures are seminally represented by the physical sciences; an example of a feud in said field would be Einstein’s critique of Newton’s mechanics. Einstein used the same specialized language as Newton (mathematics) and did not abandon the classical theory entirely–he integrated it into a larger theory of relativity. Even current theoretical conflicts between quantum mechanics and relativity are aiming for integration, as we see in the controversial attempt of string theory (so far as I understand the physics).
Following Bernstein, I contend that the Reading Wars occurred because, as Christie & Macken-Horarik (2007) argued, Literacy Studies is a horizontal knowledge structure, and the theoretical relations within such structures–far different from those in vertical knowledge structures–gave rise to the feud and enabled a war. Before engaging in an in-depth analysis, I want to review briefly what Bernstein’s theory predicts about a horizontal knowledge structure. If these predictions hold for the Reading Wars, it would further validate Christie & Macken-Horarik’s (2007) horizontal knowledge structure analysis and serve as a basis for explaining specifically what mechanisms in the structure constructed the conflict. I am basing these predictions on Maton & Muller’s (2007) review of Bernstein’s “sociology of knowledge” research program–specifically their subheading Differences between knowledge structures.
In Horizontal Knowledge Structures:
- “Knowledge develops through the addition of new languages.”
- Theories are “characterized by short-term obsolescence, only to reappear again some time in the future in a new guise”
- Specialized languages used for theoretical discourse are “based on different, often opposing assumptions, making it less clear that one is indeed speaking or writing” in a given language.
- “‘Each language [has] its own criteria for legitimate texts, what counts as evidence and what counts as legitimate questions or a legitimate problematic’” (Bernstein 1999, p. 163)
- “Integration and subsumption of past ideas [occurs] within each language. However, the capacity for such development across languages is limited.”
- Languages have relatively ‘weak grammars,’ which are ‘…explicit conceptual syntax[es] capable of “relatively” precise empirical descriptions and/or of generating formal modelling of empirical relations’ (Bernstein, 1999, p. 164)
- “Relations between languages or [theories] cannot be settled by empirical research and are confined to critique.”
- “Languages [do not] purport to share the same empirical referents.“
- “Significant changes… are all too often ideological rather than rational revolutions.”
- “Choices between competing claims to insight are based more on a ‘knower code’, that is to say, on who is making knowledge claims rather than on what is being claimed and how.”
In the tradition of Christie & Macken-Horarik (2007), I will now explain the conceptual basis for classifying Literacy Studies as a horizontal knowledge structure by analyzing the McKenna et al. and Edelsky (M-E) discourse.
In the table below, I provide representative examples from the M-E discourse that illustrate the features of a horizontal knowledge structure. I have numbered each feature for ease of explanation later. For the sake of clarity, I refer to McKenna’s first article as “McKenna et al., 1990a” and their rejoinder “McKenna et al., 1990b.” Citations from quotes are not linked. They can be found at McKenna et al., 1990a; Edelsky, and McKenna et al, 1990b.
|1. Multiple Languages||2. Theory Reoccurance||3. Opposing Assumptions|
|“[McKenna et al.] use incorrect discourse for whole language” (Edelsky, p. 7)
“If comparative effectiveness is not the ideal question, we would suggest that for the vast majority of present-day practitioners it is the question” (McKenna et al., 1990b, p. 13).
|“The authorities cited for opining that whole language is not well defined are two skills-based researchers who, in a recent meta-analysis, could not distinguish whole language from language experience approaches from activity approaches of the 1950s (Edelsky, p. 8)||“A view of how one comes to know is crucial to the credibility accorded evidence, and Edelsky is partially correct in pointing to paradigmatic differences as a source of problems” (McKenna et al., 1990b, p. 12)|
|4. Criteria for Texts||5. Integration||6. Weak Grammar|
|“Which works better? is a question emanating from the instrumental rationality of the dominant paradigm. It foregrounds method and efficiency rather than purpose-How to do it? rather than Why do it? or Whose interests does it serve?” (Edelsky, p. 9)||“Harste expressed the optimistic view that researchers from a variety of perspectives are moving toward a single model of literacy. He cautioned that we are not yet there and that it would be premature to describe such a model. He harbored healthy, scientific doubts, as do we. In contrast, a divisive, unscientific certitude has been widely observed among whole language proponents” (McKenna et al., 1990b, p. 13)||“Whole language is not well defined (Stahl & Miller, 1989). Rich (1985) described it as “an attitude, not methods” (p. 718), and Clarke (1987) described it as “[a] philosophy rather than a methodology” (p. 386). Goodman (1986) suggested that many methods are possible within a whole language classroom and that varieties of whole language are acceptable. Newman (1985) acknowledged, “I find myself in the uncomfortable position of being unable to tell you succinctly what ‘whole language’ is” (p. 1). Watson (1989) suggested that “most whole-language advocates reject a dictionary-type definition that can be looked up and memorized” (p. 131) because a given teacher’s definition be personal and unique” (McKenna et al., 1990a, p. 4).
“Their call for conceptual rigor in a definition of whole language and for procedural rigor in research design is no more than a thinly disguised demand that whole language be translated into terms that a skills model of reading and a positivist model of research. Ironically, what has been threatening to traditionalists about qualitative research in whole language may seem welcome by comparison with some of the new directions whole language research is taking” (Edelsky, p. 10).
“Just as a skills viewpoint does not require a belief in skills hierarchies but rather simply a belief in separable, acontextualized components, so too does a whole language viewpoint not preclude an attention to form”(Edelsky, p. 9).
|7. No Empirical Recourse||8. Different Empirical Referents||9. Knower Code|
|“One of the anonymous reviewers of our article offered this observation: ‘In essence, it is no longer
a scientific issue since the whole language people share a system of beliefs and they claim they have evidence to support their beliefs. But, when you look up what they cite as evidence, it is often just someone else’s published beliefs’” (McKenna et al., 1990b, p. 12)
|“Traditionalists have in fact maintained that whole language proponents (and their antecedents) have tended to confuse the process of fluent reading with the process of learning to read (Carroll, 1976; Chall, 1933; Stott, 1981)” (McKenna et. al, 1990a, p. 8)
“[Whole language] takes seriously a distinction between using language and doing language exercises, between doing science or history and doing exercises in science or history” (Edelsky, p. 8
|“Whose Agenda Is This Anyway? (Edelsky, p. 7, emphasis mine)
“[McKenna et al.] are ignorant that their presumptuousness in speaking for ‘the other side’ is one example of the kind of silencing within the academy that anthropologists (Clifford & Marcuss, 1986), sociologists (Denzín, 1990), critical educators (Brodkey, 1986; Gitlin, Siegel, & Boru, 1989), and others are now critiquing outside it. Just who gets to speak for whom?” (Edelsky, p. 8)
The M-E discourse shows all the features of a traditional academic feud in a horizontal knowledge structure. Some features are better illustrated than others, and my analysis will acknowledge areas where either Bernstein’s model does not adequately reflect the reality of the structure or it is unclear how to interpret the data. In the tradition of the qualitative paradigm, I will acknowledge my own approach to this research as being post-positivist, making me possibly more sympathetic to McKenna et al.’s perspective. I believe that Bernstein’s model can explain real social phenomena, such as the Reading Wars. I strive to maintain objectivity in applying this model, while acknowledging that Bernstein’s research was not exclusively post-positivist.
I will now provide an analysis of each of the different horizontal knowledge structure features evident in the M-E discourse with the aim to find how they contributed to the development of the Reading Wars.
1. “Knowledge develops through the addition of new languages.”
Perhaps the most explicit statement regarding the divergent discourses being written in these three articles is Edelsky’s: “[McKenna et al.] use incorrect discourse for whole language” (Edelsky, p. 7). Edelsky, at the very least, implies that there is a correct discourse for whole language–that is, Whole Language as a theory has a discourse strategy or strategies it employs to maintain and develop knowledge. Edelsky claims that McKenna et al. do not use this discourse, instead utilizing the discourse of Traditional Literacy, or Skills-Based Literacy. McKenna et al. do not dismiss the possibility of their Whole Language discourse being incorrect (p. 12), but they do challenge Edelsky’s claim about what kinds of questions are appropriate for education. This seems to imply some kind of metadiscourse into which both Whole Language and Traditional Literacy fit. It is, however, arguable whether McKenna et al.’s claim that relative effectiveness is on the minds of most teachers could be substantiated; at least in literacy educational research, comparative effectiveness of whole approaches fell under scrutiny after Bond & Dykstra’s (1967) first-grade studies, a massive comparative effort that yielded few results that could be generalized across all the programs that participated. Still, the two statements show that both sides seem to agree that they are disagreeing about certain constructions in the discourse, suggesting that they have different discourse strategies.
But are these multiple languages developing knowledge in the field? To consider this question, I examine the origin claims of these theories to determine how they emerged in the Literacy Studies field, thus demonstrating what they are adding to the conversation. Edelsky cites Goodman’s (1969) psycholinguistic work as the origin of Whole Language; this seems a legitimately novel development in Literacy Studies, as psycholinguistics itself was a relatively new field of study in the 20th century. However, Edelsky’s comment regarding Stahl & Miller’s (1989) relative inability to distinguish Whole Language from Language Experience (p. 12) may suggest that this discourse is an iteration of a previous discourse, namely Language Experience. McKenna et al. cite Samuels and Laberg (1974) in developing the Skills-Based theory of reading. Because of its structuralist nature, it seems less likely that Skills-Based theory is entirely novel. Whether these origins suggest novel development of knowledge is arguable; it is possible, even probable, that the theories are iterations of previously dominant theories. But both theories have productive research programs, as evident in their various duelling research citations. This suggests that Feature #5 is in play, because knowledge builds vertically within each theory, but they do not integrate findings across theories. The simultaneous theory building of divergent theories does challenge the pendulum metaphor so often invoked in education; the “vogue” theory does not necessarily mean that other theories are not being developed; McKenna et al., citing Stanovich (1990), argue similarly.
In short, it appears both Whole Language and Skills-Based Literacy approaches constitute two distinct specialized languages for developing knowledge in Literacy Studies. A more in-depth linguistic analysis of the discourse strategies would be necessary to fully establish this criteria solidly, as Christie & Macken-Horarik demonstrate(2007).
2. Theories are “characterized by short-term obsolescence, only to reappear again some time in the future in a new guise.”
This feature is, perhaps, the most difficult to demonstrate in the M-E discourse. The difficulty exists despite the fact that Whole Language in education policy is in relative decline (Stahl, 1999). The institution of NCLB was an instantiation of Skills-Based assessment and corresponding “teaching-to-the-test” Skills-Based instruction. Whole Language still exists as a part of NCTE, but in terms of national education policy, its influence appears diminished.
The difficulty in demonstrating this feature of horizontal knowledge structure in the M-E discourse lies not in a claim to obsolescence but in suggesting that either Whole Language or Skills-Based Literacy are “new guises” for older theories. Stahl (1999) traces some of the origins of Whole Language to earlier theories than the canonical Goodman origin–including Language Experience. He cites Y. Goodman (1989) for the conceptual tie to Language Experience. But it would be erroneous to conflate them as the same theory in different “guise.” Edelsky critiques Stahl & Miller (1989) for equating Whole Language and Language Experience. To suggest there are similarities does not seem uncontroversial, but to say that Whole Language is merely Language Experience in a “new guise” is more difficult to determine. The Skills-Based approach may find its intellectual precursors in Tylerism. But it would be difficult to argue that Skills-Based is a “new guise” of Tylerism as well.
The “new guise” feature may be understood if we consider the Integration feature of horizontal knowledge structures (Feature 5). Bernstein argued that Integration may happen within a language, but not across languages. Therefore, it is appropriate to interpret both Whole Language and Skills-Based theories as “new guises” in the sense that they integrate models from their intellectual predecessors. Certainly it is unnecessary to require all theories to be old; some may be novel. But there does seem to be evidence that both of these theories are part of existing theoretical traditions, and that they develop their knowledge vertically within those traditions.
3. Specialized languages used for theoretical discourse are “based on different, often opposing assumptions, making it less clear that one is indeed speaking or writing” in a given language.
That there are opposing assumptions, both McKenna et al. and Edelsky agree on. McKenna et al. (1990b) states: “A view of how one comes to know is crucial to the credibility accorded evidence, and Edelsky is partially correct in pointing to paradigmatic differences as a source of problems” (p. 12). The opposing assumptions are epistemological, with the Skills-Based Literacy aligning with (post?)positivist epistemology and Whole Language aligning with social constructivist epistemology. This categorization contrasts with both McKenna et al.’s (1990b) and Edelsky’s claims that the other theory is “essentially aparadigmatic” (McKenna et al., 1990b, p. 12). Edelsky is rather explicit on defining their epistemology, locating the Whole Language epistemology within social constructivism: “More recently, whole language has come to see reading as a social practice (Bloome & Solsken, 1988) entailing historically and culturally shared social activity and socially shaped psychological and linguistic processes” (p. 9). This statement is consistent with a social constructivist epistemology, as it defines a phenomenon as a “social practice,” thus implying that knowledge of the phenomenon is socially constructed.
Edelsky’s accusation of aparadigmaticity of McKenna et al. (p. 7) is more difficult to disprove. McKenna et al. never mention positivism or explicitly state their epistemological paradigm other than using the “skills-based” or “traditional” nomenclature. This does not mean they are being aparadigmatic if such silence on epistemology is a convention of positivist discourse. In this case, McKenna et al.’s objectivity, a feature of positivism (Creswell, 2012), seems to be constructed through the absence of language identifying with a particular paradigm. In other words, by NOT mentioning their epistemological paradigm, they may be engaging in positivist discourse strategies devised to avoid bias.
In short, both the Theory Reoccurance and Opposing Assumptions features seem to be present in the M-E discourse, further developing the classification of the horizontal knowledge structure of Literacy Studies. The Opposing Assumptions feature certainly may be seen as contributing to the Reading Wars feud, as both McKenna et al. and Edelsky acknowledge. It remains to be seen whether general knowledge Integration is possible in Literacy Studies if fundamental assumptions of epistemology cannot be reconciled.
4. “‘Each language [has] its own criteria for legitimate texts, what counts as evidence and what counts as legitimate questions or a legitimate problematic’” (Bernstein 1999, p. 163)
This feature of horizontal knowledge structures contains multiple parts, all of them specifying differing criteria between languages for determining various aspects of knowledge production and maintenance, such as texts, evidence, questions, and problems. While the criteria of each of these aspects of knowledge production and maintenance could be delineated for each language, Whole Language and Skills-Based Literacy, I will elect for a cursory statement of each of these from the M-E discourse.
In terms of texts, Whole Language seems to value the novel and ethnography Edelsky, p.10); Skills-Based Literacy favors the experimental report (McKenna et al., 1990a, p. 5). Perhaps alternative genres are accepted as research in each of these languages, but they are not as highly valued as those I’ve delineated here.
In terms of evidence, the divide is primarily between qualitative–valued by Whole Language–and quantitative–valued by Skills-Based Literacy (McKenna et al., 1990a, p. 4-5). Skills-Based Literacy seems possibly open to accepting qualitative data, but likely favors quantitative over qualitative. Whole Language mounts strong critiques of quantitative data (McKenna et al., 1990a, p.4-5); this form of data seems disfavored more than qualitative is disfavored by Skills-Based Literacy.
In terms of questions and problems (which I am conflating in the absence of significant distinction for the purposes of this analysis), Skills-Based Literacy values the comparative effectiveness question; Whole Language favors questions regarding purpose and power relations. Edelsky summarizes this effectively: “Which works better? is a question emanating from the instrumental rationality of the dominant paradigm. It foregrounds method and efficiency rather than purpose-How to do it? rather than Why do it? or Whose interests does it serve?” (p. 9). Edelsky thus shows that Whole Language favors addressing problems of power–”Whose interest does it serve?”–over the comparative question that McKenna et al. (1990a) so strongly advocate (p. 5).
The criteria for texts, evidence, and problems in both Whole Language and Skills-Based Literacy have thus been shown to contribute to the horizontal structure of knowledge in Literacy Studies. Because the criteria diverge in many respects, this feature certainly contributed to the Reading Wars. Whole Language’s disvaluing of quantitative research contributed to the feud on their end. However, as qualitative research has gained more acceptance, the distance between the two languages on the criteria of evidence has narrowed. Perhaps one early sign of such narrowing was the fact that McKenna et al. (1990a) proposed a research agenda with qualitative research designs as a major component, albeit one couched in skepticism (p. 5). If qualitative research gained more acceptance in Skills-Based research, the associated text genres and problems such evidence can answer would likely also transfer, leading possibly to Integration.
5. “Integration and subsumption of past ideas [occurs] within each language. However, the capacity for such development across languages is limited.”
This feature is critical to understanding the overall goal of horizontal knowledge structures, but it also serves as a primary mechanism in creating the theoretical divide of the Reading Wars. As I explained in my first post, the counterpart to horizontal knowledge structures are vertical, or hierarchical, knowledge structures. A vertical knowledge structure subsumes knowledge from the past, developing new knowledge–even new theories–based on the existing knowledge. The ultimate goal of this integration is the ability to produce abstract knowledge that can explain and predict a wide variety of phenomena within the scope of inquiry in the discipline. Bernstein’s (1999) explanation reveals this goal of producing abstract “propositions:”
“This form of knowledge attempts to create very general propositions and theories, which integrate knowledge at lower levels, and in this way shows underlying uniformities across an expanding of apparently different phenomena. Hierarchical knowledge structures appear, by their users, to be motivated towards greater and greater integrating propositions, operating at more and more abstract levels. Thus, it could be said that hierarchical knowledge structures are produced by an ‘integrating’ code” (as cited in Wignell, 2007)
This goal of Integration of knowledge is what we mean by scientific progress. Muller (2007) defines this progress from a positivist perspective: “The piled-up structure of laws related to one another by strict definition, in strict order of explanatory integratedness, that is, in strict order of their approximation to the truth” (“Kinds of Verticality”). Thus, integration is not a goal unto itself in vertical knowledge structures; it is a means by which we approximate the truth about the nature of a phenomenon–in the case of the Reading Wars, knowledge about the nature of reading.
Muller (2007) contends that constructivism, to which Edelsky and Whole Language proponents subscribe, is a critique of the progress concept in science. Constructivism’s critique of progress, according to Muller (actually Muller, 2007 citing Berlin, 2000 citing Vico, 1708–let’s see what APA style has to say about THAT monstrosity of a citation), included the claim that the divide between truth and knowledge was false; in Muller’s words, “we only know what we create” (“Progress: the very idea, and its sceptics”). It is unclear from the M-E discourse to what form of social constructivism Edelsky espouses; certainly there are strong and weak forms–Muller’s assertion seems to be a strong form. Regardless, the principle of integration in vertical knowledge structures–and its corresponding progress and positivist corollaries–stands opposed to the alternative knowledge development method utilized by horizontal knowledge structures, which is proliferation. Proliferation means that, rather than more levels of abstraction being added to a particular language, more distinct languages instead are produced. Language proliferation can occur during a language-internal theoretical conflict, which can sometimes force a schism and thus a new language (Muller, 2007). But the Reading Wars do not represent a schism of language-internal conflict; both languages have, by the time of the M-E discourse (1990), developed a grammar (which will be explained in the next post) and, as was demonstrated in my first post, are separate languages. How can Bernstein’s feature dichotomy of integration/proliferation account for the Reading Wars?
Wignell (2007) argues that:
“The stronger the boundaries around a discipline and the stronger the concord within that discipline, then the easier it will be for that discipline to prevent outside voices from intruding and for it to expel or re-educate internal dissenting voices…. Conversely, the weaker the boundaries around a discipline (the more it overlaps with other disciplines), the more it listens to other voices, and the more dissenting voices there are within that discipline, then the more likely it is that it will evolve as… a horizontal knowledge structure” (“Summary and conclusions”).
O’Halloran (2007), in summarizing Wignell’s (2007) conclusion, describes the effort to suppress dissenting voices as being “quietened” (“Bernstein’s grammaticality and knowledge structures”). Similarly, Edelsky suggests that McKenna et al. is “silencing,” (p. 8) Whole Language. McKenna et al. (1990b) protests this claim (p. 13), arguing that they seek to collaborate with Whole Language. Citing Stanovich, they suggest a “peaceful coexistence” (p. 12) and even go as far to suggest that Whole Language and Traditional Literacy are not paradigmatically “incommensurable” (p. 13). This notion of epistemological compatibility was seriously challenged in my third post. McKenna et al.’s (1990b) then make a case for Integration, deputizing a Whole Language proponent, Harste, into their argument:
“Harste expressed the optimistic view that researchers from a variety of perspectives are moving toward a single model of literacy. He cautioned that we are not yet there and that it would be premature to describe such a model. He harbored healthy, scientific doubts, as do we” (p. 13).
In their entreaty to Integration and collaboration, McKenna et al. (1990b) treat Literacy Studies like a vertical knowledge structure. This element of the M-E discourse may be seen as evidence against my thesis that Literacy Studies is, indeed, a horizontal knowledge structure. Furthermore, McKenna et al. (1990a) advocate specific tests for measuring comprehension, which illustrates a methodological Integration and an subsumption of the knowledge gained from those tests (p. 5). However, McKenna et al. concede: “Perhaps the main lesson from the present interchange is that it occurred. This, surely, is a step in the right direction” (p. 13). Their first sentence recognizes the difference of opinions as represented by the discourse; their second sentence aims to frame such a discourse as “a step in the right direction.” The word “direction” certainly presupposes a goal–and that goal, as suggested earlier, may be Integration. Thus, even as they recognize the feud, they seem to express an implicit belief in progress, as consistent with their positivist epistemology.
Edelsky is describing the discipline as a horizontal knowledge structure. McKenna et al. is recognizing the feud and suggesting the languages can coexist–which implies they see the field as a horizontal knowledge structure. But they also indicate that Integration is a goal of research in the field, which implies they see the field as a vertical knowledge structure. This dissonance must be resolved for the horizontal knowledge structure thesis to hold.
This dissonance is based on the fact that McKenna et al. and Edelsky hold opposing assumptions about epistemology. The debate between the positivists and the constructivists is a historical scientific feud that, in Muller’s (2007) analysis, is reframed thusly by Bernstein:
“The logical positivists (or realists) were right, but only in respect of hierarchical knowledge structures; the non-realists (Kuhn and after) were likewise right, but only in respect of horizontal knowledge structures” (“Knowledge and the dilemma of progress”)
Thus, McKenna et al. and Edelsky can both be said to have an accurate analysis insofar as the object of their description is a vertical knowledge structure or horizontal knowledge structure respectively. I propose that Edelsky is describing Literacy Studies as a whole–which is a horizontal knowledge structure. Each characteristic feature of horizontal knowledge structures, including this one, is represented in the M-E discourse, which is a discourse within Literacy Studies–not just within a single language. Conversely, McKenna et al. is describing Skills-Based Literacy when discussing features of vertical knowledge structures. As Maton & Muller (2007) assert, individual languages within a horizontal knowledge structure may be internally structured vertically. Knowledge within Skills-Based Literacy may be integrated with past iterations of Skills-Based Literacy. But verticality–that is, Integration–does not occur across languages–that is, Integrating knowledge from Skills-Based Literacy and Whole Language is difficult because of the opposing epistemologies. McKenna et al. acknowledge this much by expressing “healthy scientific doubts” (p. 13) about Integration. Yet they do seem to describe a hope for Integration and suggest methodological Integration with the comprehension tests, implying some degree of Integration across languages in Literacy Studies as a discipline. Within my analysis of Literacy Studies as a horizontal knowledge structure, the goal of the discipline cannot be Integration across languages because scientific progress is not the “direction” Literacy Studies as a horizontal knowledge structure is going. Literacy Studies is aiming to Proliferate more languages.
McKenna et al.’s imposing of the progress goal on the horizontal knowledge structure has two possible explanations: either Bernstein’s model is too restrictive–that is, Literacy Studies is some kind of hybrid between a horizontal and vertical knowledge structures; or McKenna et al. are engaging in O’Hollaran’s “quieten[ing]“. Edelsky’s analysis supports the latter conclusion, arguing that Skills-Based Literacy has achieved dominance in the field, which gave them a position to attempt “silencing” Whole Language (p. 8). This silencing may be represented in the imposing of a progress/integration goal on Whole Language, but also in rejecting and replacing Whole Language’s criteria for a problem. Recall that Whole Language, by the Criteria feature, reject comparative effectiveness as a problematic. Yet McKenna et al. (1990b) insist on it: “If comparative effectiveness is not the ideal question, we would suggest that for the vast majority of present-day practitioners it is the question” (p. 13). In this rejoinder to Edelsky’s critique of appropriate questions, McKenna et al. make a bandwagon appeal, suggesting that most teachers favor the comparative effectiveness question over the questions of power that are favored by Whole Language, implying that the Whole Language criteria for question does not align with the majority of the clientele the field serves. It is an interesting move because it invokes, on some level, a conceptualization of knowledge–in this case, knowledge of criteria for problems in the field–by social construction. That is, the comparative effectiveness question is more important because most people think so. In short, there is some evidence to suggest McKenna et al. are attempting to quiet Edelsky and Whole Language by substituting Whole Language questions for their own, imposing methodological integration, and attempting to establish a goal of progress and integration–all characteristics of Skills-Based Literacy and vertical knowledge structures in general. By attempting to quiet the dissenting voice of Whole Language, Skills-Based Literacy can, by virtue of dominance, describe the discipline as a vertical knowledge structure so long as the dissenting voices are quiet. But the reality of the Reading Wars as a theoretical conflict between demonstrably different languages with different assumptions and criteria for questions, texts, and evidence cast doubt on such a description and point towards Literacy Studies being a horizontal knowledge structure.
After discussing the issue of Integration with some colleagues in linguistics via Twitter, I would like to respond to three Tweets because they raise problems and objections to my analysis that are productive.
I have argued that one possible explanation for the dissonance between McKenna et al.’s and Edelsky’s descriptions of the Literacy Studies knowledge structure–which are vertical and horizontal, respectively–may be attributed to the inadequacy of Bernstein’s model. Grieve-Smith’s Tweet raised this objection, framing it as deconstructing a binary: “This vertical/horizontal binary seems to limiting. Why not just talk about how many overlapping theories there are in a field?” (Grieve-Smith, 2013). This raises the question of whether Bernstein’s model might be better represented as a cline, rather than a binary. That is, “horizontality” might be a matter of degree rather than a binary feature. Martin (2007) suggests as much when he argues that knowledge structures may be categorized by degrees of verticality–that is, the amount that the discipline tries to Integrate ideas (“Taking Stock: Future directions in research on knowledge structure”). This seems to align with Grieve-Smith’s idea of “overlapping theories.” Wignell’s (2007) analysis of the structure of sociology offers another perspective: orthodox schools of a discipline may be vertically structured, but the field itself has a horizontal knowledge structure. This might be modeled as a funnel–where certain elements of the discipline comprise the thin, vertical spout and the less-agreed-upon parts comprise the conical lip.
I certainly cannot rebut entirely Grieve-Smith and Martin’s critique of the structuralist horizontal/vertical binary. There do seem to be degrees to which disciplines agree/disagree and overlap/dissent on issues. For example, the particular issue at hand in this analysis–the Reading Wars–are arguably over. If the issues represented in this feud are settled, does this mean that knowledge has been Integrated? It is possible that the “balanced-literacy” paradigm represents this Integration. But a more recent debate between McVee et al. v. Gredler v. Krasny et al. v. McVee et al. in Review of Educational Research(2007) shows that comprehension is a less-vertical element of the field. So perhaps the funnel model of Wignell (2007) has some merit for Literacy Studies–because while the Reading Wars are over, the Comprehension language is less established.
I only have two responses to this critique: first, deconstruction necessarily limits the predictive power of any particular structuralist theory. If we represent verticality as a cline, we can be less sure about the predictions we can make about disciplines we attempt to describe with the model. This seems consistent with the assumptions of post-positivism, the philosophical tradition from which I anchor my analysis. From this perspective, I would argue that while there is a cline of verticality, it is still useful to describe a field as more or less a vertical/horizontal knowledge structure for the purposes of understanding phenomena in a field. It would be helpful to be able to determine what the “tipping point” of verticality is in order to classify a field as either.
To this end, I want to address another problem that arose in Twitter discussions of my previous post. That is, intentionality. The progress goal of Integrated vertical knowledge structures is described adequately by Muller (2007). But is there an intention of horizontal knowledge structures? In Johnson’s (2013) Tweet lies a hint: “i think in linguistics the disparate theories all think they’re right and the other guy’s wrong? also some gaps got really big?” If verticality is measured by a level of intentional Integration by its users driven by a progress (dare I say) “narrative”, then what is the goal of users of horizontal knowledge structures? Johnson suggests that in linguistics, a horizontal knowledge structure (at Purdue, I am a Hallidayan studying among Chomskyans in a relationship with Kachruan), the goal is being “right.” Unlike Integration in vertical knowledge structures, we cannot simply ascribe the process of Proliferation of horizontal knowledge structures as the intention of their users. Linguists do not aim to create new theories; they aim to be “right.” But within horizontal knowledge structures, “rightness” is a function of power. Thus I propose that, in conjunction with the degree of Integration as a measure of verticality, the measure of horizontality may be the degree of consolidated power within the field. McKenna et al. (1990b) may or may not be right about denying a “zero-sum” feature of the Literacy Studies field (p. 11); but I argue that regardless, the degree to which a language in a knowledge structure has influence and adherents, both individual and corporate, determines their consolidated power in the field and thus their status as a horizontal-ish knowledge structure.
This framework may explain the dissonance between the degree of Integration ascribed to Literacy Studies by both McKenna et al. and Edelsky. Because more power was consolidated within Skills-Based Literacy, McKenna et al. were more inclined either to 1) further consolidate the power by silencing Edelsky (as Edelsky maintained) or 2) interpret their power as representative of the structure of the field. This latter interpretation seems more fair and admittedly accurate for the data presented in the M-E discourse. There are multiple signals from McKenna et al. that at times conflict with a horizontal or vertical structure analysis. It seems more likely, based on their appeal to collaboration (1990a, p. 6), that McKenna et al. extended their own goal of scientific progress to also be the goal of the Literacy Studies field–while ignoring the positivist philosophical implications progress espoused. Meanwhile, Edelsky and Whole Language were motivated by power because of their understanding of the discipline as a horizontal knowledge structure. This should not be interpreted as insidious; when a lower-status group seeks more power, it may be an indicator of disenfranchisement or oppression. Certainly both had reasons to believe that the field was structured the way the thought it was. But it may be that Power Consolidation and not Knowledge Integration is the cline along which the horizontal/vertical knowledge structure rests. A particular language may seek to Integrate knowledge all it wants, as Skills-Based literacy attempted to do through collaborating with Whole Language during the Reading Wars. But if a particular language seeks power in the field (not necessarily dominance, but at least equality), then the result may be horizontality rather than verticality. Power rather than Integration may be the more predictive feature of knowledge structure.
6. “Languages have relatively ‘weak grammars,’ which are ‘…explicit conceptual syntax[es] capable of “relatively” precise empirical descriptions and/or of generating formal modelling of empirical relations’ (Bernstein, 1999, p. 164)
Both Grieve-Smith and Johnson’s (2013) Tweets illustrate the sixth feature of horizontal knowledge structures: weak grammar. Martin (2007) argues that the fundamental disagreement among linguists is the definition of “language.” This, in turn, creates the horizontal knowledge structure within linguistics in part because the “empirical description” of language varies across conceptual languages in linguistics. Ginsberg’s (2013) Tweet summarizes this reality well: “Srikant Sarangi once said to me that what we do is essentially methodology. Days like this, I think what we do is terminology.” Ginsberg suggests that defining terminology is often the work required within in a horizontal knowledge structure. It is in this work of defining terminology that the phenomeon of “weak grammar” becomes evident as a mechanism in constructing the Reading Wars in general and the M-E discourse in particular.
The varying degree to which Whole Language has (Edelsky, p. 8) or has not (McKenna et al., 1990a, p. 4) been defined has long been debated. This seems an obvious feature of weak grammar. However, what is often not noted is that the language of Skills-Based Literacy also lacks terminological specificity: “Just as a skills viewpoint does not require a belief in skills hierarchies but rather simply a belief in separable, acontextualized components, so too does a whole language viewpoint not preclude an attention to form” (Edelsky, p. 9). Edelsky reveals in this passage the fact that the concept of “skills” is itself amorphous, not constrained by empirical referents or hierarchies, but nebulously defined as “separable, acontextualized components.” Likewise, just as Martin (2007) argued with linguistics, so too we find in Literacy Studies the fundamental phenomenon–that is, reading–is defined differently by each group. Whole Language defines it as a “sociopsycholinguistic process” (Edelsky, p. 8), whereas Skills-Based literacy seems to define it as a mastery of various skills (McKenna et al. 1990a, p. 3). The difference in empirical referent is, perhaps, not irreconcilable, but at the point of the M-E discourse and the Reading Wars, the difference was enough to distinguish the languages from each other.
This differing of empirical referents leads to the next feature of horizontal knowledge structures, which I will address in the next post. But first, I would like to summarize the issues addressed in this post. First, I attempted to address the hybrid explanation, recognizing that verticality is a matter of degree and less a binary of a knowledge structure. I argued that even if verticality is a cline, there may be a “tipping point” at which a given structure may be more or less a horizontal/vertical knowledge structure. Second, I contended that another way to determine this “tipping point” may be the degree to which power is consolidated within a specific language. If power is highly consolidated within a given language (e.g. the Standard Model of physics or Evolution of biology), then we may argue that it is vertical. Third, I recognized Wignell’s (2007) hybrid explanation and proposed a funnel shape as a model, where orthodoxy represents the spout and heterodoxy represents the lip. Lastly, I argued that Literacy Studies, by virtue of unclear empirical referents, illustrates the feature of weak grammar. The weak grammar of both Whole Language and Skills-Based Literacy contributed to the Reading Wars by providing space for negotiation over the definition of the fundamental phenomenon in question–that is, reading. As an extension, all relations between the fundamental concepts of the discipline were thrown into question. The weak grammar component of Literacy Studies’ horizontal knowledge structure significantly contributed to the traditional academic feud that became the Reading Wars.
7. “Relations between languages or [theories] cannot be settled by empirical research and are confined to critique.”
8. “Languages [do not] purport to share the same empirical referents.”
In a vertical knowledge structure, conflicts are resolved by empirical research. Maton & Muller (2007) explain that during a feud where Integration occurs, the new theory must be able to “explain [its] predecessor’s success” (“Differences between knowledge structures”). Following Popper (1994), the new model must explain a phenomenon at least as well as an old model–plus provide better predictions than the old model. In order to do this, conflicts must be able to compare different theoretical predictions of the same empirical phenomenon. The comparison question is essential in vertical knowledge structures to resolve feuds and Integrate knowledge.
But comparison did not happen in the Reading Wars. McKenna et al. (1990a) argue for a research agenda aimed at the comparison task, although their question asks to compare effectiveness of pedagogical methods (p.5-6). Edelsky rejected the research agenda. Why? McKenna et al. (1990b) offer one possible answer: “One of the anonymous reviewers of our article offered this observation: ‘In essence, it is no longer a scientific issue since the whole language people share a system of beliefs and they claim they have evidence to support their beliefs. But, when you look up what they cite as evidence, it is often just someone else’s published beliefs’” (p. 12, emphases mine). The word scientific here might be read as empirical–that is, the issue cannot be resolved by appealing to comparative research on some common empirical referent. The issue is less one-sided than McKenna et al.’s reviewer describes it: the system of belief is not just the efficacy of Whole Language as a pedagogy, as may be extrapolated by the comparative effectiveness focus of McKenna et al. overall, but includes the definition of the fundamental phenomenon of reading. Skills-Based Literacy’s definition of reading and Whole Language’s definition of reading are not the same; therefore, any comparison is not valid if made between each model’s predictions about reading (or, jumping ahead with McKenna et al., between each model’s methods of teaching reading) since predictions would be of different empirical phenomenon. Put more simply, McKenna et al. is making predictions about apples; Edelsky is making predictions about oranges. If we try to compare the predictions to see which one is more accurate, we will be conflating apples with oranges.
Both McKenna et al. and Edelsky recognized, on some level, there was definitional differences between them. McKenna et al. (1990a) argued that
“whole language proponents (and their antecedents) have tended to confuse the process of fluent reading with the process of learning to read (Carroll, 1976; Chall, 1933; Stott, 1981)” (p. 8). In this passage, McKenna et al. critiques Whole Language’s definition of “fluent reading,” arguing that in the Skills-Based Literacy’s language, the phenomenon deemed “fluent reading” by Whole Language would be defined as “the process of learning to read.” The relationship between these definitions cannot be resolved through appeal to “empirical research” because in order to do “empirical research,” one needs to have a common understanding of the object of study–thus, as Maton & Muller (2007) argue, the issue can only be resolved through “critique.” In like fashion, Edelsky highlights another definitional divide: “[Whole language] takes seriously a distinction between using language and doing language exercises, between doing science or history and doing exercises in science or history” (p. 8). Edelsky argues here for the social context of reading being integral to the definition of reading, and thus critiques the Skills-Based inclusion of acontextual “language exercises” as part of “reading.” Again, the difference can only resolved by argument.
If we take my last post’s explanation of the intentions of horizontal knowledge structures, there is one other way to resolve the issue: consolidation of power. That is, rather than a definition being agreed upon by rational argument, the definition that amasses the most adherents and political power may become the common referent upon which empirical research may be based. This would allow for the process of Integration to begin and verticality to arise. But verticality itself, as this discussion has shown, is dependent upon strong grammar. Integration cannot happen without referring to a common, explicit, non-overlapping phenomenon. But strength of grammar may itself be a function of consolidated power. Tying power to empirical referents is, admittedly, a part of the social constructivist critique of positivism. But the social constructivist paradigm that seems to describe knowledge structuring and making in horizontal knowledge structures may itself limit the ability of the field to make rational predictions. At times it seems that research within horizontal knowledge structures is circular: that is, we often find results that our models predict–not because some objective reality creates only these results–but because our definitions of phenomenon constrain the observer to seek only data that satisfy the features of the model. It may be that Skills-Based Literacy research finds that students demonstrate skills in reading because they have defined reading as unified use of skills. Similarly, Whole Language research may find that students engage in contextualized meaning-making in reading because they have defined reading as contextualized meaning-making.
But from a post-positivist perspective, I am inclined to suggest there are two objective possibilities. Either 1) both definitions are useful and can be productive for empirical research if they recognize the definitional differences and acknowledge that they are studying different phenomenon Or, as perhaps is the aim of Balanced-Literacy, 2) both definitions are accurate descriptions of the same phenomenon–reading–and that, by combining the definitions, the phenomenon might be more restrictively defined, which could lead to a stronger grammar and the possibility of scientific progress on reading. But, as was argued in my third post, the fundamental epistemological assumptions and methodological criteria of each language are not obviously compatible. If Balanced-Literacy seeks to Integrate knowledge, resolving the philosophical differences would be necessary. That endeavor is not within the scope of this blogpost.
9. “Choices between competing claims to insight are based more on a ‘knower code’, that is to say, on who is making knowledge claims rather than on what is being claimed and how.”
Bernstein’s final feature of horizontal knowledge structures has been further developed my Maton (2007), and I will use part of his description of “knower code” to analyze the M-E discourse here (his full model of knowledge-knower structures is beyond the scope of this analysis). To be a legitimate “knower” in a horizontal knowledge structure is to have the “sensibilities, character, and personal attributes of an ideal knower” within the knowledge structure (“Knower structure”). The idea that there is a “knower code” for Literacy Studies in general, and each language in particular, is quite evident in Edelsky’s title: “Whose Agenda Is This Anyway?” (Edelsky, p. 7, emphasis mine). If we consider alternative wh-word possibilities that might be substituted in Edelsky’s title–”What Agenda Is This Anyway?”; “Why This Agenda Anyway?”; “Where Is This Agenda From Anyway?”–it is the choice of the genitive “Whose” that locates the focus of the field, from Edelsky’s perspective, not in objective knowledge, but in the makers of it. Edelsky argument,challenging McKenna et al. highlights not just the other features of horizontal knowledge structures that this series has attempt to elucidate, but also the personal qualities that make McKenna et al. not “ideal knowers” in the Whole Language community: “[McKenna et al.] are ignorant that their presumptuousness in speaking for ‘the other side’ is one example of the kind of silencing within the academy that anthropologists (Clifford & Marcuss, 1986), sociologists (Denzín, 1990), critical educators (Brodkey, 1986; Gitlin, Siegel, & Boru, 1989), and others are now critiquing outside it. Just who gets to speak for whom?” (p. 8, emphasis mine). Edelsky deputizes the personal characteristic of “presumptuousness” of the researchers as a rationale for rejecting their proposals. Thus, rather than the strictly logos appeal of the vertical knowledge structure discourse like physics, the horizontal knowledge structure of Literacy Studies allows appeals to ethos as valid moves in its discourse.
Edelsky is not alone; in one passage, McKenna et al. (1990b), perhaps signalling a horizontal knowledge structure interpretation of the discipline, describe one of the personal characteristics that disqualifies Edelsky from being an “ideal knower” in Skills-Based Literacy: “A divisive, unscientific certitude has been widely observed among whole language proponents…. Unreflective certitude, and the arrogance it fosters, are attributes we can ill afford at present” (p. 13). McKenna et al., despite espousing positivism and its conventions of logos-centric discourse, engage in an ethos appeal, invoking the personal characteristics of arrogance and “unreflective certitude” as antithetical to the sensibility of an “ideal knower” in the discipline. If nothing else, this display of ethical appeals stands as an aberration to normally logos-centric discourse in the sciences; yet Bernstein’s model of horizontal knowledge structures, combined with Maton’s extension, predict that this behavior will happen if knowledge in a discipline is so structured.
Maton also argues that the “ideal knower” may have some “biological/social” basis (“Knower structures”). For example, within the disciplines of divinity–again, a horizontal knowledge structure–there may be a sex basis for “ideal knower,” where only men are permitted or encouraged to participate in knowledge construction and maintenance. Similarly, class may play a role in determining the “ideal knower;” in mid-20th century humanities, Maton argues, the “ideal knower” was an “English gentleman,” suggesting higher social status as a factor (“Knower structures”). It may be that the genders represented in the M-E discourse are not coincidental; however, more evidence than can be presented here would need to be gathered to make any particular gender basis for an “ideal knower” in Whole Language or Skills-Based Literacy. Similarly, Whole Language’s political leanings might suggest there is a social basis for the “ideal knower” in that particular community. In short, aspects of character combined with biological/social features are present in constructing “ideal knowers” in the M-E discourse, supporting the horizontal knowledge structure analysis. However, I have left unexamined the possibility that opposing conceptions of “ideal knowers” contributed to the Reading Wars. Because this feature seeks to define a character, more evidence across multiple discourse would be necessary to begin to construct an outline of the “ideal knower” in these different schools of thought.
In a revision of my thesis, which originally but no longer contains a second analysis of the debate I highlighted in the Review of Educational Research, I will conclude by reviewing the main features of horizontal knowledge structures and argue that they are the primary mechanisms that facilitated the Reading Wars. I will then offer suggestions based on this analysis on how we might avoid wars and have more productive feuds in Literacy Studies
I have demonstrated in this series the various features of horizontal knowledge structures, abstracted from Maton & Muller (2007) and based on Bernstein’s theory, that are present in the M-E discourse and that indicate that Literacy Studies is a horizontal knowledge structure. These features included: 1) multiple languages, 2) theory reoccurrence, 3) opposing assumptions, 4) criteria for texts/data/problems, 5) lack of integration, 6) weak grammar, 7) lack of common empirical referents, and 8) knower code. This horizontal knowledge structure model reveals the primary mechanisms that contributed to the Reading Wars, showing what the fundamental disagreements were: the two approaches spoke different discourses; they held opposing epistemological philosophies; they had different criteria for what constitutes a problem; they did not seek equitable integration; and they talked past each other by virtue of imprecise empirical descriptions and non-aligning empirical referents for the same phenomenon–reading. On the whole, the Reading Wars happened because Literacy Studies is a horizontal knowledge structure; many of the disagreements that occurred in the Reading Wars would likely not been possible in a vertical knowledge structure because they lacked the particular theoretical mechanisms to trigger such disagreements.
My final conclusion, as consistent with my revised thesis above, is an admonition based on my analysis of this traditional academic feud so intense it achieved “war” status. What can we do to avoid wars? It seems that we need to work to resolve some of the disciplinary conflicts that Bernstein’s model reveals. If we want to avoid wars, we need to work towards a common discourse, a common epistemology, a common set of criteria for determining problems in the field, and a common and rigorously defined conceptual grammar. But most importantly, we need to establish a way to achieve equitable integration of knowledge. We need to agree on, as McKenna et al. (1990b) and Harste suggested (p. 13), a single model. But we should not seek to integrate inequitably, through consolidating power unilaterally. The Reading Wars are a testament to the power consolidation intention in horizontal knowledge structures. Equitable integration can only happen when we achieve all of the previously stated goals–and when we make both social justice and scientific progress equal goals in the discipline.
All major sources are linked. Unlinked citations have either been linked previously or can be found in the bibliographies of the sources that are linked.
Tweets Cited (Extrapolated APA Style)