27 August 2010

ELT in Japan (Issue #3, August 2010) Feature 1: Schema theory: Is this set of concepts relevant to TEFL?

ELT in Japan (Issue #3, August 2010)

Feature 1: Schema theory: Is this set of concepts relevant to TEFL?
Charles Jannuzi, University of Fukui, Japan


Schema theory (or actually theories) is an important area of discussion that could be added to ELT's growing catalogue of 'key concepts'. This is especially true because of the prevalence of 'top down' approaches to teaching communicative activities such as EFL 'listening' and 'reading'.

It should be emphasized from the very beginning of this discussion that, however we define it, a 'schema' is NOT something that is real in the sense that neuroscientists can or will ever literally put their finger on it. It is basically an empirically 'unverifiable object', like a myriad of other 'things' that we talk about stipulatively in education, linguistics and ELT, e.g., IQ, linguistic competence, universal grammar, general learning ability, learning styles, multiple intelligences, the Language Acquisition Device (LAD), and so on.

Still, however problematic or intangible the the nature of the things and structures behind the concept are, the impact schema theory has had on language teaching is very real. Therefore, as classroom teachers we might ask ourselves, "Is schema theory relevant and useful in my teaching? Does it help me to teach so my students learn a foreign language?" In other words, just what is the pragmatic value of this well-known theory? If valuable, just how so?"


Keeping also in mind that a definition of something is not the thing itself, let us proceed with some attempts to define a schema. Cook (1997) defines the concept as "a mental representation of a typical instance" which helps people to make sense of the world more quickly because "people understand new experiences by activating relevant schemas in their mind" (p. 86). Thus, according to the theory, we can understand a new situation or a new set of raw data (to include linguistic input, oral, written, or multimedia/ hypertextual) before all of that data has been received via bottom-up processing by matching parts to known whole types.

So prior learning helps us to understand new data and experiences. That is, unless of course, we have insufficient schemata for application or are misapplying the ones we do have, in which case we are more likely to misperceive and misunderstand something in a top-down way. This is what happens when we say, "My mind seems to be playing tricks on me," or "It was right under my nose, but I never saw it." In the case of linguistic processing and perception required in everyday communication, we often end up apologizing because we misheard or misread something: "I'm sorry, I thought you said..." or "I thought what I read was..." These are also examples of the influence of schemata on language processing and memory.

The earlier and more influential source texts for the extension and application of schema theory to Foreign Language Literacy (FLL) go back to Goodman (1967)--who posited that L1 reading was a "psycholinguistic guessing game--as well as Widdowson (1983), and Carrell and Eisterhold (1983, 1984). More or less, because of the impact and influence of these and subsequent works referring to them, schema theory and top-down approaches have been very much a part of mainstream ELT, perhaps especially noticeable in mainstream approaches to teaching EFL listening and reading.

Types of schema

In literacy and ELT it is often said that there are two types of schema: formal and content. Formal schemata are described as abstract, encoded, internalized, coherent patterns of meta-linguistic, discoursal, and textual organization (e.g., rhetorical patterns, story grammar, narrative scripts) that guide expectations in our attempts to understand a meaningful piece of language. Content schemata are less abstract and must presumably be about the physical world of discernible objects and actions.

An interesting personal example of the effect of formal schemata for me is televison news. Partly because of the way it is presented, I find the TV news from Japan often harder at first to grasp for gist than I do broadcasts from Germany, despite the fact that I know much more Japanese than I do German. I have to assume that this is because German news programs are structured more like the American and British news programs with which I am familiar.

Still, not all of the difficulty in understanding the news is attributable to how the stories are put together and delivered over the TV. Also different are the things actually depicted in the news, which illustrates as well the importance of content schemata. In terms of the TV news example, I would assert that it is most definitely the case that the subject matter of German TV news is also much more like my native culture's (America) than that of my current resident culture (Japan).

Cultural influences on schemata

Some schemata are said to be culturally specific. Presumably such cultural schemata could be either of the formal or content type. For example, much has been made of the way different cultures organize language into meaningful written texts and how members of these cultures then approach reading, interpreting, and using these texts. Or to give a very down-to-earth, everyday example, having lived in Japan for twenty years, I can assure you that the typical "script" (a type of FORMAL schema) for ordering a meal in a restaurant is different than it is in the USA. For example, I still sometimes have trouble with the question where I am asked if I want the beverage I have ordered with or after the meal.

Obviously, content schemata could be specific to a culture, as culture does help to determine our life experiences and how we make meaning of them. What would it mean to teach the word 'cheese' in an EFL class where this food isn't very popular or where no such food existed? Could the word and the concepts and associations it might come to represent mean the same thing to someone who has never eaten it? Or, imagine teaching the words to describe the four seasons of a temperate climate to someone who lives in a tropical desert.

A new type of schema: Abstract

Breaking from the established formal- content schemata distinction, Oller (1995) argues for making a three-part distinction amongst schemata: formal, content, and abstract.. Using terms from Peircean logic, he associates each type of schemata with a particular kind of inference: deduction, induction, and abduction. Deduction is most often identified with logical inference. Induction is really empirically probabilistic reasoning (i.e., directly experiential). The last, 'abduction', is supposed to be a type of reasoning that appears to be experientially informed but intuitive, that is, beyond both logical inference and probability derived only from lived experience.

Although acknowledging that formal schemata and content schemata are well known in applied linguistics literature, he argues that the relatively unknown abstract type are logically necessary for the theory to be complete. He writes:

     Abstract schemata must constitute a third class, and unlike the other two this class of 
     schemata has not been recognized previously in schema theory as a distinct category (unless
     grammars themselves are taken to be schemata). Abstract schemata carry the inductive
     integration to the completely general (abstract, non-material, non-syntacticized) level of pure
     symbols (in Peirce's sense of the term 'symbol'). Such a schema is necessary, for instance, if
     we are to draw inferences from representations that are independent of any particular  
     case or  any finite number of actual cases in the material world. For instance, if hotels
     are businesses that aim to make a profit they must generally charge more for their services
     than those services cost the owners. Thus, deductive inferences give us a great deal of
     information about all possible hotels that could not be acquired by merely examining or
     auditing the records of however many individual cases we might gain access to by whatever
     methods might be applied. (pp. 286-287)

One immediate challenge for classroom teaching here is to apply the concept of abstract schemata and deductive reasoning to our own pedagogical procedures and decision making. The numerous empirical/ quantitative studies that Applied Linguistics provides and the inductive reasoning underlying them, without benefit of deductive reasoning, may yield nothing more than rather empty formalisms about teaching. More than any other type, the truly powerful, operable schemata telling us how to proceed in the classroom come more from abstraction and deductive reasoning consistent with material reality and our own actual experiences as language learners and teachers.

Oller himself is not afraid to draw implications for literacy and language learning when he writes:

     Hence in relative degrees of theoretical power, the connectedness afforded by deduction, is  
     higher than any afforded by induction, which in turn is higher than any afforded by abduction.
     As a result, some interesting hypotheses about language acquisition and use, and especially
     about discourse processing can be formulated. The theory predicts, for instance, that formal 
     schemata will, other things being equal, have greater power than content schemata in 
     facilitating discourse comprehension and hence communication, language acquisition, etc.
     Also, other things being equal, abstract schemata, involving symbols and their definitions, will
     be more helpful still. (pp. 287-288)

Interestingly, and of special note for teachers of literacy and ELT, Oller's three-part schema theory is essentially an interactive one, involving top-down and bottom-up processing: "it comes out there in a completely general way that all comprehension and learning must be grounded ultimately in bottom-up processing of perceptual representations linked to actual material facts" (p. 299). Such a re-conceptualization of schema theory could have profound applications for ELT.

Schemata and the top-down vs. bottom-up processing debate

Oller's work can be fit into the long-running debate about whether reading predominately involves top-down or bottom-up processing. Top-down approaches to teaching reading and text comprehension are undoubtedly very popular in ELT. In these approaches to reading and listening, there is an emphasis on what a language user/learner psychologically brings to a given communicative situation. This prior knowledge that the individual systematizes and internalizes is equated with schemata. One reason for ELT's almost unexamined faith in top-down approaches might be that bottom-up approaches are associated with an often discredited behaviorism and its LT analog, the audio-lingual approach. It could also be argued that many of the psycholinguistic details of bottom-up accounts of language processing have never been well understood by actual classroom practitioners.

Bottom-up approaches to FLT and FLL might minimize the importance of activating schemata because with these approaches comprehension and learning from text are for the most part viewed as being driven by the proper perception and processing of text-based clues (either written or spoken, e.g., units of perceived sound, letters/letter combinations, syllables/word parts, whole sight words, etc.). Remember that within a top-down view of comprehension, the key to how understanding is possible is with the emphasis on the individual language user's/learner's background knowledge (of the target language, of the target language's culture or cultures, of the world) and their ability to direct these cognitive resources toward the text so as to reconstruct the message--write it, if you will, into working and short-term memories.

The entire process of comprehending a reading or listening text requires top-down processing in order to meet and make sense of incoming data and to fill in where the bottom-up information (or reader/listener perception of it) is incomplete or garbled.

Interactive models

Of course, numerous compromise models have been offered in order to overcome the simple dichotomy of bottom-up vs. top-down; these are versions of the very appealing interactive (or two-way) model. Interactive theories attempt to reconcile and combine the apparent strengths of the two opposing views (top-down vs. bottom-up, that is) while eliminating the weaknesses of both. An interactive model of text processing and comprehension holds that we process information in both ways simultaneously--from the top down and the bottom up. One danger here is that we make a 'Rube Goldberg' model for the sake of theory when what actually need to proceed in the classroom is more like 'Occam's Razor' instead.

In current ELT methodology what is often seen as one version of the mainstream top-down models (but which is not specific to ELT) might actually be more accurately called a "compensatory, interactive model" (first advanced as a complete model by Stanovich, 1980). Basically what "compensatory, interactive" means is that top-down cognition (e.g., active, outwardly directed use of pragmatic expectations and semantic knowledge toward the text) is used to fill in the gaps that FL learners inevitably experience as they incompletely and erroneously process L2 text. This in theory is possible because the interactive processing of parallel codes (e.g., phonological/orthographic, lexical, syntactical, semantic, pragmatic) supposedly allows these different levels to compensate for loss of information at any one level.

According to advocates of this theory, teachers are doing the right thing by placing an emphasis on a more efficient use of top-down skills (e.g., pre-reading or pre-listening activities to call up prior knowledge, skimming, reading for gist, etc.) in order to boost students' performance with the bottom-up skills (e.g., decoding letter for sounds or recognizing sounds as parts of meaningful chunks of language).

None of the above!

Unfortunately, as you might expect if you are a classroom teacher used to trying to make your way through the thicket of applied and pure research, there are problems even with this seemingly reasonable compensatory, interactive model. An outstanding example of pure (not applied) research that certainly complicates and weakens--if not entirely undercuts--both the top-down and the compensatory, interactive models can be found in Townsend and Bever (1991), who "test and disprove the common assumption that pragmatic probability facilitates the processing of lower linguistic levels" (p. 49). In their research, what the application of pragmatic considerations in comprehension tasks seems to do is actually overwhelm working memory and crowd out bottom-up processes altogether. There is no smooth, complete, complementary synthesis of parallel processes to compensate for incomplete bottom-up processing. The implications for the language and literacy classrooms are enormous: encourage top-down processes in order to boost comprehension and we may distract our students from fluently processing the text.

This is not to say, however, that top-down approaches have no use whatsoever. Such approaches could prove invaluable in getting students interested in and focused on a language learning or literacy task, since reading in a foreign language is not a matter of course because of a lack of the foreign language in the native culture. Also, getting and holding students・attention so that attention can be directed to focussed reading and listening tasks is arguably necessary for effective instruction. Managing a classroom discussion (for both language practice and for thinking while using the L2) would be impossible without such approaches. Moreover, in face-to-face encounters or when watching a movie or TV program, for example, not all the information crucial to understanding is locked up in the linguistic encoding.

Schema theoretic approaches in the classroom may well boost comprehension (but so, too, does translation and interpretation into L1), but they may not boost linguistic processing the way we have been led to hope. In fact, they may well distract students from actually trying to fluently read or listen to the whole text.

Where do top-down and bottom-up meet? 

While some of the above discussion may work to undercut the mainstream acceptance of top-down, schema-theoretic approaches to FLT and FL literacy, I would argue that there still is room for them in the classroom. In current discussions of ELT methodology, there is much talk of lexical approaches. Maybe it is at the level of words and lexical phrases that meaning can be said to be perceived in both top-down and bottom-up fashions.

A lexical approach can be said to be top-down in the sense that a person's prior word, conceptual, and world knowledge come into use when attention is directed toward a text and it is processed with comprehension. However, it could also be viewed as bottom-up in the sense that a word is the smallest unit of meaning in a text to occur on its own. In other words, the processing and automatic recognition of separate lexical items might act as a bridge between the lower levels of processing (e.g., phonemes/graphemes, syllables, major and minor spelling patterns, sight words) and the upper ones (e.g., syntax, text semantics, pragmatic considerations).

In actual classroom practice, teachers who want to reconsider what improved vocabulary instruction might include would do well to heed Paran's (1996) call for activities as seemingly mundane, passe, and uncommunicative as timed word recognition exercises, a classic example of what has been called "unenlightened" bottom-up language practice. Still, they need not be uncommunicative if situated properly in a communicative classroom environment.


Much of the this discussion might have seemed abstract--abstracted from the reality of any classroom. My goal was to take an abstract concept--one that many teachers make assumptions about, assumptions which have huge implications for the classroom--and examine it more closely in order to see the implications, problems and promise that it holds. My point is that while some abstraction belongs in the classroom, the theories and concepts that are out there in the world of academia do not simply apply to our classrooms because applied linguists say they should or do. As teachers we really do have to work out such theories and concepts for ourselves. We are the all important connection that allows theories and concepts to serve a pragmatic purpose: improved instruction.


Carrell, P. and Eisterhold, J. (1983). Schema theory and ESL reading pedagogy. TESOL  
   Quarterly, 17
(4), 553-73.

Cook, G. (1997). Key concepts in ELT: Schemas. ELT Journal, 51(1), 86.

Goodman, K.S. (1967). Reading: A psycholinguistic guessing game. Journal of the Reading
   Specialist, 4,

Jannuzi, C. (1994). Team teaching the reading class. In M. Wada & A. Cominos (Eds.), Studies in
   team teaching
(pp. 119-131). Tokyo: Kenkyusha.

Oller, J.W. (1995). Adding abstract to formal and content schemata: Results of recent work in
   Peircean semiotics. Applied Linguistics, 16(3), 273-306.

Paran, A. (1996). Reading in EFL: Facts and fictions. ELT Journal, 50(1), 25-34.

Rumelhart, D.E. (1994). Toward an interactive model of reading. In R. Ruddell, M. Ruddell, & H.
   Singer (Eds.), Theoretical models and processes of reading (4th ed.) (pp. 864-894). Newark,
   Del.: International Reading Association.

Stanovich, K.E. (1980). Toward an interactive compensatory model of individual differences in 
   the development of reading fluency. Reading Research Quarterly, 16, 32-71.

Towsend, D.J. & Bever, T.G. (1991). The use of higher-level constraints in monitoring for a
   change in speaker demonstrates functionally distinct levels of representation in discourse
   comprehension. Language and Cognitive Processes, 6(1), 49-77.

Widdowson, H.G. (1983). Learning purpose and language. Oxford: Oxford University Press.

Creative Commons Licence 3.0 and its provisions apply.

1 comment:

  1. Thank you. I will share with your creative commons 3.0 licence. Good info!