Language Teaching Research:

Language Teaching Research:
In the Literature, but not always in the Classroom









Abstract
This paper examines applied linguistic research done in four areas: 1) behaviorism, cognitivism, contrastive analysis, error analysis, 2) purpose and expectation, 3) Krashen's natural order of acquisition, and 4) reading for pleasure and language acquisition.

The paper then examines a conundrum: the compelling findings offered in the research, findings which have practical application in the classroom, and the fact that these findings have, in the real world, rarely been adopted by teachers or school districts.

The paper posits possible reasons for this "foot dragging" by educators, and closes by citing the ongoing increase of second language research and encouraging educators to apply the more relevant research to their classroom methodology.








Background

In 1957 Noam Chomsky changed the face of linguistics when he published Syntactic Structures, but it wasn't until the 1960s (and more prominently in the 1970s) when applied linguistic research started to change the prevailing views of classroom methodology; notably, the behaviorists, following B.F. Skinner, came under attack. Behaviorists, supporters of audiolingualism, embraced Skinner's Verbal Behavior, which suggested that language was a form of conditioned behavior. According to Skinnerian behaviorists, language could be acquired like a habit by a program of stimulus, response, and reinforcement. Audiolingualism used consistent and unending drilling of the students followed by positive or negative reinforcement.

Chomsky cited the fact that children spontaneously utter sentences they have never heard, much less practiced, before.
No, language isn't just a habit, there must be a cognitive factor, a language acquisition device in the brain.

Later, linguists such as Stephen Krashen and the late Tracy Terrell conducted classroom research on input focus; like a child learning its native language (L1), the listening input is key, they said. Give students massive amounts of comprehensible input, and have them focus on meaning, not form, like a child listening to its parents.

Most recently, Steven Pinker's controversial The Language Instinct, has contended that language is an instinct, as automatic to humans as flying is to geese. And although researchers have not found a "language gene" which programs language, they well might in the near future.

As research piles up and moves in a linear direction, there is still no definitive answer which can explain the process of either first or second language acquisition. Controversy still abounds and while a majority of applied linguists might agree that habit formation is a theoretically unacceptable explanation for language acquisition, many still feel that a traditional audiolingual classroom may help some students acquire a second language. Indeed, with all the research and counterclaims, we might one day discover that language acquisition is a combination of multiple forces, psychological, physiological, and environmental, which interact in different ways with different students: how else can we explain why language students show a wide range of outcomes across the spectrum of classroom methodologies?

If the research has progressed linearly, each new wave adding a little to the corpus, classroom pedagogy seems to have moved in a rather small circle, especially in the developing world.

Students tend to practice grammar rules, memorize vocabularly lists, and translate sentence level discourse in order to traverse the long journey to second language fluency. In many of these classrooms, the compiled research of the last forty years remains untapped; many of the classes I have seen use the same techniques today that were used in 1961 when I started learning Spanish at Walt Whitman Junior High School in New York.

Although the behaviorists are considered passe, much of the research undertaken from their era up to the present has potentially salubrious implications for language teachers and learners. Yet, many language curricula are designed oblivious to this fact. In other words, there is a lot of research available which can improve the language acquisition environments of students, but much of it is not utilized.

Behaviorism, Cognitivism, Contrastive Analysis, Error Analysis

Contrastive analysis, a comparative analysis of two languages, their similarities and their differences, was thought by many in the 1940s, 50s, and 60s to be a useful predictor of where second language students would likely encounter problems in learning a second language. It stood to reason that if certain elements of a second language differed greatly from the student's native language, that student would likely encounter difficulties. Lado (1957) stated that his book Linguistics across Cultures,
"...rests on the assumption that we can predict and describe patterns that will cause difficulty in learning and those that will not cause difficulty." Oller (1971) echoed, speaking of CA as "...a device for predicting points of difficulty and some of the errors that learners will make." It seemed a plausible assumption, but as research piled up it showed that contrastive analysis predicted students errors sometimes and didn't at others. Although Politzer (1968) found some predictive elements in CA, Wardhaugh (1970), Selinker (1972), Richards (1974), Taylor (1974), and Tran (1975) all identified errors which were of a non-contrastive origin, thereby lowering expecatations of CA as a predictor of language error. This encouraged the critics of behaviorism, the cognitivists, since the theory of transfer was often associated with the behaviorists. A counter-theory, although not strictly associated with the cognitivists, certainly more compatible to them, was error analysis, which treated second language errors as similar to errors encountered in first language acquisition, or what the linguists referred to as "developmental errors." These errors divided into three sub-categories: overgeneralization, incomplete rule application, and the hypothesizing of false concepts and, according to Richards (1971), reflected a learner's competence at a certain stage and thereby differed from learner to learner.

The debate raged on, notably through the 1970s, and contrastive analysis took quite a pummeling. Error analysis did not escape unscathed, suffering notably from Schacter's (1974) study which showed error analysis misdiagnosing student learning problems due to their "avoidance" of certain difficult L2 elements.

The result today is that both contrastive analysis and error analysis are rarely used in identifying L2 learner problem areas and "customizing" curricula for homogeneous L2 groups, one exception being Pierson (1982) who used error analysis as a basis for developing curriculum items for Cantonese speakers. I contend that the research done in both CA and EA can be utilized in curriculum development to the benefit of the L2 learner.

At Tunghai University in Taiwan, informal error analyses were conducted on student writing continuously through the mid 1980s. The taxonomy didn't encompass the traditional error analysis categories such as verb tense error and possessive marker error. According to error analysis theory the errors identified should represent "developmental errors," errors similar to those made in L1, but differing from student to student due to the students' differing developmental levels. Remarkably, the corpus collected showed high rates of overlap: in one class 50% to 60% of the identified errors were common to the whole class. Although flying in the face of EA which would expect differing errors depending on developmental stage, this 50% to 60% overlap itself could have served as a valuable pedagogical tool. That was a large body of language that none of the students had acquired: if they didn't know it, at some point it should have been taught. If we can positively identify large chunks of common non-acquired language in homogeneous L2 learners, we can "customize" their curriculum to ensure they are dealing with unmastered language items.

Furthermore, even though, according to Richards (1971), research shows that contrastive analysis may be most predictive at the level of phonology and least predictive at the syntactic level, many of the common mistakes by the Taiwanese students were syntactic errors found in written work:


*Because I enjoy eating; therefore I am fat
Yinwei wo chrde dwo, swoyi wo hen pang
*Although I washed the clothes, but I still didn't get them clean Sweiran wo yijing syi gwo, keshr wode yifu haishr bu ganjing
*I like to sing songs
Wo syihwan changge


The examples here are pattern sentences, structurally sound in Mandarin, but awkward when directly translated into English -- therefore, but, and songs are all obligatory in Chinese, but render the English sentences awkward if not ungrammatical. The superimposition of the Mandarin structure onto the English strongly hints of L1 interference here. In addition, the persistent and stubborn problem of verb tense and aspect errors in Taiwanese and Macanese students could possibly be predicted by contrastive analysis: the Chinese and English verb systems differ greatly.

The debate over contrastive anaysis and error analysis has virtually disappeared in the last ten years. Most researchers agree that contrastive analysis and error analysis alone can't predict or account for the myriad errors encountered in learner English. But Candlin in his preface to Richards (1974) states "...far from their being a fundamental opposition between Contrastive Analysis and Error Analysis, the latter serves as an important source of corroboration to contrastive linguistic analyses in their claims for predictability of error." And Norrish (1983) adds, "Although this 'strong predictive' claim for contrastive analysis can hardly be sustained any longer, it is certainly true to say that analysis has a useful explanatory role. That is, it can still be said to explain certain errors and mistakes."

Most young language teachers today are unaware of the CA-EA controversy and the research it engendered. Many of the original "overblown" claims of the two research methodologies have been debunked, but even a superficial grasp of these two theoretical constructs of language learning can be applied in the classroom to identify and explain some of the problems the students are experiencing.

Purpose, Expectation -- An Alternative to Traditional Comprehension

The cognitivist Noam Chomsky implied more than once that his theories of language acquisition had nothing to do with adult language learning, and language teaching has never adopted a set methodology based on Chomsky's work. But, as cognitivism gained credibility relative to behaviorism, it was natural that researchers should turn their attentions to the role of thought processes in language acquisition: this can cover broad areas including psychology, physiology, areas most often associated with psycholinguistics, and I hardly want to bore the layperson by getting into that. However, I would like to address the thought processes of the L2 learner, what's going on when the student is accessing the target language.

L2 comprehension skills are most often classified as type 1 or type 2. According to Harmer (1983), type 1 skills are "...those operations that students perform on a text when they tackle it for the first time." This basically includes the two
traditional tasks of reading to extract specific information and reading for general understanding. Type 2 skills are, quoting Harmer (1983) again "...those that are subsequently used when studying reading or listening material and they involve detailed comprehension of the text (after the students have performed type 1 skills)." These type 2 tasks include inferring opinion and attitude, deducing meaning from context, and recognizing function and discourse patterns and markers.

Traditionally, learners have encountered the two type 1 skills involved with extracting specific information and getting the "general picture." This is especially true for young learners and those at the lower levels. These skills have been so emphasized that reading strategies have been developed, skimming for general understanding and scanning for specific information, to deal with them. The more practitioners have examined learners' motivations and thought processes, the more these two traditional skills have been questioned.

In many classrooms around the world, students are given either written or listening material and asked to regurgitate specific information or the "main ideas" of the material. The student may or may not be interested in the topic and, if interested, may wish to pick up some general understanding or some specific information. In real life, people often don't approach reading or listening this way.

Let's take an example. You walk by a newsstand and see the headline on a newspaper, "Hong Kong Stock Prices Surge Upward." You own Hong Kong stocks, so you buy the paper and start scanning for specific information (your stocks). Or you overhear some friends talking about your boss, who was involved in a car accident last night. You quickly move over to where they are talking and inquire for more details about your boss's misfortune. Nobody "assigned" the listening or reading to you, initial exposure created a desire to read (to see if your stocks rose) and to listen (to find out the condition of your boss). This exposure created what Harmer (1983) refers to as a purpose to listen or a purpose to read. Quoting Harmer, "In real life people read or listen to language because they want to and because they have a purpose for doing so.... In real life, therefore, readers and listeners have a purpose that is more fundamental than the typical language learners'" Other researchers (see Coe, 1978) have also addressed the issues of interest, expectation and purpose, and proposed an addition to the type 1 questions, namely, predictive skills.

Going back to our examples, when the reader saw the Hong Kong stock headline or heard the bad news about his boss, he started subconsciously to predict, perhaps how much his stocks had gained or how seriously his boss was hurt. These predictions are initially the result of the expectations the initial exposure has raised in him. "Stock Prices Surge" may create an expectation of big one-day profits for him. However, as he reads further his predictions will change as he receives more information. This is what Harmer (1978) refers to as interaction with context, and what naturally happens when one reads or listens in one's native language. Thus, this type of exercise becomes more realistic and natural than simply assigning a random reading and asking students to answer general and specific questions.

How is this type of activity set up in the classroom? First you solicit a list of topic areas your students are interested in: people usually don't listen or read when they are not motivated. Then you spend time generating interest in the lead-in part of the exercise. Interest leads to prediction, prediction based on expectation, and this in turn creates a purpose to read, to confirm or deny predictions and expectations.

Let us walk through a practical example. My students have a keen interest in Canto Pop, the local pop music of Hong Kong and Macau. I assign a reading about one of the top singers in Hong Kong headlined, "Jacky Cheung Finally Marries at 35." The headline is enough to generate interest, especially among female students, but a continued lead-in discussion about the article is necessary. Rossner, et al (1979a) would set up a three-box chart, one box headed "things you know about Jacky Cheung," the second headed "things you are not sure of," the third headed "things you would like to know." As the boxes fill up, the students will make subconscious predictions based on the expectations that the headline and the lead-in have created. For example, under "things you are not sure of," some students may have written "who the bride is." This will stimulate certain predictions, e.g. it must be _____________, if not her who else? If it's someone new what kind of person is she? how old is she? This creates a purpose to read -- to find out the answers to those questions and, hence, confirm or deny the predictions and expectations.

Type 1 skills which stimulate prediction and which are designed to confirm or deny a student's expectations are both interesting and realistic, interesting in that they present content the students want to access, and realistic in that they map the subconscious cognitive processes involved in reading and listening to native language. Sadly many schools and language programs fail to tap into this valuable resource.

The Krashen Era and the Natural Order of Acquisition

In the 1970s, an applied linguist from the University of Southern California started raising eyebrows by, like Chomsky, making some controversial assertions about language. Unlike Chomsky, though, Stephen Krashen's linguistic theories had a direct relationship to language learning and acquisition, thereby bringing them to the attention of language teachers around the world.

Steve Krashen, along with the late Tracy Terrell, is the father of "input theory," which stresses massive amounts of passive language or what Krashen (1979) refers to as i+1, language that is just a little beyond comprehension. Krashen contends that through context and extralinguistic information, like a mother talking to her child, learners will climb to the next level and then repeat the process. Terrell (1982) seconded Krashen's findings and coined the term, "natural approach"; that is, an approach that is like a child learning its L1 in the home.

The central core of Krashen's work involves his nine hypotheses, the main two being the input hypothesis and the acquisition-learning distinction hypothesis. But it is a lesser emphasized hypothesis I would like to address here.

One of the more interesting discoveries in language acquisition has been the finding that acquisition of grammatical structures proceeds in predictable order. Krashen (1987) cites research that shows both learners of L1 and of L2 tend to acquire grammatical elements in a fairly predictable order. Brown's (1973) study of children learning English as an L1 shows a distinct order with -ing progressive and plural being early acquired and regular past and possessive s being late acquired. Later Dulay and Burt (1974, 1975) showed that children studying English as a second language also showed a natural order of acquisition, no matter what their native language. Both the Brown study and the Dulay and Burt studies have been replicated (see de Villiers & de Villiers, 1973; Kessler & Idar, 1977; Fabris, 1978; Makino, 1980). In addition, Bruce (1979) and van Naerssen (1981) confirm a natural order of acquisition for other languages.

Does this mean that research should be applied to sequence grammar teaching according to the natural order of acquisition? Not really, according to Krashen. While arguing that grammatical sequencing will not lead to fluent acquisition of language, Krashen (1987) does hedge by saying, "...that we should present rules one at a time in some order when the goal is conscious learning..." Many orders of grammatical sequencing have been suggested over the years, including frequency of occurrence, grammatical simplicity, and utility, with most traditional grammar texts attempting to order along the lines of grammatical simplicity. But does this make sense when research has shown that the irregular past tense is most often acquired before the regular past tense? Partly. Krashen suggests that simplicity does play a role, but natural order does also. Krashen (1987) contends that we can only teach what is learnable and what is portable (that which can be carried around in the students' heads). The natural order studies can provide us with at least some of the information we need in that rules to be learned should meet three criteria: learnable, portable, not yet acquired. The sequencing issue then revolves around which items meeting all these criteria should be presented first. Granted, this is not a "magic bullet" for sequencing grammatical items in the classroom, but it contributes by limiting the set of items that must be sequenced.

Many textbooks are sequenced according to grammatical simplicity which is part of the equation, but "perceived" grammatical simplicity may vary from country to country; that is, a materials developer in Malaysia may arrange English syntactical items in a far different simple to complex order than a materials developer in Hungary. Here is where an understanding of the "natural order of acquisition" may have practical application in the classroom. However, it has yet to make any meaningful inroads in most traditional curricula.

Reading for Pleasure and Language Acquisition

Perhaps the most ironic body of research to be stymied at the classroom door is the research on reading for pleasure or, what some refer to as extensive reading or free voluntary reading. Ironic in that few need research to corroborate that reading is good for you. Krashen has also been very involved in this area, compiling a research review, The Power of Reading, which cites over 200 studies from the 1890s to the present. The conclusions drawn from that highly readable research review are noteworthy. To wit, reading habits and literacy are postively correlated. Reading for pleasure is more effective in improving comprehension than traditional skill-based reading instruction. Reading for pleasure, even comic books, is the most effective way to increase vocabulary. Extensive reading is nearly always superior to direct instruction on tests of reading comprehension, vocabulary, writing, and grammar, and reading as a leisure activity is the best predictor of comprehension, vocabulary, and reading speed.

Schackne (1994) and Sims (1996) looked at several EFL classrooms in Taiwan and Macau, and came to the same conclusion: there is strong evidence that reading for pleasure can promote substantial language acquisition both on a short and long term basis.

A technique that is effective obviously has many applications. Here is an activity that is not only student centered, but an activity a student can pursue independently and be relatively sure of positive results: that would make it not only effective, but cheap and convenient as well. Also, it is an activity that supplies teachers with an effective weapon, a trump card to use when confronted with stagnant, ineffective curricula.

Extensive reading is uniformly accepted among langauge teachers: there is virtual consensus, backed by quantitative evidence, among professionals and lay people that extensive reading has a salubrious effect on language development. And although the number of extensive reading components in schools has been increasing worldwide, it is still, considering its effectiveness, underutilized. Why? Three possible reasons come to mind.

First, expense. A fairly healthy library is needed to implement this program, and the cost can be daunting for many schools, especially those in the developing world. Second, extensive reading is a student centered activity which takes control out of the teacher's hands and puts it into the student's hands. While this frees teachers for other activities, it is also a radical, and sometimes threatening, change for traditional school systems. Third, it simply appears too simple and too unstructured: students lounging around reading comic books, magazines, and novels doesn't fit many pre-conceived notions of what a language program should be about. Many traditional educators balk at this approach even though it is meant to be a relatively small part of an overall language program.

Conclusion

On the TESL-L bulletin board, on "the net," a netter recently posted a message promoting the "mishmash" school of language teaching: try several different methods and see what works. An unfortunate choice of words because mishmash implies a random, disorganized approach. Years ago, teflers referred to it as the "golden nugget" approach. Students learn in many different ways; if you give them a wide range of options, they will selectively choose those which are most suitable to them: picking up the "nuggets" that appeal to them as they walk through the gold mine. As research in applied linguistics accumulates, this approach becomes more feasible because more options are continuously coming on stream. However, the number of schools and programs which offers the student a choice is still quite limited.