The Language Resource is a monthly publication of the National Capital Language Resource Center, http://www.nclrc.org.

Should We Teach Grammar? (part I)
Stephen Krashen
Professor of Education, University of Southern California

Dr. Krashen is the nation's leaders in the fields of bilingual education, neurolinguistics, second language acquisition and
literacy. He has authored over 250 articles and books and has received numerous awards. Dr. Krashen is also an indispensable
champion of foreign language education in the United States.

At one time, it was assumed that the only way of developing grammatical competence in a second language was through direct teaching of grammar. We
all held the "skill-building" position: We learn language by first learning the rules consciously, then practicing them in output exercises, and we
fine-tune our knowledge of rules by getting our errors corrected.

This axiom has been demoted to the level of hypothesis: It has been argued that we develop competence in second languages in another way. We
acquire the grammatical rules of a language by understanding input containing these rules. Our attention is not on consciously learning the rules but on
understanding the message, and we subconsciously absorb the rules the same way children absorb the rules of their first language. Conscious knowledge
of grammar has a limited function: It is used to edit or monitor our second language production. We can use the conscious grammar to make small
grammatical repairs when we have time, when we are thinking about correctness, and when we know the rules. Conscious knowledge is thus not
useless, but it has a limited function.

The evidence for the "comprehension" or "input" hypothesis includes studies showing that students in comprehension-based second language classes
consistently outperform those in traditional classes, at both the beginning and intermediate levels, and includes studies showing the powerful impact of
recreational reading (Krashen, 2003).

There is also strong indirect evidence supporting the comprehension hypothesis. The grammatical system of any language is far too complex to be
consciously learned, and many people develop high levels of competence without formal instruction. Quite often, those who have not reached the highest
levels of competence in second languages, despite what seems to be a great deal of exposure, have not been readers.

In recent years, a number of studies have been published contesting the input hypothesis. These studies typically show that after we provide students with
direct instruction, they improve in the use of the rules taught, and show more improvement than comparison groups. But in all studies,

1.Students were experienced "learners." They expected direct teaching and were good at it;
2.Students showed only modest improvement in the rules taught, even after a considerable amount of practice, and even these modest gains were
typically short-lived;
3.Tests used were not communicative. In all cases, students were focused on rules during the test, had plenty of time to access the rules, and were
tested on rules they had just covered in class (Krashen, 2003).
4.Comparison groups in these studies did not receive quality comprehensible input.

The results of these studies are thus fully consistent with the input hypothesis.

Some grammar proponents argue that we must teach grammar because accuracy is so important. I agree that accuracy is important. We are discussing
how to achieve it. The comprehension hypothesis claims that comprehensible input is the best way to develop grammatical accuracy.

Note: I do not think that grammar teaching should be at the core of curriculum, but there are good reasons for including it. First, grammar teaching can be
an excellent introduction to the study of linguistics, which has obvious value, e.g. the study of universals, language change, and dialects. Second, even
with massive reading, complete acquisition of the conventions of writing may not take place; even very well-read people may have gaps in their first
language (e.g. the it's/its distinction). These gaps are typically small and rarely interfere with the clarity of the message. Conscious knowledge of
grammar rules can help fill at least some of these gaps, and can be used in the editing stage of the composing process, after ideas are on the page.

Krashen, S. 2003. Explorations in Language Acquisition and Language Use: The Taipei Lectures. Portsmouth, NH: Heinemann.

************************

Should We Teach Grammar? (part II)

Dr. Richard Robin is Associate Professor and Chair of the German and Slavic Department at The George Washington University. He has coauthored a number of books including Russian Listening Comprehension, Political Russian and Russian for Russians.

Grammar is like sex. It's a basic facet of the life for both language teacher and language learner. It casts a huge shadow over both teacher and learner, often dominating the relationship. And it has become taboo.

In determining how much grammar to teach explicitly, how much accuracy to demand, and at what stages and in what contexts, we should take our inspiration from the decisions we make in our private lives: one size does not fit all. Teachers must juggle three important elements: (1) the goal of instruction / time for instruction, (2) the structure of the target language, and (3) the style of the learner. These factors rarely line up neatly, which is why after hundreds of studies and dozens of hypotheses, we're still debating.

Goal of instruction / time for instruction. Is the end goal in speaking ACTFL Intermediate (sentence-level creative communication, transactional survivability)? If you teach in a college like mine, which has a only one-year language requirement amounting to less than 100 hours of face-to-face instruction, you should consider a special section for those who do not plan to continue and dump some of the grammar. If, on the other hand, your audience is likely to stay put for a few hundred hours of instruction, and higher proficiency levels are within sight, explicit instruction in grammar not only results in some fleeting accuracy in largely memorized role-plays; it creates a mindset for useful processing of input years later. Overt grammar instruction and practice, coupled with explicit instruction in other language learning strategies, helps to create expert learners who go beyond the range of ACTFL Intermediate on to the holy grail of Advanced.

Structure of the target language. The profession is in general agreement that grammar should not drive the curriculum. But for a morphologically complex language such as Russian, explicit grammar instruction occupies the front passenger seat. Not surprisingly, longitudinal studies based on study-abroad data at the American Council of Teachers of Russian show a positive correlation between pre-country test scores of explicit grammar and improvements on post-country oral proficiency interviews. In languages such as Japanese and Chinese, where input through reading is necessarily reduced, explicit instruction may be required to fill the gap. In Chinese, that practice will of course be concentrated on prosodic accuracy.

Learner style and background. Ten-year olds don't want the big picture, but college students are more likely to demand an analytical treatment of the morphological system of the target language. It behooves teachers to play to the best styles of each learner. A few charts and basic manipulation drills on each point does take away from valuable time that might otherwise be spent in communicative activities. But this kind of basic structural handholding appeals to many learners, especially older ones, who find more chaos than solace in a less structured environment.

I am not arguing for a return to the grammar and drill driven curriculum of decades past. Rather I am suggesting that a number of factors help determine what grammar we teach, how we teach it, and to which learners of which languages. If our goal is ultimately to produce Advanced speakers of the target language in a relatively short period of time - one hundred to several hundred contact hours, we need our students to become not only proficient speakers, but expert learners as well. Explicit grammatical instruction and practice, coupled with instruction in strategies is a path to the kind of language learning awareness that such a daunting task requires.