Krashen and Terrell’s “Natural Approach”
by Ken Romeo
The Acquisition-Learning Hypothesis
The Natural Order Hypothesis
The Monitor Hypothesis
The Input Hypothesis
The Affective Filter Hypothesis
Curriculum Design

Return to Language and Literacy  ->  LAU top


        The influence of Stephen Krashen on language education research and practice is undeniable.  First introduced over 20 years ago, his theories are still debated today.  In 1983, he published The Natural Approach with Tracy Terrell, which combined a comprehensive second language acquisition theory with a curriculum for language classrooms.  The influence of Natural Approach can be seen especially in current EFL textbooks and teachers resource books such as The Lexical Approach (Lewis, 1993).  Krashen’s theories on second language acquisition have also had a huge impact on education in the state of California, starting in 1981 with his contribution to Schooling and language minority students: A theoretical framework by the California State Department of Education (Krashen 1981).  Today his influence can be seen most prominently in the debate about bilingual education and perhaps less explicitly in language education policy:  The BCLAD/CLAD teacher assessment tests define the pedagogical factors affecting first and second language development in exactly the same terms used in Krashen’s Monitor Model (California Commission on Teacher Credentialing, 1998).
        As advertised, The Natural Approach is very appealing – who wouldn’t want to learn a language the natural way, and what language teacher doesn’t think about what kind of input to provide for students.  However, upon closer examination of Krashen’s hypotheses and Terrell’s methods, they fail to provide the goods for a workable system.  In fact, within the covers of “The Natural Approach”, the weaknesses that other authors criticize can be seen playing themselves out into proof of the failure of Krashen’s model.  In addition to reviewing what other authors have written about Krashen’s hypotheses, I will attempt to directly address what I consider to be some of the implications for ES/FL teaching today by drawing on my own experience in the classroom as a teacher and a student of language.  Rather than use Krashen’s own label, which is to call his ideas simply “second language acquisition theory”, I will adopt McLaughlin’s terminology (1987) and refer to them collectively as “the Monitor Model”.  This is distinct from “the Monitor Hypothesis”, which is the fourth of Krashen’s five hypotheses.

The Acquisition-Learning Hypothesis

        First is the Acquisition-Learning Hypothesis, which makes a distinction between “acquisition,” which he defines as developing competence by using language for “real communication” and “learning.” which he defines as “knowing about” or “formal knowledge” of a language (p.26).  This hypothesis is presented largely as common sense: Krashen only draws on only one set of references from Roger Brown in the early 1970’s.  He claims that Brown’s research on first language acquisition showed that parents tend to correct the content of children’s speech rather than their grammar.  He compares it with several other authors’ distinction of “implicit” and “explicit” learning but simply informs the reader that evidence will be presented later.
        Gregg (1984) first notes that Krashen’s use of the Language Acquisition Device (LAD) gives it a much wider scope of operation than even Chomsky himself.  He intended it simply as a construct to describe the child’s initial state, which would therefore mean that it cannot apply to adult learners.  Drawing on his own experience of learning Japanese, Gregg contends that Krashen’s dogmatic insistence that “learning” can never become “acquisition” is quickly refuted by the experience of anyone who has internalized some of the grammar they have consciously memorized.  However, although it is not explicitly stated, Krashen’s emphasis seems to be that classroom learning does not lead to fluent, native-like speech.  Gregg’s account that his memorization of a verb conjugation chart was “error-free after a couple of days”(p.81) seems to go against this spirit.  The reader is left to speculate whether his proficiency in Japanese at the time was sufficient enough for him to engage in error-free conversations with the verbs from his chart.
        McLaughlin (1987) begins his critique by pointing out that Krashen never adequately defines “acquisition”, “learning”, “conscious” and “subconscious”, and that without such clarification, it is very difficult to independently determine whether subjects are “learning” or “acquiring” language.  This is perhaps the first area that needs to be explained in attempting to utilize the Natural Approach.  If the classroom situation is hopeless for attaining proficiency, then it is probably best not to start.  As we will see in an analysis of the specific methods in the book, any attempt to recreate an environment suitable for “acquisition” is bound to be problematic.
        Krashen’s conscious/unconscious learning distinction appeals to students and teachers in monolingual countries immediately.  In societies where there are few bilinguals, like the United States, many people have struggled to learn a foreign language at school, often unsuccessfully.  They see people who live in other countries as just having “picked up” their second language naturally in childhood.  The effort spent in studying and doing homework seems pointless when contrasted with the apparent ease that “natural” acquisition presents.  This feeling is not lost on teachers: without a theoretical basis for the methods, given any perceived slow progress of their students, they would feel that they have no choice but to be open to any new ideas
        Taking a broad interpretation of this hypothesis, the main intent seems to be to convey how grammar study (learning) is less effective than simple exposure (acquisition).  This is something that very few researchers seem to doubt, and recent findings in the analysis of right hemisphere trauma indicate a clear separation of the facilities for interpreting context-independent sentences from context-dependent utterances (Paradis, 1998).  However, when called upon to clarify, Krashen takes the somewhat less defensible position that the two are completely unrelated and that grammar study has no place in language learning (Krashen 1993a, 1993b).  As several authors have shown (Gregg 1984, McLaughlin 1987, and Lightbown & Pienemann 1993, for a direct counter-argument to Krashen 1993a) there are countless examples of how grammar study can be of great benefit to students learning by some sort of communicative method.

The Natural Order Hypothesis

        The second hypothesis is simply that grammatical structures are learned in a predictable order.  Once again this is based on first language acquisition research done by Roger Brown, as well as that of Jill and Peter de Villiers.  These studies found striking similarities in the order in which children acquired certain grammatical morphemes.  Krashen cites a series of studies by Dulay and Burt which show that a group of Spanish speaking and a group of Chinese speaking children learning English as a second language also exhibited a “natural” order for grammatical morphemes which did not differ between the two groups.  A rather lengthy end-note directs readers to further research in first and second language acquisition, but somewhat undercuts the basic hypothesis by showing limitations to the concept of an order of acquisition.
        Gregg argues that Krashen has no basis for separating grammatical morphemes from, for example, phonology.  Although Krashen only briefly mentions the existence of other parallel “streams” of acquisition in The Natural Approach, their very existence rules out any order that might be used in instruction.  The basic idea of a simple linear order of acquisition is extremely unlikely, Gregg reminds us.  In addition, if there are individual differences then the hypothesis is not provable, falsifiable, and in the end, not useful.
        McLaughlin points out the methodological problems with Dulay and Burt’s 1974 study, and cites a study by Hakuta and Cancino (1977, cited in McLaughlin, 1987, p.32) which found that the complexity of a morpheme depended on the learner’s native language.  The difference between the experience of a speaker of a Germanic language studying English with that of an Asian language studying English is a clear indication of the relevance of this finding.
 The contradictions for planning curriculum are immediately evident.  Having just discredited grammar study in the Acquisition-Learning Hypothesis, Krashen suddenly proposes that second language learners should follow the “natural” order of acquisition for grammatical morphemes.  The teacher is first instructed to create a natural environment for the learner but then, in trying to create a curriculum, they are instructed to base it on grammar.  As described below in an analysis of the actual classroom methods presented in the Natural Approach, attempting to put these conflicting theories into practice is very problematic.
        When one examines this hypothesis in terms of comprehension and production, its insufficiencies become even more apparent.  Many of the studies of order of acquisition, especially those in first language acquisition, are based on production.  McLaughlin also points out that “correct usage” is not monolithic – even for grammatical morphemes, correct usage in one situation does not guarantee as correct usage in another (p.33).  In this sense, the term “acquisition” becomes very unclear, even when not applying Krashen’s definition.  Is a structure “acquired” when there are no mistakes in comprehension?  Or is it acquired when there is a certain level of accuracy in production?  First language acquisition is very closely linked to the cognitive development of infants, but second language learners have most of these facilities present, even as children.  Further, even if some weak form of natural order exists for any learners who are speakers of a given language, learning in a given environment, it is not clear that the order is the same for comprehension and production.  If these two orders differ, it is not clear how they would interact.

The Monitor Hypothesis

The role of conscious learning is defined in this somewhat negative hypothesis: The only role that such “learned” competence can have is an editor on what is produced.  Output is checked and repaired, after it has been produced, by the explicit knowledge the learner has gained through grammar study.  The implication is that the use of this Monitor should be discouraged and that production should be left up to some instinct that has been formed by “acquisition”.  Using the Monitor, speech is halting since it only can check what has been produced, but Monitor-free speech is much more instinctive and less contrived.  However, he later describes cases of using the Monitor efficiently (p. 32) to eliminate errors on “easy” rules.  This hypothesis presents very little in the way of supportive evidence:  Krashen cites several studies by Bialystok alone and with Frohlich as “confirming evidence” (p.31) and several of his own studies on the difficulty of confirming acquisition of grammar.
        Perhaps Krashen’s recognition of this factor was indeed a step forward – language learners and teachers everywhere know the feeling that the harder they try to make a correct sentence, the worse it comes out.  However, he seems to draw the lines around it a bit too closely.  Gregg points (p.84) out that by restricting monitor use to “learned” grammar and only in production, Krashen in effect makes the Acquisition-Learning Hypothesis and the Monitor Hypothesis contradictory.  Gregg also points out that the restricting learning to the role of editing production completely ignores comprehension (p.82).  Explicitly learned grammar can obviously play a crucial role in understanding speech.
        McLaughlin gives a thorough dissection of the hypothesis, showing that Krashen has never demonstrated the operation of the Monitor in his own or any other research.  Even the further qualification that it only works on discrete-point tests on one grammar rule at a time failed to produce evidence of operation.  Only one study (Seliger, cited on p.26) was able to find narrow conditions for its operation, and even there the conclusion was that it was not representative of the conscious knowledge of grammar.  He goes on to point out how difficult it is to determine if one is consciously employing a rule, and that such conscious editing actually interferes with performance.  But his most convincing argument is the existence of learners who have taught themselves a language with very little contact with native speakers.  These people are perhaps rare on the campuses of U.S. universities, but it is quite undeniable that they exist.
        The role that explicitly learned grammar and incidentally acquired exposure have in forming sentences is far from clear.  Watching intermediate students practice using recasts is certainly convincing evidence that something like the Monitor is at work: even without outside correction, they can eliminate the errors in a target sentence or expression of their own ideas after several tries.  However, psycholinguists have yet to determine just what goes into sentence processing and bilingual memory.  In a later paper (Krashen 1991), he tried to show that high school students, despite applying spelling rules they knew explicitly, performed worse than college students who did not remember such rules.  He failed to address not only the relevance of this study to the ability to communicate in a language, but also the possibility that whether they remembered the rules or not, the college students probably did know the rules consciously at some point, which again violates the Learning-Acquisition Hypothesis.

The Input Hypothesis

        Here Krashen explains how successful “acquisition” occurs:  by simply understanding input that is a little beyond the learner’s present “level” – he defined that present “level” as i and the ideal level of input as i +1.  In the development of oral fluency, unknown words and grammar are deduced through the use of context (both situational and discursive), rather than through direct instruction.  Krashen has several areas which he draws on for proof of the Input Hypothesis.  One is the speech that parents use when talking to children (caretaker speech), which he says is vital in first language acquisition (p.34).  He also illustrates how good teachers tune their speech to their students’ level, and how when talking to each other, second language learners adjust their speech in order to communicate.  This hypothesis is also supported by the fact that often the first second language utterances of adult learners are very similar to those of infants in their first language.  However it is the results of methods such as Asher’s Total Physical Response that provide the most convincing evidence.  This method was shown to be far superior to audiolingual, grammar-translation or other approaches, producing what Krashen calls “nearly five times the [normal] acquisition rate.”
        Gregg spends substantial time on this particular hypothesis, because, while it seems to be the core of the model, it is simply an uncontroversial observation with no process described and no proof provided.  He brings up the very salient point that perhaps practice does indeed also have something to do with second language acquisition, pointing out that monitoring could be used as a source of correct utterances (p. 87).  He also cites several studies that shed some doubt on the connection between caretaker speech in first language acquisition and simplified input in second language acquisition.
        McLaughlin also gives careful and thorough consideration to this part of Krashen’s model.  He addresses each of the ten lines of evidence that Krashen presents, arguing that it is not sufficient to simply say that certain phenomenon can be viewed from the perspective of the Input Hypothesis.  The concept of a learner’s “level” is extremely difficult to define, just as the idea of i +1 is (p.37).  Further, there are many structures such as passives and yes/no questions that cannot be learned through context.  Also, there is no evidence that a learner has to fully comprehend an utterance for it to aid in acquisition.  Some of the first words that children and second language learners produce are formulaic expressions that are not fully understood initially.  Finally McLaughlin points out that Krashen simply ignores other internal factors such as motivation and the importance of producing language for interaction.
        This hypothesis is perhaps the most appealing part of Krashen’s model for the language learner as well as the teacher.  He makes use of the gap between comprehension and production that everyone feels, enticing us with the hope of instant benefits if we just get the input tuned to the right level.  One of Krashen’s cleverest catch-alls is that other methods of teaching appear to work at times because they inadvertently provide this input.  But the disappointment is that he never gives any convincing idea as to how it works.  In the classroom a teacher can see when the students don’t understand and can simplify his or her speech to the point where they do.  Krashen would have the teacher think that this was all that is necessary, and it is just a matter of time before the students are able to express themselves freely.  However, Ellis (1992) points out that even as of his 1985 work (Krashen 1985), he still had not provided a single study that demonstrated the Input Hypothesis.  Over extended periods of time students do learn to understand more and even how to speak, but it often seems to take much longer than Krashen implies, indicating that there are perhaps many more factors involved.  More importantly, even given this beginning of i, and the goal of i + 1, indefinable as they are, the reader is given no indication of how to proceed.  As shown above the Natural Order Hypothesis holds no answers, especially as to how comprehension progresses.  In an indication of a direction that should be explored, Ellis’s exploratory study (ibid.) showed that it is the effort involved in attempting to understand input rather than simple comprehension that fuels acquisition.

The Affective Filter Hypothesis

        This concept receives the briefest treatment in “The Natural Approach”.  Krashen simply states that “attitudinal variables relate directly to language acquisition but not language learning.”  He cites several studies that examine the link between motivation and self-image, arguing that an “integrative” motivation (the learner want to “be like” the native speakers of a language) is necessary.  He postulates an “affective filter” that acts before the Language Acquisition Device and restricts the desire to seek input if the learner does not have such motivation.  Krashen also says that at puberty, this filter increases dramatically in strength.
        Gregg notes several problems with this hypothesis as well.  Among others, Krashen seems to indicate that perhaps the affective filter is associated with the emotional upheaval and hypersensitivity of puberty, but Gregg notes that this would indicate that the filter would slowly disappear in adulthood, which Krashen does not allow for (p.92).  He also remarks on several operational details, such as the fact that simply not being unmotivated would be the same as being highly motivated in this hypothesis – neither is the negative state of being unmotivated.  Also, he questions how this filter would selectively choose certain “parts of a language” to reject (p.94).
        McLaughlin argues much along the same lines as Gregg and points out that adolescents often acquire languages faster than younger, monitor-free children (p.29).  He concludes that while affective variables certainly play a critical role in acquisition, there is no need to theorize a filter like Krashen’s.
        Again, the teacher in the classroom is enticed by this hypothesis because of the obvious effects of self-confidence and motivation.  However, Krashen seems to imply that teaching children, who don’t have this filter, is somehow easier, since “given sufficient exposure, most children reach native-like levels of competence in second languages” (p.47).  This obviously completely ignores the demanding situations that face language minority children in the U.S. every day.  A simplification into a one page “hypothesis” gives teachers the idea that these problems are easily solved and fluency is just a matter of following this path.  As Gregg and McLaughlin point out, however, trying to put these ideas into practice, one quickly runs into problems.

Curriculum Design

        The educational implications of Krashen’s theories become more apparent in the remainder of the book, where he and Terrell lay out the specific methods that make use of the Monitor Model.  These ideas are based on Terrell’s earlier work (Terrell, 1977) but have been expanded into a full curriculum.  The authors qualify this collection somewhat by saying that teachers can use all or part of the Natural Approach, depending on how it fits into their classroom.
        This freedom, combined with the thoroughness of their curriculum, make the Natural Approach very attractive.  In fact, the guidelines they set out at the beginning– communication is the primary goal, comprehension preceding production, production simply emerge, acquisition activities are central, and the affective filter should be lowered (p. 58-60) – are without question, excellent guidelines for any language classroom.  The compilation of topics and situations (p.67-70) which make up their curriculum are a good, broad overview of many of the things that students who study by grammar translation or audiolingual methods do not get.  The list of suggested rules (p.74) is notable in its departure from previous methods with its insistence on target language input but its allowance for partial, non-grammatical or even L1 responses.
        Outside of these areas, application of the suggestions run into some difficulty.  Three general communicative goals of being able to express personal identification, experiences and opinions (p.73) are presented, but there is no theoretical background.  The Natural Approach contains ample guidance and resources for the beginner levels, with methods for introducing basic vocabulary and situations in a way that keeps students involved.  It also has very viable techniques for more advanced and self-confident classes who will be stimulated by the imaginative situational practice (starting on p.101).  However, teachers of the broad middle range of students who have gotten a grip on basic vocabulary but are still struggling with sentence and question production are left with conflicting advice.
        Once beyond one-word answers to questions, the Natural Approach ventures out onto thin ice by suggesting elicited productions.  These take the form of open-ended sentences, open dialogs and even prefabricated patterns (p.84).  These formats necessarily involve explicit use of grammar, which violates every hypothesis of the Monitor Model.  The authors write this off as training for optimal Monitor use (p.71, 142), despite Krashen’s promotion of “Monitor-free” production.  Even if a teacher were to set off in this direction and begin to introduce a “structure of the day” (p. 72), once again there is no theoretical basis for what to choose.  Perhaps the most glaring omission is the lack of any reference to the Natural Order Hypothesis, which as noted previously, contained no realistically usable information for designing curriculum.
        Judging from the emphasis on exposure in the Natural Approach and the pattern of Krashen’s later publications, which focused on the Input Hypothesis, the solution to curriculum problems seems to be massive listening.  However, as noted before, other than i + 1, there is no theoretical basis for overall curriculum design regarding comprehension.  Once again, the teacher is forced to rely on a somewhat dubious “order of acquisition”, which is based on production anyway.  Further, the link from exposure to production targets is tenuous at best.  Consider the dialog presented on p.87:

. . . to the question What is the man doing in this picture? the students may reply run.  The instructor expands the answer.  Yes, that’s right, he’s running.
The exchange is meant to illustrate how allowing for errors, while at the same time providing corrected input can help students in “acquisition”.  To the student, however, the information in the instructor’s response is completely contained in the word “Yes”.  Krashen makes no comment on how, even if it is comprehended, the extra information of “He’s running” enters the student’s production.  If simple exposure is the answer, then “that’s right” is more likely to be “acquired” given its proximity to the carrier of meaning “Yes”.
        This issue is the subject of extensive psycholinguistic research in sentence processing and bilingual lexical memory, and conclusive answers have not yet been found.  The length of the path from 1) understanding the above question to 2) giving a one-word answer, to 3) being able to give a full sentence answer, and then 4) being able to ask a similar question is quite unclear.  Especially if the teacher is to rely on input alone, it is very conceivable that the students could be working their way through the intermediate steps for quite some time.  Teachers would perhaps be better served by a less dogmatic approach that informed them of not only single steps, but what exactly has been found in current research.  This of course includes hypotheses and findings that have not been conclusively proven yet, but a more balanced approach than the present one would allow teachers to use their valuable experience in the classroom to make informed judgments about curriculum.  In attempting to teach a subject whose process is not clearly known, it seems obvious that a well-rounded awareness of the theoretical issues involved is necessary.  For this reason concurrent teacher education in language education is essential to insure the needs of all students are met.


        Krashen seemed to be on the right track with each of his hypotheses.  Anyone who has learned a language, and especially those who have seen the grammar-translation method in action seems to have a gut level feeling that the road to proficiency runs somewhere outside of textbooks and classrooms.  Indeed, in the literature, every reviewer makes a special effort to acknowledge the incredible contribution that Krashen had made to language education.  Kramsch (1995) points out that the input metaphor may be a relic of the prestige of the physical sciences and electrical engineering, but that Krashen’s acquisition-learning dichotomy cuts at the heart of academic legitimation.  She advocates a more productive discourse between applied linguists and foreign language teachers to explore and question the historical and social forces that have created the present context.
        Krashen’s conclusion to his presentation at the 1991 Georgetown University Round Table on Languages and Linguistics (Krashen, 1991) is especially telling about what he is trying to achieve:  “It is possible that ‘no pain, no gain’ does not apply to language acquisition” (p. 423).  Certainly this may be true for some learners and in all likelihood it is true for more communicative methods when compared to older methods.  But the majority of us have had to struggle to be able to understand and speak a language, no matter how much exposure to “comprehensible input” we have had.  And the particular circumstances of language minority students in the U.S. and many other countries certainly indicate that those children have formidable barriers to overcome just to understand the first things their teacher is saying.  To propagate such an “easy way” philosophy in the policy of state educational boards, EFL textbooks and general teacher guides is to demean the effort that less able students have to make every day.  To institutionally impart such a concept to new teachers whose responsibility it is to understand these adults and children is a disservice to all parties involved.  Despite the pressing need of policy to provide a workable teacher training system, it is imperative that, at the very least, there is no misinformation.  Second language learning is a very complex process, with many make or break factors involved and there is simply no comprehensive theory to guide teachers and students at the moment.
        This does not mean, however, that teachers should be sent to their classrooms with no direction, or worse yet, back to a grammar-based or audiolingual approach.  The issue of exactly what and how to tell teachers to teach is one of the most complex and sensitive issues that policy has to implement.  It is only through basic research into a wide variety of areas such as the role of exposure in comprehension and production that we can begin to develop the policies to create the best practices for the classroom.


California Commission on Teacher Credentialing Home Page. Revised Knowledge and Skill Areas Assessed on Tests 1-3 of the (Bilingual) Crosscultural, Language and Academic Development (CLAD/BCLAD). 1998. California Commission on Teacher Credentialing.  14 December 2000.

Ellis, R. (1992). Comprehension and the acquisition of grammatical knowledge in a second language. In Courchãene, R.J. Comprehension-based second language teaching. Ottawa : University of Ottawa Press.

Gregg, K. (1984). Krashen’s monitor and Occam’s razor. Applied Linguistics, 5, 79-100.

Kramsch, C. (1995). The applied linguist and the foreign language teacher: Can they talk to each other? Australian Review of Applied Linguistics, 18,1. 1-16.

Krashen, S.D. (1981). Bilingual education and second language acquisition theory.  In Schooling and language minority students: A theoretical framework. (p.51-79). California State Department of Education.

Krashen, S.D. (1985). The input hypothesis: Issues and implications. New York, Longman.

Krashen, S.D. (1991). The input hypothesis: An update. In James E. Alatis (ed.) Georgetown University Round Table on Languages and Linguistics 1991. Washington, D.C.: Georgetown University Press. 409-431.

Krashen, S.D. (1993a). Teaching issues: Formal grammar instruction.  Another educator comments . . . . TESOL Quarterly, 26, No. 2. 409-411

Krashen, S.D. (1993b). The effect of formal grammar teaching: Still peripheral.  TESOL Quarterly, 26, No.3. 722-725.

Krashen, S.D. & Terrell, T.D. (1983). The natural approach: Language acquisition in the classroom. London: Prentice Hall Europe.

Lewis, M. (1993). The Lexical Approach: The state of ELT and a way forward. Hove: Language Teaching Publications.

Lightbown, P.M. & Pienemann, M. (1993). Comments on Stephen D. Krashen’s “Teaching Issues: Formal grammar instruction.”  TESOL Quarterly, 26, No.3. 717-722.

McLaughlin, B. (1987). Theories of second language learning. London: Edward Arnold.

Paradis, M. (1998). The other side of language: Pragmatic competence. Journal of Neurolinguistics, 11, Nos. 1-2. 1-10.

Terrell, T.D. (1977). "A natural approach to the acquisition and learning of a language". Modern Language Journal, 61. 325-336.

Return to I.C. Language and Literacy  ->  LAU top