During the 1970's we witnessed an explosion of research on second language acquisition, first mainly concerning immigrants to the United States, later also many studies among migrant workers in Western Europe. This research developed largely independently from the classroom «method» studies that drew so much attention in the late 60's and early 70's. While the latter took their design from existing paradigms in educational psychology, the former benefited from the progress made in the study of child language acquisition. Studies on «naturalistic» language acquisition, as it came to be called, quickly outnumbered those on classroom variables, and the debates on language teaching methods became increasingly devoid of empirical support.

Throughout the 60's and 70's, researchers failed to link classroom studies with research in the native-speaking environment. The early 80's were characterized by a methodological convergence. An increasing number of studies began to describe language development in the classroom with the methodology and terminology of research on «naturalistic» acquisition, but there are still very few empirical studies that purport to be a direct comparison of the psycholinguistic processes characterizing foreign or second language learning in the classroom and in the native-speaking environment, either with the same learners in consecutive stages of their experience with the new language, or with comparable (groups of) learners.

There is, however, a considerable body of theory and speculation based on a posteriori comparisons of data drawn from the two separate bodies of research concerning formal instruction, on the one hand, and natural acquisition, on the other. The most radical position concerning the psycholinguistic processes involved in these two different experiences has been taken by Krashen (e. g., Krashen 1980, 1981, 1982, 1985; Dulay et al., 1982). Krashen argues that language learning outside the formal classroom draws on largely unconscious processes of acquisition, and that classroom instruction can only be successful to the extent that it stimulates these unconscious acquisition processes. The monitoring ability that results from conscious learning contributes only to correctness, not to fluency, and is operational only when the learner has enough time, is focused on form, and knows the relevant grammatical rules. Fluency, on the other hand, results from acquisition processes that require two conditions: «comprehensible input», i. e., language that the learner can understand but that is enough beyond his present level of competence so that he can learn from it, and a «low affective filter», i. e., an attitude, motivation and personality that allow input to become intake.

The monitoring theory and the studies that tend to support it have been criticized by a number of researchers, both from a substantive and from a methodological point of view. The substantive criticisms mainly concern the radical distinction Krashen makes between learning and acquisition. Although he himself stresses that implicit rules can be acquired in a formal classroom situation, while others have pointed out that natural exposure to the language may contain a lot of formal learning (e. g., Taylor 1978), Krashen insists that the very processes of acquisition are radically different from those of learning.

Stevick (1984) proposed to replace the concept of two discrete processes, learning and acquisition, with a contrast between «poorly integrated» and «well-integrated» configurations in long-term memory. Learning, he argues, can readily produce a configuration in which the form of a new item is clear, but the item will be less well-integrated «with the full range of circumstances and purposes for which it would be appropriate in the spontaneous

––––––––   239   ––––––––

usage of language in life outside the classroom» (34). The degree of integration represents a continuum, and varies from learner to learner and from situation to situation.

Other researchers have made similar points. Bialystok (1981, 1982; cf. also Bialystok & Sharwood Smith 1985; Chaudron 1983, fn. 5) refers to different knowledge representations as falling on an «unanalyzed» to «analyzed» dimension, conceived of as a continuum, and argues that the importance of unanalyzed knowledge gradually increases as the linguistic task varies from «structural» (language-oriented) to «instrumental» (communication-oriented). She recognizes the possibility that both aspects of proficiency may be stimulated during any learning experience. Ellis (1984, 1985) also argues for a «variable competence model» in which more or less controlled or automatized rules prevail at a particular point in time as a function of socially determined discourse requirements.

Similarly, Tarone (1983, 1984) sees the learner's competence as a «capability continuum», one extreme of which is represented by the learner's conscious intuitions about correctness¹³⁸. Furthermore, even when the learner «monitors», his degree of confidence in his rules or intuitions can vary. Therefore, saying that «knowing the rule» is a condition for monitoring is another simplification (Kohn 1982).

Krashen's categorical distinction between learning and acquisition has led him to a series of related points of view, which have been strongly criticized by a number of applied linguists. McLaughlin (1978) points out that there is no evidence for what he calls Krashen's main hypothesis: that what is learned is not available for initiating utterances and can only be used to monitor the output of the acquired system. McLaughlin (1978) proposes to replace the dichotomy of conscious versus unconscious rules by a distinction between controlled and automatic processes, and subsequently claims that learning can become acquisition, i. e., initially controlled processes can gradually become automatized. This point of view is shared by Carroll (1981) and Rivers (1980). Bialystok (1981, 1982) combines the automatic-controlled and the analyzed-unanalyzed distinctions in a two-dimensional model.

Krashen (1979) rejects the possibility that learning becomes acquisition arguing that there are cases of learning without acquisition and of acquisition without learning, and that learning only works for a small set of simple rules. McLaughlin (1978) points out, however, that empirical studies that tend to confirm the last point, such as Krashen et al. (1978) are methodologically unsound: when a learner is asked whether he operated by rule or by feel, he may say «by feel» to avoid phrasing a difficult rule, on the one hand, and «by rule» when he finds it easy to state, on the other hand, even if he did not consciously use it.

The radical distinction between learning and acquisition further implies that monitoring plays no role in the building of an internal grammar, which is the product of unconscious acquisition that takes place through mere comprehension. This amounts to a claim that, on the one hand, transfer from controlled to automatic processes as it is known in many areas of cognitive functioning does not obtain in language learning, while, on the other hand, transfer between two different skills such as language comprehension and language production -yes, even between reading and speaking (Krashen & Terrell 1983, 131)- is possible. Gregg (1984) takes issue with this view by arguing that it makes no sense to hypothesize, as Krashen does, that acquisition can occur through comprehension alone. Swain (1985) elaborates further on this point by introducing the notion of «comprehensible output»: Being pushed in output to make the message comprehensible, not only in the sense of «getting it across», but in the sense of conveying it «precisely, coherently and appropriately» (249), she argues, is a notion parallel to that of i + 1 in output. Therefore, «it can be argued that one learns to speak by speaking» (248). A variety of further arguments for and against Krashen's input hypothesis can be found in Pienemann (1985, 47-49) and Long (1985, 84-88), and can be summarized by saying that comprehensible input is a necessary, but not sufficient condition for successful second language development.

A stronger, expanded version of the monitor theory might give monitoring a crucial role in the acquisition process. Morrison and Low (1983), for instance, hypothesize that, in mediating between conflicting rules, the monitor might eventually help to settle these differences. Sharwood Smith (1981) takes a somewhat intermediary position, granting that comprehensible input is necessary for

––––––––   240   ––––––––

acquisition, but hypothesizing that utterances pieced together by the monitor, with or without recourse to implicit knowledge, constitute a source of input in their own right. Ellis (1984), Gregg (1984) and Schmidt and Frota (1986) take similar stands. Krashen hedges on this point by arguing that one's own output only becomes comprehensible input for acquisition when it coincides with the structure currently being acquired (1982, 1985). Bialystok (1982), on the other hand, allows not only for practice through which the acquired system draws on the learned, but also for inferencing through which transfer takes place from the acquired to the learned system. For a comparison of Krashen's and Bialystok's views, see Tollefson et al. (1983).

The methodological criticisms raised against the monitor theory concern a variety of issues. The most general criticism has been made by McLaughlin (1978): the monitor theory, he argues, is only an a posteriori framework for restating a variety of phenomena (child-adult differences, the natural order of morpheme acquisition, etc.). Little empirical work has been done, indeed, to test the monitor theory directly, although Krashen (1979) claims that it provides a series of testable hypotheses.

The empirical work aimed at investigating the monitoring process itself has been restricted to errors and their correction (e. g., Krashen et al. 1978, Stafford and Covitt 1978; cf. Kohn 1982). Moreover, the distinction between conscious and unconscious rule application, which these studies rely on, ultimately requires introspection, with all the problems this entails. This is one of the reasons why McLaughlin (1978) proposed to replace the learning-acquisition dichotomy by a distinction between automatic and controlled processes: if a theory is to be falsifiable, it has to rely on behavioral acts.

It is doubtful, however, that self-correction on the basis of the learned system (Krashen's monitoring) can be distinguished from self-correction on the basis of the acquired system (what Krashen has called «monitoring by feel»: cf. Krashen et al. 1978; McLaughlin 1978; Morrison & Low 1983) by any direct empirical means, either by introspection or by the analysis of behavior (Hulstijn & Hulstijn 1984). Moreover, Hulstijn and Hulstijn's own experimental results showed that two groups of learners, with and without explicit knowledge of the rules tested, were influenced by time and attention to the same extent, which seems to suggest that not only monitoring on the basis of learned knowledge, but also monitoring by feel is a function of two factors identified by Krashen: time and attention to form¹³⁹.

In conclusion, then, as indirect experimental evidence seems to suggest that monitoring in Krashen's sense and monitoring by feel are not necessarily as different as Krashen assumes, as there are theoretical arguments against their hypothesized separateness, as the hypothesized differences between these two processes cannot be observed directly in the learner's behavior, and as introspection is not a very reliable source of information about such processes, it may be preferable to collapse them into one category, which we could call monitoring in the wider sense, and to try to investigate what role both of these processes together play in foreign language development in different settings and in different stages of the learning process. This leads us to the following questions:

1. To what extent do learners monitor in the foreign language classroom and in the native-speaking environment?
2. How does monitoring evolve over time?
3. Is there a relationship between the amount of monitoring and the development of the internal grammar of a foreign language?

In order to answer these questions, we need to be able to rely on direct evidence, not on a posteriori interpretations of related phenomena. Furthermore, it is necessary to analyze real communication and not only artificial tasks such as error correction. Language behavior has to be observed directly, and actual language production has to be distinguished from the underlying internal grammar (whether this be a learned or acquired system) in order to assess the prevalence of monitoring in the wider sense. The next section presents a study that attempted to meet these requirements.

2.1 Subjects

The study dealt with two groups of learners. The first group consisted of American students participating in a six-month overseas program in Spain, who were observed during the fall quarter of 1984-85. They had all gone

––––––––   241   ––––––––

through the two-year basic course sequence at their home university, but some of them had skipped the first quarters because they had had the equivalent in high school. Their proficiency level was intermediate. The second group consisted of five students of the same university who were planning to study in Spain or Latin America in 1985-86; they were observed in their second-year, second-quarter Spanish classes in the U. S. during the winter of 1984-85. These students had taken the four preceding quarters of the course sequence (most of the students had skipped the first quarter because they had had the equivalent in high school).

All subjects were volunteers. None of them had spent any substantial time in a Spanish-speaking country before. While these «convenience samples» were small and somewhat heterogeneous, they did present the advantage of having two groups with a substantial amount of shared foreign language learning experience in college, one group having the overseas experience in addition. Further information about the comparability of the two groups can be found in Sections 2.2.5 and 2.3.1.

The U. S. group attended five hours of class a week, representing a fairly balanced offering of grammar theory and exercises, reading and writing, and oral activities. The students in Spain took three courses designed especially for them and taught in Spanish by local professors: Spanish language (two hours per week), Spanish history (three hours), and Spanish literature (four hours). Each individual learner also chose two other classes from the university's regular offerings, and attended these classes together with the local students. While they were in Spain, the Americans lived in university-affiliated dormitories, where they were virtually the only foreigners.

2.2.1 Grammar test

A grammar test was devised in order to assess the base-line grammar knowledge of the participants. It seemed to us that the only valid way of operationalizing the concept of monitoring was by testing the learners on a series of grammar items in a highly monitored task, and by subsequently assessing to what extent the learners adhered to their own knowledge in a variety of communicative settings.

Our test included four typical problems of intermediate-level Spanish: the subjunctive after verbs of volition and verbs of opinion, conditional sentences, the copula, and relative pronouns. These problems had been dealt with in the students' first- and second-year courses. Each of these structures was tested by means of 20 four-choice items. For three of the constructions, the choice was actually a combination of two two-way options: indicative versus subjunctive and present versus past in the first case, subjunctive versus conditional and simple tense versus compound tense in the second, and imperfect versus preterit and ser versus estar in the third. The test was so designed that each of the 16 answers (= four problems x four options) would be correct five times. To summarize, the test contained 80 items representing four grammatical problems consisting of a four-way choice corresponding to four obligatory contexts each occurring five times.

The test items pertaining to the respective grammar problems were scrambled and for each item the order of answer alternatives was randomized. The test was administered to both groups in the first weeks of the quarter.

2.2.2 Interview

In the U. S. as well as in Spain, all participants were interviewed three times by the researcher at approximately three-week intervals, starting about three weeks into the quarter. The interviews were about 10 to 15 minutes in length, and dealt with the students' own daily life and their learning experiences, as well as with topical issues. The interviews took place in the researcher's office in the U. S. and in his dorm room in Spain. They were conducted as casually as possible, but they resembled an interview more than a casual conversation. Each interview contained a few questions meant to elicit past tense verbs, relative clauses and conditional clauses. These questions, however, changed from interview to interview and were geared toward the students' recent experiences or topical issues, in order not to draw the learners' attention to points of grammar any more than would be the case in a natural conversation with a native speaker¹⁴⁰.

On a first level then, the interview yielded data about the students' performance in Spanish under semi-controlled conditions. On a

––––––––   242   ––––––––

second level, some of the questions elicited information on the students' learning behavior and communicative experiences.

2.2.3 Picture description

Around the same dates the interviews were taped, the students participated in a picture-description task, which lasted about 15 minutes each time. The learner was given a drawing, which he had to describe to a native speaker of Spanish, who was paid for his participation. The native speakers tried to copy the drawing from the descriptions given by the learner. The learner and the native speaker could not see each other's drawings.

The main purpose of this task was to elicit a substantial number of communication strategies (hereafter CS) under fairly controlled conditions, in order to facilitate comparisons across individuals and groups. As the learners had not mastered the advanced vocabulary needed for this task, they had to resort to CS continuously¹⁴¹.

2.2.4 Participant observation

The researcher tried to observe as wide a range of the students' use of Spanish as possible. In the U. S., the observation was limited to the classroom, as the students had virtually no contact with Spanish speakers outside class. In Spain, the students were also observed in a variety of other situations. As the researcher lived in the same dormitory as the male students, it was relatively easy for him to observe their interaction with the Spanish residents in an unobtrusive way by sharing their meals in the dorm's cafetería, hanging around in the hallways, joining the students in the traditional bar-hopping, etc. Fieldnotes were kept throughout the quarter.

2.2.5 Control variables

To assess whether the two groups were comparable on a series of variables known to affect foreign language learning in general and monitoring in particular, the subjects were given the Modern Language Aptitude Test (Carroll and Sapon 1959) and two questionnaires: one concerning attitude, motivation and risk-taking, and one about attitude toward correctness.

2.3.1 Group equivalence

The differences between the two groups regarding the control variables were quite small, and none of the t-tests for these five variables (language learning aptitude, attitude, motivation, risk-taking, attitude toward correctness) reached a significance level of alpha = 0.05. The two groups, then, were comparable for the purpose of this study, i. e., to shed light on the questions formulated at the end of section 1.

2.3.2 Monitoring

The extent of monitoring in both groups was assessed by a quantitative comparison of the results of the multiple-choice grammar tests with those of the interviews and picture descriptions. All the structures on the grammar test had been defined quite narrowly to avoid the problem of comparative fallacy (Bley-Vroman 1983). For most structures and most individuals, in the U. S. as well as in Spain, the students' performance on the test, although under conditions conductive to monitoring (multiple-choice, focusing on form, no time limit), reflected considerable uncertainty about what was correct. It seldom happened that one option was consistently marked as correct and/or the others as incorrect. This inconsistent performance on the grammar test, together with the fact that the narrow definition of the structures tested made their occurrence in the interviews and picture descriptions rather infrequent, made reliable comparisons impossible in most cases. Therefore, data will be presented here for the one grammar problem only for which, at the same time, sufficient data were found in the recordings, and fairly consistent results were found on the grammar test, at least for certain individuals.

The problem that yielded the most consistent answers on the grammar test was the choice of the copula. Each of the relevant items had the form of a four-way option combining the ser/estar and the imperfect/preterit dichotomies. In the tables that follow, the asterisk designates the correct option. Figure 1 on the following page lists the values of correctness for the copula among the seven students in Spain. These values range from -1 (for answer alternatives consistently marked as incorrect) to 1 (for alternatives consistently marked as correct). To increase the number of occurrences in each category, the difference in aspect was uncoupled from the noun/location distinction. Thus, as can be seen from Figure 1, students 2, 5 and 6 never used ser for a location, and students 1, 2 and 7 never used estar before a predicate noun.

Figure 2 presents the frequencies of the copulas used before a noun in the recordings (interview + picture descriptions), for the three students with a high degree of consistency on the grammar test. As can be seen from the table, all used ser correctly in all instances recorded.

Figure 3 presents parallel data for the copula before an expression of location. While the data are less clear-cut in this case, the frequencies for ser are generally quite low; the percentage of errors varies between a maximum of 20% (October) and a minimum of 11% (November). In other words, those students who made consistently correct judgments on the grammar test for this construction used the right copula in more than 80% of all cases in the recordings.

Figure 4 lists the values of correctness for the copula among the five students in the U. S. The data in this table are generally inconsistent, except for the fact that students 1 and 5 almost consistently rejected the use of estar before a predicate noun (a value of correctness of -0.8 means that in one out of five cases the student had doubts about the correctness of the alternative in question, but never marked it as simply incorrect). The frequencies in Figure 5 pertain to the recordings for these two students with consistent judgments on the grammar test for ser/estar before a predicate noun. The table shows that both students used ser with a high degree of consistency in the recorded oral tests.

To summarize, those students in Spain who made consistent judgments about the use of ser before a predicate noun, always used it correctly in communicative tasks, and those who made consistently correct judgments about estar before an expression of location used it correctly in the vast majority of cases. The students in the U. S. with consistently correct judgments for ser before a predicate noun used it correctly in almost all cases. Students in both groups appeared to monitor their knowledge about ser and estar quite efficiently, at least those students whose knowledge appeared to be solid.

2.4 Discussion

The first question concerning monitoring was: To what extent do learners monitor in the foreign language learning context and in the native-speaking environment? The study addressed this question by comparing, for both groups, the students' performance on a multiple-choice test concerning intermediate-level grammar problems and on a series of interviews and picture descriptions. While the students' judgments on the grammar test were usually rather inconsistent, and while the absolute frequencies of the forms under investigation in the communicative tasks were too low for a reliable comparison with the grammar test in most cases, it was found that those students with consistently correct judgments for certain constructions complied with their own knowledge during communicative tasks in the vast majority of obligatory contexts¹⁴². This could be shown for the use of ser before a predicate noun and for the use of estar to express a location among the students in Spain, and for the use of ser before a predicate noun among the students in the U. S.

The students' inconsistent judgments on other parts of the grammar test were reflected in their choice of the default option where that was possible, i. e., choosing que as relative pronoun or choosing the indicative after affirmative verbs of opinion, or in avoidance strategies and a variety of mistakes, where there was no clear default option, i. e., in conditional sentences. When students overseas were asked about differences in their learning or speaking since they had come to Spain, they never mentioned anything like the difference between consciously learned grammar in the U. S. and picking it up automatically in Spain. When asked more explicitly whether they agreed that language development abroad was largely automatic, they all rejected this idea. No evidence was found, then, to suggest that foreign language development in the native-speaking environment involves less monitoring then learning in the classroom context. Our answer to the question «from learning to acquisition?», then, can only be «no», at least for the students observed.

The second question concerning monitoring was: How does monitoring evolve over time? As no differences in degree of monitoring were found between the two groups, it comes as no surprise that no clear evolution during one quarter was found either, neither in the U. S. nor in Spain. The only case where improved monitoring of a known rule was documented was for the use of estar in Spain, and even there the evidence was scant. In all other cases, the numbers were too small to allow for comparison, or the students' usage was either completely consistent (for ser) or completely inconsistent (for conditionals) throughout the quarter. Even after the use of conditional sentences was explicitly reviewed in class for the group in the U. S., no noticeable improvement took place.

Classroom instruction is often equated with a large amount of grammar drills and vocabulary explanation, while residents in a foreign country are supposed to master the language through communication, with little conscious attention paid to structure. Students blame classroom instruction for failing to bring about oral fluency, claiming «we don't want to learn grammar; we want to learn how to speak»: The results of our study, however, do not suggest such a strong dichotomy between learning language in the classroom versus picking it up abroad, or between grammar and oral proficiency. What we found, rather, were two gaps: one within written achievement, between the student's satisfactory performance on tests and exams on the one hand, and their generally inconsistent judgments on our multiple-choice test, on the other hand; a second one within oral performance, between students' satisfactory proficiency on our interviews, on the one hand, and the inability of most students to sustain a prolonged conversation with native speakers, at least during the first few weeks of their stay abroad,

––––––––   245   ––––––––

on the other hand.

What is lacking, then, is not so much the controversial interface between what is learned and what can be used for communication, but rather a thoroughly learned system that would allow students to perform consistently on any kind of grammar test, and that could be drawn on during oral communication activities of various sorts.

The third question was: Is there a relationship between the amount of monitoring and the development of an internal grammar of the foreign language? While our research design was not primarily oriented toward answering these questions, our observational data on individual differences (cf., DeKeyser 1986, in press) made it clear that one of the students, who monitored very conspicuously, was perceived by the Spaniards as harder to talk with than another student, whose monitoring style and use of CS were more geared toward dissimulating his problems of grammar and vocabulary. The latter student, it should be noted, monitored constantly too, but he managed to dissimulate this through a variety of CS. As a result, he interacted far more with the Spaniards, and in the long run, this should differentially affect the interlanguage system of the two learners, if not necessarily through internal psycholinguistic processes such as automatization, then at least indirectly, i. e., through different exposure and interaction.

It was shown that the intermediate Spanish learners in this study monitored extensively, both in the U. S. and in Spain. In those few cases where they had unequivocal knowledge of target language structures, this knowledge was applied with a high degree of consistency during oral communication tasks. For most structures tested, however, the students' monitored knowledge was more limited than expected, as became clear from the results on the grammar test, where students' answers were highly unsystematic for most of the 16 structures on the test, despite the fact that these structures were very narrowly defined.

The implications for practice formulated below should be read with two caveats in mind. First, the conclusions of this study pertain in the first place to two particular groups of learners who studied at a particular university in the U. S., and who attended a particular overseas program in Spain¹⁴³. Second, due to the students' weak performance on the grammar test, the data showing successful monitoring during communicative tasks are scant.

If it is correct that students can make only limited use of their knowledge of grammar during communicative tasks, not because of a lack of interface between learned knowledge available for monitoring and acquired knowledge available for communication, but simply because their knowledge of grammar is extremely shaky, the implication for educational practice is quite straightforward: students need to know their grammar better. This does not mean that grammar should be a goal in itself, it may imply that more time should be spent on certain points of grammar, or that grammar should be taught in a different way, or that it should be better integrated with the rest of the language course.

As our study mainly dealt with structures the students had not sufficiently learned, we could not marshall any substantial evidence for or against the notion of automatization. We were able to document, however, the fact that students were able to draw on their grammar knowledge during certain communicative tasks, on the one hand, but that most students felt far from comfortable interacting with several native speakers at once, especially during the first weeks of their stay in Spain, on the other hand.

These facts do suggest that a more gradual transition from the explicit learning of grammar, at one extreme, to spontaneous interaction with a group of native speakers, at the other, via one-to-one conversation with a native speaker or fluent teacher in a relatively anxiety-free context, would enhance the usefulness of students' knowledge of grammar for communicative purposes. This implies a change at several stages of current practice, because even in the communicative classroom, the teaching of grammar and unstructured conversation are often juxtaposed activities, rather than the begin and end points of a carefully planned progression.

This paper is based on parts of my Ph. D. dissertation submitted to Stanford University in August 1986. I would like to thank my advisor Dr. Shirley Heath, and the other members of my reading committee, Dr. Charles Ferguson

––––––––   246   ––––––––

and Dr. Thomas Huebner, for their comments on my dissertation, and Prof. M. Debrock and Dr. N. Delbecque of the Katholieke Universiteit Leuven for their comments on an earlier draft of this paper. All remaining errors of fact or interpretation are, of course, my own.

Bialystok, E. 1981. «The Role of Linguistic Knowledge in Second Language Use». Studies in Second Language Acquisition 4:31-46.

___. 1982. «On the Relationship Between Knowing and Using Linguistic Forms». Applied Linguistics 3: 181-206.

Bialystok, E. and M. Sharwood Smith. 1985. «Interlanguage is Not a State of Mind: An Evaluation of the Construct for Second Language Acquisition». Applied Linguistics 6: 101-17.

Bley-Vroman, R. 1983. «The Comparative Fallacy in Interlanguage Studies: The Case of Systematicity». Language Learning 33: 1-17.

Carroll, J. B. 1981. «Conscious and Automatic Processes in Language Learning». Canadian Modern Language Review 37: 462-74.

Carroll J. B. and S. Sapon. 1959. Modern Language Aptitude Test. Form A. New York: The Psychological Corporation.

Chaudron, C. 1983. «Research on Metalinguistic Judgments: A Review of Theory, Methods and Results». Language Learning 33: 343-77.

Corder, S. 1983. «Strategies of Communication». In C. Faerch and G. Kasper (eds.), Strategies in Interlanguage Communication. London: Longman, 15-19.

DeKeyser, R. 1986. «From Learning to Acquisition? Foreign Language Development in a U. S. Classroom and During a Semester Abroad». Unpublished Ph. D. dissertation, Stanford University.

___. In press. «Communicative Behavior in a Foreign Language: Individual Differences in Monitoring and Communication Strategies». IRAL.

Dulay, H., M. Burt and S. Krashen. 1982. Language Two. Oxford: Oxford University Press.

Ellis, R. 1984. Classroom Second Language Development. Oxford: Pergamon.

___. 1985. Understanding Second Language Acquisition. Oxford: Oxford University Press.

Gregg, K. 1984. «Krashen's Monitor and Occam's Razor». Applied Linguistis 5: 79-100.

Hulstijn, H. J. and W. Hulstijn. 1984. «Grammatical Errors as a Function of Processing Constraints and Explicit Knowledge». Language Learning 34: 23-43.

Hyltenstam, K. 1985. «L2 Learners' Variable Output and Language Teaching». In K. Hyltenstam and M. Pienemann (eds.), Modelling and Assessing Second Language Acquisition. Clevedon: Multilingual Matters, 113-36.

Ingram, D. 1985. «Assessing Proficiency: An Overview of Some Aspects of Testing». In K. Hyltenstam and M. Pienemann (eds.), Modelling and Assessing Second Language Acquisition. Clevedon: Multilingual Matters, 215-76.

James, J. 1980. «Learner Variation: The Monitor Model and Language Learning Strategies». Interlanguage Studies Bulletin Utrecht 2: 99-111.

Kohn, K. 1982. «Beyond Output: The Analysis of Interlanguage Development». Studies in Second Language Acquisition. 4: 137-52.

Krashen, S. 1979. «A Response to McLaughlin, "The Monitor Model: Some Methodological Considerations"». Language Learning 29: 151-67.

___. 1980. «The Input Hypothesis». In J. Alatis (ed.), Current Issues in Bilingual Education. Washington, DC: Georgetown University Press, 168-80.

___. 1981. Second Language Acquisition and Second Language Learning. Oxford: Pergamon.

___. 1982. Principles and Practice in Second Language Acquisition. Oxford: Pergamon.

___. 1985. The Input Hypothesis: Issues and Implications. London: Longman.

Krashen, S., J. Butler, R. Birnbaum and J. Robertson. 1978. «Two Studies in Language Acquisition and Language Learning». ITL 39-40: 73-92.

Krashen, S. and T Terrell. 1983. The Natural Approach. Oxford/Hayward: Pergamon/Alemany.

Long, M. 1985. «A Role for Instruction in Second Language Acquisition: Task-based Language Training». In K. Hyltenstam and M. Pienemann (eds.), Modelling and Assessing Second Language Acquisition. Clevedon: Multilingual Matters, 77-99.

McLaughlin, B. 1978. «The Monitor Model: Some Methodological Considerations». Language Learning

––––––––   247   ––––––––

28: 309-32.

Morrison, D. and G. Low. 1983. «Monitoring and the Second Language Learner». In J. Richards and R. Schmidt (eds.), Language and Communication. London: Longman, 228-50.

Palmer, A. 1979. «Compartmentalized and Integrated Control: An Assessment of Some Evidence for Two Kinds of Competence and Implications for the Classroom». Language Learning 29: 169-80.

Pienemann, M. 1985. «Learnability and Syllabus Construction». In K. Hyltenstam and M. Pienemann (eds.), Modelling and Assessing Second Language Acquisition. Clevedon: Multilingual Matters, 23-75.

Rivers, W. 1980. «Foreign Language Acquisition: Where the Real Problems Lie». Applied Linguistics 1: 48-59.

Schmidt, R. and S. Frota. 1986. «Developing Basic Conversational Ability in a Second Language: A Case Study of an Adult Learner of Portuguese». In R. R. Day (ed.), Conversation in Second Language Acquisition. Rowley, MA: Newbury House.

Selinker, L. 1972. «Interlanguage». International Review of Applied Linguistics. 10: 209-30.

Sharwood Smith, M. 1981. «Consciousness-raising and the Second Language Learner». Applied Linguistics 2: 159-68.

Sorace, A. 1984. «Connaissance et usage dans l'apprentissage d'une langue seconde: une interprétation de la variabilité». Travaux Neuchâtelois de Linguistique Appliquée 7: 43-92.

Stafford, C. and G. Covitt. 1978. «Monitor Use in Adult Second Language Production». ITL 39-40: 103-25.

Stevick, E. 1984. «Memory, Learning and Acquisition». In F. Eckman, L. Bell and D. Nelson (eds.), Universals of Second Language Acquisition. Rowley, MA: Newbury House, 24-35.

___. 1985. «Elephant Lore: A Circle of Astonishments». In B. Wheatley, A. Hastings, F. Eckman, L. Bell, G. Krukar and R. Rutkowski (eds.), Current Approaches to Second Language Acquisition. Bloomington: Indiana University Linguistics Club, 155-93.

Swain, M. 1985. «Communicative Competence: Some Roles of Comprehensible Input and Comprehensible Output in its Development». In S. Gass and C. Madden (eds.), Input in Second Language Acquisition. Rowley, MA: Newbury House, 235-53.

Tarone, E. 1983. «On the Variability of Interlanguage Systems». Applied Linguistics 4: 142-63.

___. 1984. «On the Variability of Interlanguage Systems». In F. Eckman, L. Bell and D. Nelson (eds.), Universals of Second Language Acquisition. Rowley, MA: Newbury, 3-23.

___. 1985. «Variability in Interlanguage Use: A Study of Style-Shifting in Morphology and Syntax». Language Learning 35: 373-403.

Taylor, I. 1978. «Acquiring vs. Learning a Second Language». Canadian Modern Language Review 34: 455-72.

Tollefson, J. B. Jacobs and E. Selipsky. 1983. «The Monitor Model and Neurofunctional Theory: An Integrated View». Studies in Second Language Acquisition 6: 1-16.

Yule, G., J. Yanz and A. Tsuda. 1985. «Investigating Aspects of the Language Learner's Confidence: An Application of the Theory of Signal Detection». Language Learning 35: 473-88.

Hispania [Publicaciones periódicas]. Volume 73, Number 1, March 1990