[Abstract]
[Bibliography - publications in German only]
1. The principles of 'constructivist' foreign language teaching (FLT)
In two recent articles (1994, 1997), Dieter Wolff, a prominent German FLT researcher, outlined the features of foreign language teaching on constructivist lines. The first essay in particular has been both widely read and frequently quoted. In launching constructivism as the new paradigm for FLT, Wolff places "constructivist" FLT in opposition to the traditional forms of "instructivism" in language teaching. Most probably, Wolff borrowed the binarism "constructivism vs. instructivism" from American colleagues who had drawn it from the discussion concerning science education theory, where this term has quite frequently been used since the 1980's. In concordance with these disciplines, Wolff defines instructivism as directed by a paedagogical-psychological concept, according to which learners are considered as "reactive beings" ("reaktive Wesen") either accepting or not accepting the material offered by the teacher. According to Wolff (1997: 47), constructivists, however, postulate that learning is a process which is highly subjective, autonomous, and active.
Michael Wendt, who holds the Bremen chair of education in the romance languages, is another prominent exponent of 'constructivist' FLT (1996a, 1998). But both in his theoretical groundwork and in the teaching propagated in his writings, he emphasises a position which is quite distinct from that of Wolff. Nevertheless, there is a great affinity between the constructivist approaches of both of them.1 This can also be inferred from the principles of teaching propagated by Wolff and Wendt. These are by no means idiosyncratic or unusual, but they more or less correspond to the commonly 'progressive' approaches frequently displayed in FTL reviews in recent years. On the whole, the approaches focus on mediation, which, currently (at least in the literature), is a prevailing theme and which, in my opinion, is moving in thedirection to a more open methodological approach.2 Graph no. 1 illustrates the most important features of this approach in a schematic drawing.
The major principle to which Wolff's (1994: 21 f.) as well as Wendt's (1996a: 77 ff.) lines of reasoning can be related is action-orientated methodology including cooperative learning, (i.e. social forms of learning such as work with partners or in groups), active and creative forms of work as well as teaching by projects. Furthermore, Wendt (1996a: 75) argues for an approach called 'learning by teaching' ("Lernen durch Lehren"; "LdL"), which encourages the pupil to take over the teacher's role. Additionally, Wendt favours intercultural projects and the Freinet approach, which is a method based upon the writing and printing of the pupil's own texts, as he holds in high esteem its possible "fusion of individualising and cooperative forms of learning" (Wendt 1996a: 79).
The individualisation of learning, which, in my schematic drawing, is centred on the learner, is one of the most important principles of constructivist FLT. Wolff (1997: 47) argues that learning can only be influenced by teaching in a very restricted way. The individual learner should be allowed to choose his own selection from exercises and texts presented to him. This choice is in line with the principle of learner autonomy (cp. Wolff 1994: 426 f.; Wendt 1998: 7). Thus the learner's decision will take into account the respective type and style of learning.
One prerequisite, however, is that the learner is instructed on how to become aware of his own learning behaviour, i.e. why he or she selects and applies particular strategies and techniques. Wolff (1994: 421) suggests using the computer as a tool, while Wendt (1996a: 76) has a rather sceptical attitude with regard to the use of electronic media. On the one hand, Wolff (1994: 424 f.) maintains that learning awareness should be complemented by language awareness, whereas Wendt stresses the intercultural awareness and the "construction awareness" aspects of the learning process.
|
Principles of "constructivist" FLT
|
Another major principle of constructivism is holistic language experience, as shown above. It links up to content-oriented FLT, which generally takes place in project instruction or, as Wolff underlined (1994: 424), in bilingual classes. According to this approach, foreign language acquisition will thrive in an authentic and complex learning environment or situation (Wolff 1994: 418). Wendt (1996a: 82 f.) similarly emphasises that authentic contacts with foreign language and culture have crucial relevance because the non-directed acquisition of language skills is made possible at the level of trial and error. According to Wendt (1996a: 40 f.), even authentic fictional texts and movies offer the opportunity to the learner to examine his or her constructions of reality in the foreign culture and to 'negotiate' them within their group of learners.
It is evident that these principles of methodology and procedure discussed so far were not developed by Wendt and Wolff themselves. As shown above, these principles of FLT reflect the general consensus in the 1990's, although their feasability has not yet been tested. The novelty of Wendt's and Wolff's concept consists in their having brought together the various principles and procedures under the one roof of constructivist FLT. The framework of constructivist theory offers coherence and a superstructure for the various concepts of learning and teaching involved in this definition. In Wolff's opinion (1994: 412, 421), the FLT theoretical framework will meet the approval of both the radical philosophers of constructivism and the more moderate representatives of constructivist approaches in cognitive psychology. Wolff believes that an amalgam of these two stances forms a suitable basis for a coherent theory of cognition, perception, and learning and will provide a foundation of FLT.
To Wolff, Radical Constructivism appears convincing particularly because it makes the individuality and autonomy of learning explicit. In this context, Wolff (1994: 412, 421) sees "the constructivist and thus subject-related character of human perception and cognition, the closed circuit of human organisms with regard regard to processing information, and finally the impossibility of directing perception and cognition from the outside" as the decisive points. Furthermore, Radical Constructivism would do justice to the active role of the learner, as this theory particularly emphasises the fact that "the learner needs to have contact with others in order to validate his assumptions of the environment, to reach a consensus on the way the environment is constructed."
Contrary to Wolff, Wendt exclusively relies upon the tenets of Radical Constructivism and puts its epistemological core problems into the centre of his FLT theory. In the following, I am going to briefly outline its theoretical abstractions and examine the exact constituents of his ideas.
2. The central ideas of Radical Constructivism
The term Radical Constructivism was coined by Ernst von Glasersfeld, a philosopher and psychologist teaching in the U.S.A. It stands for an epistemology which had evolved in an interdisciplinary discourse at the Biological Computer Laboratory (University of Illinois at Urbana-Champaign) and, also, in this institution's "intellectual environment" during the sixties and seventies. At this institute, which was directed by Heinz von Foerster, a number of neurologists, philosophers and psychologists worked together on the problems of cybernetics and on the foundations of AI research (cp. von Foerster/von Glasersfeld 1999: 172 ff., 232 ff.). The intellectual origins of Radical Constructivism can be traced to systems theory, which was established by research in biology and neuro-science.
Apart from Ernst von Glaserfeld, the Chilean neuro-biologist, Humberto Maturana, who was employed by the Biological Computer Laboratory for a certain period of time, belongs to the founding fathers of Radical Constructivism. Maturana's theory starts from very general reflexions about the material 'functioning' of all living beings on this earth. He describes organisms as autopoetic, i.e. self-creating systems which keep alive for a certain amount of time by materially reproducing themselves. As autopoetic systems, organisms are principally autonomous and neurally cut off from the environment, although they are still connected with it by exchanging matter and energy. Species need to adapt to their external environment to secure their survival (for a limited period oftime). This process is thus evolution for the individual species, whereas individual beings operate by cognition and the resulting change of behaviour vital for their survival. At both levels this is not a question of determinism. The general evolutionary framework of the environment permits different modes of development to the species. This openness of phylogenesis corresponds to the level of cognition and behaviour of individual organisms whose perception of one and the same environment and whose reaction to the very same stimulus may differ widely from each other; the range of principally equal possibilities will only be restricted by those ways of interacting with the environment which are detrimental to the organism or which threaten its chances of survival (see e.g. Maturana 1991; Maturana/Varela 1991).
After this brief outline of the tenets of Radical Constructivism, by which I wanted to show both its genesis and its roots in topical issues of theoretical biology, we can now turn to the epistemological implications for human cognition, which is the point of interest and discussion for many disciplines and which is particularly relevant to FLT. Even in this field, the above-mentioned idea that living systems are autonomous is applied on the central nervous system. Radical Constructivists consider the human brain a self-referential and self-explicative system. As the Bremen neuro-physiologist Gerhard Roth explains, "This means that the brain is functionally closed and that it can only interact with its own states of being. There is no direct access to the outside world, so that its states are only examined for their meeting the criteria of internal consistency."
In order to prove his thesis that the brain attributes "to its own states meanings which have been taken out of itself" (Roth 1991a: 241), Roth - like other Radical Constructivists - argues that the brain does not "know" any information input if the verb "know" is taken in its strictest sense, which is the reason for the human brain's "self-referentiality". The brain is kept apart from the outside world by the fact that "physico-chemical events in the environment are translated by the sense organs into the 'language of the brain'", which means a "radical break" with the environment (Roth 1998: 115, 334). This separation is seen as a fundamentally radical break, because the "language of the brain" is restricted to the various states of neuronal electric excitement of the neurons, to their chemical equivalents and to the transmitters in the synapses, which either pass on the electric stimuli of excitement or block them. Thus the encoding of the brain is "neutral" (Roth 1998: 249) or "undifferentiated", as Heinz von Foerster put it (1991: 137). Sense modalities which are taken up by the brain and the different qualities of sense perception will not result in different forms of neural processing. Whether it be the noise heard by a person or the visual impression perceived, e.g. the blue car or the red rose, the brain can process the different stimuli only by exciting a complete set of nerve cells, which, principally, cannot be distinguished from each other.
The translation of the physico-chemical events in the environment into the language of the brain requires a highly complex form of processing and of analysing incoming information, a process which is interpreted as "construction" by the adherents of Radical Constructivism. Fundamentally, perception itself is highly selective, which means that the constancy of colour, shape, and materiality is generated by the human brain (Roth 1998: 114, 343). Furthermore, the perception of a concrete object simultaneously activates numerous cell groups, each of which separately process only a very limited aspect of the given object. For instance, the colour of the animal you spot, its shape, spatial depth, location and movement will be processed in different areas of the cortex. The cell units involved in information processing are spread out through many parts of the brain (Roth 1998: 253 ff.). In the last analysis, it is the stored knowledge that exerts a very strong influence on perception. For roughly one hundered thousand neurons in the cortex, which both assess and interpret incoming visual information, there is only one single nerve cell involved in this process (Roth 1998: 124).
The brain's 'self-referentiality' is related to the idea that, while processing information, the brain, as one might say, circles within itself, when it transforms self-produced neural states into differing neural states (see e.g. Roth 1991: 235, 240 f.). Consequently, it cannot know anything reliable about the true origin and meaning of the neural stimuli, as it is unable to compare them with the real objects beyond its own perception. It is not the existence of a reality which exists unconnected with the subjectthat is denied, but, instead, the recognition of this reality. In this context different points of views are discussed. To some Radical Constructivists (e.g. Richards/von Glasersfeld 1991: 218), the whole structure of reality represents a hypothetical construct; to other Radical Constructivists (e.g. Scheffer 1990: 50) there are discrete (i.e. clearly distinguishable) but unspecific stimuli produced by the world outside, the quality of which is determined by subjective interpretation; other constructivists (e.g. Watzlawick 1994: 124 f., who normally figures among the Radical Constructivists) make a distinction between "properties of things exclusively physical and thus more or less objectively definable" and the (constructed) social significance of these phenomena.
Consequently, Radical Constructivists argue that the perception and cognition of the world does not reflect any properties of its matter, as the followers of traditional representational theory have assumed (adaequatio rei et intellectus). The impression caused by the world is produced exclusively by the subject's act of cognition. In other words, this is valid not only for the transphenomenal world, e.g. the atomic microcosm, but also for the everyday world of our perception (Rusch 1991: 378). Accordingly, the checking of propositions and concepts does not depend upon truth or adequacy, for Radical Constructivists do not see any "possibility" of distinguishing "between propositions on 'reality' being true or untrue" (Schmidt 1991a: 41). The feeling of insecurity caused by the human perception of reality results in in all kinds of confusion such as paradoxes and mental delusions (cf. von Foerster 1995: 51 ff.; Schmidt 1990: 64).
Viability, a concept which has been borrowed from evolutionary theory, thus replaces truth. As mentioned above, Radical Constructivism is based on the fundamental notion that organisms have to adapt to their environment, which, in turn, means that they have to cope with restrictive conditions in order to survive; there is quite a number of different possibilities to ensure survival, and only those creatures which are unable to meet the restrictions will be eliminated. Perception and understanding in communication are supposed to operate in a similar way. Ernst von Glasersfeld argues (1987b: 140) that our ideas are continuously checked against and adapted to our experiences which are derived from acting: "We construct ideas, hypotheses, theories and models, and as soon as these survive, i.e. as long as our experience can be fitted into them, they are viable."
Consequently, viability is inextricably tied to a continuous checking of information. Constructs of meaning are tested "for internal consistency and for external viability or suitability for functional existence"(Wendt 1999a: 1). New information may lead either to extending or to questioning older knowledge.3 Inversely, a new piece of information can be examined to see whether it is credible, plausible and compatible with the knowledge already stored and so, can be discarded. Knowledge can be tested within an individual as well as between several individuals. Siegfried J. Schmidt (1992a: 431, 1992b: 25) considers socialisation by linguistic communication the main cause for cognitive agreement between humans, which is developed by Gerhard Rusch (1990: 27) who emphasises the importance of positive and negative feedback in this context. By and large, to the majority of the Radical Constructivists, the process of learning involves the principles of trial and error (see, e.g., von Glaserfeld 1987a: 132). This concept of learning is derived from pragmatism, according to which the acquisition of knowledge serves for reaching the goals for the individual's actions or behaviour in the external world.
There is already a radical-constructivist sign theory for the initial stages. According to Wolfram Köck (1991: 368), signs do not refer principally to definable objects or categories of objects, but only to the representations of things. Michael Wendt (1996a: 36 ff.) uses the term "signal" for "signs which stimulate either the testing processes or the extension processes of reality constructions" and he has devised a typology of signals. Referring to Susanne Langer, Heinz von Foerster distinguishes the denotative or referential function of language, (which he defines as 'monologue') from a connotative representation-related function, which he characterises as a 'dialogue', because it takes into account the epistemological aspect of other human beings. In his opinion, only the constructivist concept, which is the second one, can be applied on communication, as mutual understanding appears to him a purely connotative process (cp. von Foerster/von Glasersfeld 1999: 210 ff., 225 ff.). Umberto Eco, who can by no means called a Radical Constructivist, deals with a similar use of connotation without denotation in his Introduction to Semiotics (1972: 106), but he takes it to be a special case, which can occur, for example, in the proper names of unknown persons. To him, they are "significants with open denotation which are decoded like an abstruse scientific term which are encountered for the first time, but the meaning of which must certainly refer to something very fixed." However, von Glasersfeld (contrary to von Foerster and Maturana) does not reject the concept of denotation, but he re-defines it with reference to the subject (see von Foerster/von Glasersfeld 1999: 228 f.). In his concept, denotation roughly corresponds to what linguists who belong to a realistic epistemology would call concepts, perceptemes or prototypes. But he tends to name linguistic meanings 're-representations' by implying the actualisation of past experience, which can also be considered pre- or extralinguistic (von Glasersfeld 1998: 157 ff.).
3. A critique of Radical Constructivists trying to trace their roots in neurophysiology
Dieter Wolff (1994: 427) favours adapting the radical-constructivist theories of perception, understanding and learning by arguing that they "are a secure stronghold based on the explanations provided by philosophers and biologists." Can this proposition really be maintained, or shall we, after all, recognise the epistemological weaknesses of Radical Constructivism? Is it not marred by one-sidedness, exaggeration, distortion and 'blind spots'?
First of all, I would like to deal with the supposed foundation of Radical Constructivism in biological research. As argued above, Radical Constructivists frequently refer to neuro-physiological principles in order to prove that all representations are subjective. I am going to show, however, that this approach is not practicable for solving the problem of representation. Although it seems to me to be obvious - contrary to a number of philosophers4 - that the processes of semiosis also take place at a physical level of the brain. They occur from a brain-physiological point of view in the "mid-range". Up to the present time there have been no positive results for this level of reasearch into brain processes. We already know much more about the functional "micro-level" of the brain, about the electric and chemical processes in particular types of neurons and about their morphology. For quite some time, a great amount of comprehensive research has been done on the functional "macro-level" of neuronal activity in order to throw new light on how cognitive activities are located in the different areas of the brain. However, even today, there is hardly any detailed knowledge of the "meso-level" where correspondences of signs are processed and stored at the material energetic level. Patterns of neuronal activity in which images are represented in material form have not yet been decoded (Roth 1998: 291).
What are the impediments which make research on the 'language of the brain' difficult? First of all, there is the great complexity of the human brain, which consists of approximately half a billion of nerve cells in an adult. On average, each of these neurons is connected with another one thousand neurons and together with them it can make use of approximately ten thousand links called synapses. Thus information processing forms an extremely intricate and labyrinthine network, where electrochemical impulses converge, diverge or are worked out in parallel (Roth 1995: 50 ff., 1998: 33 ff. and 122).
The second difficulty is that the present test procedures are inadequate for recording physiological events at the "meso-level". Microelectrodes measure brain activity only point by point, whereas the electroencephalograph gives only a rough account of the fluctuations in the potential. Nor are the more recent image-producing procedures able to penetrate into the "meso-level" yet. The tomograph registering the emission of positrons cannot cope with the accelerated neuronal processes, and the spectroscope still produces a relatively rough spatial image. Currently, the best spatial and temporal resolution is achieved by the magnetic encephalograph, which is extremely expensive, but unable to picture areas below 5 mm (Roth 1998: 223 ff.).
An added difficulty is the absence of any more precise hypothetical model of the "brain language" based upon current physiological research. The insights gained so far remain both general and vague. The transmission of data can only be derived from the action potentials, which, however, offer only a few possibilities of differentiation, as Kochendörfer explains (1999: 26): "As the action potential is always the same or, in any case, cannot be interpreted from the content of its properties, a symbolic code can only be established by the presence or absence of an action potential, by the combination of action potentials or by using the temporal distance between them." Furthermore, the utmost importance must be attached to the local position of the nerve cell, which is usually in the centre of a tightly woven and complex electronic circuit within a cell ensemble; it is reasonable to assume that the quality of representations depends heavily upon the location where it is processed, as Gerhard Roth argues (1998: 110, 249). It seems that the place of storage definitely affects the piece of information and thus decisively contributes to its specific semiosis. However, it cannot be possible that, for instance, in all human brains a prototypical representation of the colour red will be stored at a predetermined place in the secondary visual cortex. This is opposed by the plasticity of the brain which permits that, within certain limits, the storage capacity of each human being can be used individually. Neurophysiology has not be able yet to find an answer to the aporia of the brain's plasticity on the one hand and information marked by the place of storage on the other.
For cognitive science, current research in neurophysiology leads to the following conclusion: neurophysiology is unable to provide empirical evidence for the human brain being a self-referential and self-explicative system, nor can the contrary be proved empirically at present. Thus only by precise analysis can we decide upon the nature of the relations existing between re-presentations and the outside world. For this purpose, philosophical and semiotic categories will be given proper consideration.
4. A philosophical and semiotic critique of Radical Constructivism
In his comprehensive doctoral dissertation (1998), Clemens Diesbergen dealt with Radical Constructivism both as epistemology and as a general pedagogical theory. In assessing the validity of this metatheory, he comes to the conclusion that, in some aspects, there is a considerable amount of incoherence and, in other aspects, a degree of indeterminacy which makes the value of its explanations doubtful (Diesbergen 1998: 231, 233).
Several contradictions pointed out by Diesbergen (1998: 207 ff.) can be derived from the fact that Radical Constructivists are unable to maintain their subjectivistic position; in their epistemology realistic elements do appear albeit in a disguised form. Most of the Radical Constructivists assume that the objective, ontic world, which is called "reality" by von Glasersfeld (1995b: 42) and Roth (1998: 324 f.) remains separate from the representational "reality" which depends upon the subject. Such a division diminishes the quality of their argumentation, because, if we take Radical Constructivism seriously, we are unable to know anything about objective reality - neither whether it exists nor if objective reality and the subjective construct of reality, however, converge (cp. also Bredella 1998: 38). Normally, Radical Constructivists ought to retreat to a position of radical agnosticism. For already the assumption that any construct of reality is totally subjective requires a "super" observer (Diesbergen 1998: 205), who simultaneously is able to adopt both the subjective perspective and the external role of the observer and, finally, can link both these perspectives.
Another serious incoherence of Radical Constructivism has been pointed out by Nüse et al. (1995: 251 ff.) and Bredella (1998: 39): The principle of the self-referentiality of all knowledge however systematically applied, will lead to the questioning of or even the refutation of Radical Constructivism. Provided that human knowledge is a pure construct, Radical Constructivism is in itself a construct similar to the brains that have created it. As a consequence of this view, we would all vanish in a vague haze of agnosticism.
I am not sure about what Siegfried J. Schmidt has written on the acquisition of meaning, which appears highly contradictory and thus hardly understandable to me. With regard to Maturana, Schmidt (1991a: 24) denies that there is any input or output of information; environmental stimuli lack a specific character (ib.: 17), thus there is no possibility of transmitting information by language (ib.: 28). However, he believes (in an article published later; Schmidt 1992a: 431) that knowledge is made viable "by a continuous process of interaction and communication, which partly parallels the cognitive systems of the inter-actants." On the one hand Schmidt assumes unspecific stimuli in perception, but on the other hand he postulates specific impressions left by language on the socialisation of individuals. He refuses to admit that linguistic signs are conventional and unambiguously recognisable and nevertheless considers the verbal exchange of information in a framework of action the most important foundation for the acquisition of knowledge.
In my opinion, even the pragmatism of the Radical Constructivists, for whom learning exclusively results from "the success of failure of one's own actions" (Roth 1991b: 148), leads to simplification and reduction. The notion of viability, which was explained above and which has been translated as "fitness to survive" (von Glasersfeld 1987b: 140), "practicability" (Dörfler/Mitterer 1998: 43) or as "(functional) adequacy" (von Foerster/von Glasersfeld 1999: 128). The weakness of this notion is evident in the fact that it is not suited to examine all cognitive constructs equally. As Nüse et al. (1995: 175) confirmingly remarked, the concept of viability is applied to perception rather than to general ideas: "Theories and ideas [...] are not directly relevant to survival, which is why people holding the most abstruse ideas of 'God and the world' can survive, even if their ideas do not have any practical consequences."
The notion of viability leads to a very simplistic concept of understanding and theory of learning. It appears especially reductionist with regard to the complexity of linguistic communication. Its situative holism emphasises only the use of language, and is thus supported primarily by the extra-linguistic context and only secondarily by the linguistic co-text. The internal structure of language which also arises from paradigmatic conventions (e.g. the valeur of a lexeme within a semantic field) is not taken into consideration. Furthermore, reference is not included in the process of understanding for (as argued above) in the view of most of the Radical Constructivists physical stimuli are supposed to be too unspecific for semiosis. Thus there is scant regard for the basic accessing processes, without which, in my opinion, language acquisition and (in the young child) the preceding differentiation of categories concerning the environment can no longer plausibly be explained.
In the sign theory of Radical Constructivism, there are still some more aspects which are equally problematic. Explanatory deficits frequently occur when the term denotaton is not made use of or when it is re-defined in the sense of subjectivism (see above). As Klaus Robra explained in detail in his critical contribution to Radical Constructivist semantics (forthcoming), even Michael Wendt starts out from a subjective and connotative concept of meaning (see e.g. 1996b: 130) to which he later adds meaning as "negotiated in society" and as inter-subjectively valid denotation (1996a: 70). However, the range of this 'realistic corrective' and the degree to which the conventions of society play an active role in language acquisition and restrict the individual's scope of construction remain vague.5 As far as I am familiar with the literature of Radical Constructivism, I do not see how linguistic expressions, which are generally governed by norms, can be constructed subjectively. It could be because I just have not come across the adequate explanation for this issue. When, for instance, the radical-constructivist linguist Martin Stegu (1999: 2) postulates that on the one hand the recipients do not have "any direct access to the acoustic and written forms of signs", and when, on the other hand, they are supposed to "perceptively (re-)construct" them, there is an obvious contradiction here. How can 're-construction' be achieved without any access?
5. Proposal for a critically sound constructivism
The great interest Radical Constructivism has consistently attracted in Germany during the last ten years arose in the context of the currents in postmodern science. The psychologist Norbert Groeben (1995: 157) emphasises the fact that this reception must be seen in connection both with the failure of the "great unifying theories", such as Marxism, and also with psychoanalysis to a certain extent. In a paradoxical way, Radical Constructivists not only react against the great philosophical systems, (as they are involved with expressing their doubts about recognising the world) but they are also designing the ultimate monistic plan for an epistemological system. This system is conceived as a closed system by some of its representatives. Siegfried J. Schmidt (1991a: 40), for example, believes that "radical constructivist positions cannot possibly be mixed with realistic ones"; being a constructivist means by definition being it "indiscriminately", "radically".
If, as I have argued above, according to von Glasersfeld (1991: 402), coherence is the highest standard of quality in epistemology, Radical Constructivism undoubtedly fails. More importantly, however, I reject this theory because it is unable to clarify either the origin or the acquisition of knowledge to my satisfaction. In my estimation, the weaknesses derive from the 'radicalism' of this epistemology. I agree with Wendt (1996a: 84), that Radical Constructivism cannot, however, be falsified systematically either, for the decision in favour of an epistemology will not be deduced from axioms, but it will be made from a general assessment of its explanatory potential.
A moderate form of cognitive constructivism that is based upon critical realism appears more convincing to me. In contrast to the naive realist, the critical realist assumes that ontic reality and the reality of experience must be distinguished from each other. But contrary to radical constructivists, he argues that there are correspondences between both realities and that, normally, there is a one-to-one relation between the two worlds in the mathematical (not in the colloquial) sense of the word, which means according to Nüse et al. (1995: 178): "The phenomenal world and the real world have different properties indeed, which, as a rule, co-vary." In doing so the phenomenal world substitutes the ontic world, i.e. it is a sign for it (ib.: 1995: 179, 182). Apparently Gerhard Roth also comes nearer to this position: "There must be a dependable relationship between the exterior world and the activity of the recipient for the brain's successful survival, otherwise, the brain is unable to connect the activity of the recipient to the exterior world" (1995: 51).
While the cognisant subject plays an important role in Radical Constructivism, critically realistic constructivists try to strike a balance in estimating the interaction between the world and the subject. Consciousness can be directed to an exterior world, but this is still done from a subjective point of view. This concept corresponds to the pattern of information processing, according to which language is understood at the interface of data-driven and concept-directed procedures, of bottom-up and top-down movements of understanding (see Wolff 1990: 614 ff.).6 The phenomenal world is a mixture of mimesis (the mirroring of ontic reality) and construction, in which both factors amalgamate, because they cannot clearly be separated from each other. Thus even the critical-pragmatic constructivist cannot infallibly recognise ontic reality, but he can, instead, merely build a hypothetical model of this reality.
But in this respect, everyday knowledge and even a large amount of scientific knowledge are not very problematic, because for many objects of cognition we need only an inter-subjective checking of similarity to the phenomenal worlds rather than strictly objective knowledge. By the way we intuitively know from our communicative practice that there are different modes of cognition, the inter-individual differences of which (and the extent of construction) are smaller than in other areas of cognition; this knowledge has been acquired during our socialisation by comparing our "views" with alein ones which we try to explore by appropriate indices. Therefore, for instance, the scope of interpretation in sensual perception tends to be less extended and similar knowledge is by and large less contested than in discussing abstract issues, for sensual perception has a fundamentum in re. Since Galileo, philosophers and psychologists have for this reason distinguished between the primary qualities things already have and the secondary qualities which are created by the perceiving subject (Nüse et al. 1995: 136). Even Watzlawick (1995: 71) differentiates between primary reality, which corresponds to the material world, and secondary reality, i.e. a 'purely spiritual' world.
However, it is not only the continuum consisting of the concrete and the abstract which determines the degree at which construction occurs. First of all, emotion will exert its influence on human interaction, which produces all sorts of fallacies such as projections and misunderstanding, as repeatedly described by Watzlawick (cp., for instance, Watzlawick 1994: 14 ff., 72 ff.). In spite of all these difficulties, critical realists are not willing generally to renounce the concept of truth, although the notion of truth in propositions has to be submitted to an exact definition case by case (with regard to its conditions of validity). The critical realist assumes that the world is partly recognisable, but he also propagates constant doubt, which we should experience facing our representations of reality, as well as the ensuing necessity constantly to put our own experiences to the test by using as many different sources of knowledge as possible and to connect these experiences with the various sources.
6. FTL is following suit: subjective monism rejected
The weakness of Radical Constructivism, thus in particular the fundamental thesis of the informational closedness of learning systems, has a limited use for a didactic meta-theory. The basic processes of learning, teaching, interaction and class organisation are only insufficiently taken into account. Even more seriously in the field of FTL, the central 'objects of learning' (unlike physics, for example), namely the respective languages and cultures, cannot take a meaningful part as societal constructs within the dichotomy of trans-phenomenal reality vs. subject-related reality. Whereas this dichotomy may prove helpful for presenting the epistemological possibilities and limitations of physical structures, the existence of verbal and culturally specific signs cannot be attributed to a strictly objective level, but only to an inter-subjective one. Despite all the efforts to protect Radical Constructivism from attacks with the argument that the sign (more precisely: its signifier) has no objective-real meaning, its major tenets are doomed to failure. 7
As a result of its weaknesses Radical Constructivism does not qualify as basic theory for current directions in the instruction methodology of FLT. However, this statement does not pass judgement on the methodological concepts which are sometimes linked to Radical Constructivism. Methodology can probably be better founded on psycholinguistic models or on concepts of cognitive psychology rather than on epistemological criticism. Furthermore, there are a number of pragmatic arguments which bring to the fore the methodological principles of learner-centredness, action-orientedness, awareness of the process involved and holistic experience of language.
These methodological principles, however, are not generally acceptable. In the same way, as the absoluteness of the subject makes hardly any sense in epistemology, I consider the radical orientation on the learner questionable. Certainly Wolff (1997: 47) is right when he believes that in the process of learning the contents will "undergo changes which depend upon the available knowledge of the learner, upon his strategies and style of learning." At the same time I disagree with him when he describes the didactic notion of the "instructivist" approach as obsolete, according to which "the contents of teaching are mirrored as contents of learning in the learner" (ib.: 46). For there are, in fact, empirical studies which come to the conclusion that the instructivist axiom may be useful to a certain extent. In her study of how declarative and procedural knowledge is acquired Patricia East (1992: 210), for instance, reaches the following conclusion: "There are hardly any learner-formulated rules. Pupils adopt the patterns of grammatical training into their explicit knowledge without any changes, although they do not fully understand these structures."
The contradiction between the transfer of contents and the subjective shaping of them, between understanding which benefits from good explanations by the teacher and the necessity of personal thoughtful elaboration by the learner will at first glance appear paradoxical. But there is hardly any principle of teaching nor of learning nor a didactic axiom which is universally valid. The world does not 'function' on one and the same principle. So, for instance, the axiom of the authentic and complex learning environment as propagated by Wolff (1997: 48) is put into question by the experience that the neglect of structured progression will not improve learning and will even demotivate some of the learners. At middle-ability secondary schools in particular, learner autonomy cannot be fully implemented when a certain number of pupils refuse to learn a foreign language, if great effort is needed for this task.
Philipp Aronstein, a leading specialist in the teaching of English and French, had already stated in 1917 that many principles of organising FLT are antinomical in character. As a result of this, a specific teaching or learning process reflects the tension between two possible extremes: recourse to previously learnt languages on the one hand and the exclusive orientation towards the target language on the other, and similarly between conscious learning and acquisition with a relatively low level of onsciousness on the other. As I would like to add, tension is also caused by "instructivist" and "constructivist" principles of learning. Whether "instructivist" or "constructivist" techniques of instruction appear to be more successful is still not generally clear, but it depends both on the respective aim and on the group of learners. Wendt and Wolff are certainly quite correct in stating that in German schools, FLT is, on the whole, too instructivist. This is due to the fact that many teachers are not very familiar with implementing group or partner work, projects or "learning by teaching" in FLT, because they have not yet internalised the respective techniques of instruction yet. Furthermore, useful materials and concepts for the classroom supporting the individualisation of learning, e.g. the practice of learning strategies, are slowly evolving. Nevertheless, I am beginning to agree with the analysis of the "constructivist" organisation of teaching by Rolf Dubs (1995: 901) and do not believe either that "exclusively self-directed learning will remain effective" (although empirical research has hardly been able to contribute anything to this problem); patterns of practice and usage which are centred upon the pupil and restricted to the communication of learners with each other will not succeed in entirely replacing the interaction of teacher and learner. As I understand it, however, a moderate school of constructivism will hold its place in FLT in the near future.
Notes
1 Another well-known FLT researcher favouring Radical Constructivism is Werner Bleyhl, although he avoids this label with regard to his own approach. As Bleyhl distances himself from Wendt and Wolff by supporting the non-linear, recursive cybernetics used by constructivists ('secondary cybernetics') and by leaning towards some of the American 'alternative' approaches (particularly to the comprehension approach created by Harris Winitz), he will be disregarded in the present context. Bleyhl's theory can only be discussed separately.
2 With regard to these new principles of methodology (with reference to Königs 1991), I would like to suggest the expression "neo-communicative FLT" (see Reinfriedt 1999: 336 ff. and forthcoming).
3 In Jean Piaget's theory of schemata (frequently adopted by contemporary Radical Constructivists), the first case can be likened to assimilation, as the new information is partly a repetition of the old one and, partly a complement. In the second case, it obviously is accommodation; perturbation, i.e. disappointment or surprise, will duly lead to the testing and (at least partially) to a correction of the schemata (cf. Von Glasersfeld 1998: 113 ff.).
4 The dualism of mind versus brain, which presupposes that psyche and physis belong to different "areas of being" was still rigidly applied in the 17th and 18th centuries; the most prominent representatives of this classical dualism are Arnold Geulincx, Nicole Malebranche and Gottfried Wilhelm Leibniz. In the 20th century, this dualism suffered an almost total loss of influence due to the predominance of science and its principle of general causality. But there still are representatives of a moderate form of dualism. One example is John C. Eccles who, on the one hand, postulates the autonomy of the spirit on the brain (which is just an instrument for the spirit), but who on the other hand will not deny a psycho-physical parallelism, although to him the psyche represents the controlling and bridging factor.
5 I agree to Wendt's (1999b: 2) reply to Bredella that "insignificance is a property of the objective-real (ontic) level of the physical sign". However, there is hardly any language theorist today (like some of the neo-humanists of the 19th and early 20th century) still searching for a transphenomenal 'source' of verbal meaning or for its truth 'as such'. The crucial issue is the question whether or not signs as societal constructs carry an intersubjectively conventionalised meaning which is prior to their acquisition. This has been confirmed by Wendt (1998b: 3). The problem, however, arises if Radical Constructivism and its concept of viability can convincingly explain the constitution of linguistic intersubjectivity, which, in my opinion, considerably narrows down the ideolectal scope to the individual at the lexical (rather than at the textual) level of understanding.
6 However, Wolff (1990: 614) emphasises the part of language processing which is constructive and which creates meaning. In 1994 (409), in his survey on concepts of understanding in cognitive psychology, he reaches the conclusion that there is a growing tendency to "consider one's own stock of knowledge more important than the incoming stimuli". Cognitive psychology is thus presented as a discipline, which seems to be drifting towards the epistemological direction of Radical Constructivism. In my opinion, such a generalisation is questionable, because in the 1990's, apart from the researchers on textual understanding who generally take inference to be of the utmost importance, there are others who come to the conclusion that "the result of text processing" is just "a representation of the propositional structure of the text; only in a few places is the knowledge of the world added" (Kelter/Habel 1996: 267). I believe that such controversies cannot be solved at a general abstract level at all, but that the amount of inference (and thus the share of construction) may vary greatly depending on the respective sort of text and on the users of language (with their respective 'styles of thought').
7 See my discussion of Wendt's argumentation above in note 5. In my opinion, there should principally be an intersubjective level inserted on equal terms between the level of strict objectivity and that of strict subjectivity. Regrettably, epistemological realists often avoid doing this, so that their concept of objectivity remains imprecise and hazy.
This paper offers a description of the more recent methodological principles in foreign language learning for which some theoreticians lean on radical constructivism. The radical constructivist perspective which is based on arguments taken both from neurophysiology and philosophy is submitted to a critical analysis with special reference to epistemology. The author comes to the conclusion that radical constructivism will not work as a methodological meta-theory for language learning because it has a number of serious weaknesses at the explanatory level as well as areas of inconsistency. This paper argues, however, for a realistic constructivist version of the theory, which may well serve as an epistemological basis and which, at the same time, opposes learner-centred theories taking an absolutist position in language learning.