Task-based Language Learning in Bilingual Montessori Elementary Schools: Customizing Foreign Language Learning and Promoting L2 Speaking Skills

Foreign language learning has been a part of German elementary schools for several years now. Montessori schools focusing on individual learning, i.e. mostly independent from the teacher and based on auto-education, interest, and free choice, are also asked to teach an L2. The original lack of a concept of L2 learning for this environment has brought forth different approaches. Bilingual education seems to be feasible and applicable in Montessori education. The downside to this is that even in a bilingual classroom the Montessori way of learning may not allow for very much oral production of the foreign language. The role of L2 production (cf. Swain 1985, 1995, 2005) for language acquisition has been theoretically claimed and empirically investigated. Output can have a positive influence on L2 learning (cf. e.g. Izumi 2002, Keck et al. 2006). This also applies to interaction (cf. Long 1996), where negotiation of meaning and modified output are factors supporting L2 development (cf. e.g. de la Fuente 2002, McDonough 2005). Task-based Language Learning (TBLL) presents itself as one way to promote oral language production and to provide opportunities for meaning-negotiation. Especially tasks with required information exchange and a closed outcome have been shown to be beneficial for the elicitation of negotiation of meaning and modified output. This paper argues that TBLL is a promising approach for the facilitation of L2 production and thus the development of speaking skills in a Montessori context. It also hypothesizes that TBLL can be implemented in a bilingual Montessori environment while still making the Montessori way of learning possible. Different tasks on various topics, examples of which are presented in this article, can lay the foundation for this. Offering such tasks in a bilingual Montessori elementary classroom promises to foster language production and the use of communication strategies like negotiation of meaning, both being facilitative for L2 acquisition. This hypothesis remains to be tested in future research.


Introduction
Customizing foreign language learning and catering to the individual needs of children have been discussed in educational policies and teaching practice for quite some time. Teachers are expected to optimally support each individual child. This expectation is not easily translated into practice. Montessori schools offer individual learning oriented to the children's needs, interests, and skills in many subjects in Montessori Freiarbeit 1 . Customizing foreign language learning still appears to be problematic because these schools have very limited teacher resources for foreign language learning in this context, e.g. some published teaching ideas and governmental guidelines which seem to be difficult to apply in Montessori individual learning. Therefore the question arises: How can foreign language learning best take place in such a learning environment? Various schools have developed a number of different approaches. One of these is to offer foreign language learning opportunities through bilingual learning, for example immersion, during Freiarbeit. This can be accomplished by introducing language teachers who teach content matter in the L2 2 and only communicate with the children in the foreign language. It seems that learners in such a bilingual environment do not produce much oral language output, nor do they appear to interact very much with each other in the foreign language. Sufficient opportunities for oral output (cf. Swain 1985Swain , 1995Swain , 2005 and interaction (cf. Long 1983Long , 1996 on the part of the learners are hypothesized to foster language acquisition. This paper 3 introduces Task-based Language Learning (TBLL) as a means of enabling individual foreign language learning in a Montessori environment so that the development of speaking skills is facilitated as well. Different task design variables will be discussed in terms of their effect on the learners' interaction and performance during task-based interaction. Finally, specific examples of tasks will be presented.

Montessori education
Montessori education focuses on customizing learning for each learner. It is an education vom Kinde aus ("starting at the child", Scheibe 1969: 55). The learner is seen as an individual with different knowledge, skills, and interests, but with similar developmental periods and tendencies. The Montessori approach intends to cater to these individual needs of each learner. Therefore every child is supposed to be able to learn individually which does not mean that the child necessarily works on its own but rather that it can work according to its specific needs and interests. The child can choose what it wants to work with, how long it will be occupied with this specific activity, who it will work with and where it will perform the task (cf. A. S. Lillard 2007, P. P. Lillard 1996. Such individual learning is based on "auto-education" (G. MacDonald, personal communication, June 13, 2012) in a 'prepared environment ' (cf. Standing 2009: 166). There the learner can find materials in different subject areas. In order for the learner to be able to work with specific material he has to receive a demonstration of working with the material by the teacher. This can be initiated e.g. by the teacher who sees that a child might be ready to go on to a new level or by the child who expresses his or her interest. Once the child has received a demonstration and understands how to work with the provided material for a given subject, the teacher leaves the child to work with the material independently (cf. Klein-Landeck 2009, A. S. Lillard 2007. These individual and independent learning cycles called Freiarbeit (cf. Klein-Landeck 2009) are suggested to take place "uninterrupted" and for approximately three hours every day (cf. Association Montessori Internationale 2010, P. P. Lillard 1996: 93).

3
Foreign language learning at Montessori schools Individual learning takes place in many subjects at Montessori elementary schools. There is a standardized set of materials for Freiarbeit in various areas, e.g. mathematics, geography, and the school's primary language. A standardized set of Montessori materials for foreign language learning does not exist. Maria Montessori, the developer and founder of the reformist approach described here, never developed such material or an elaborate concept for foreign language learning at Montessori schools (cf. A. S. Lillard 2007: 333-334, Klein-Landeck 2004. Where foreign language learning is included in the basic training of Montessori teachers, it is brief at best. Thus, Montessori teachers are mostly left to themselves when choosing how to implement foreign language learning. This is slowly starting to change. In Germany a small number of optional professional development courses on foreign language learning for certified Montessori teachers have been offered during the last few years (cf. Fehrer 2011, Montessori Akademie Berlin n.d., Montessori-Bildungsakademie n.d.). Additionally, some teaching ideas have been published (cf. e.g. Crane-Fisk 1986, Fehrer 2011, Klein-Landeck 2008, Kuty 1997. The only other guidelines Montessori teachers have are governmental policies; for Germany this is based on the Common European Framework of Reference for Language Learning (cf. Council of Europe 2001) which stipulates the level of competency children ought to have acquired by the end of grade four (see for example the guidelines by the School Ministry of Lower Saxony, Germany, Niedersächsisches Kultusministerium 2006).
German elementary schools are required to teach a foreign language beginning at least in grade three (cf. Roos 2006: 24-25), with a few states starting already in the first grade. Thus, Montessori schools are faced with the dilemma of how to teach foreign languages without many resources at their disposal. Some have chosen to teach language lessons or to offer foreign language activities during Freiarbeit, others to apply a bilingual program, e.g. by immersing children in the foreign language during Freiarbeit (cf. Klein-Landeck 2000, A. S. Lillard 2007. Following Swain and Johnson (1997), in this article the term 'bilingual education' includes the intensive bilingual program immersion as well as other bilingual programs. The option of implementing bilingual education in a Montessori environment seems to be convenient because the three hour-cycle of Freiarbeit does not need to be interrupted and individual learning according to Montessori is not disturbed. Among the schools attempting to teach a foreign language through bilingual learning in general or through immersion (cf. section 4) the programs differ considerably (cf. Winnefeld in press).

Immersion and its effects on L2 skills
Immersion as an intensive form of bilingual education (cf. Couve de Murville/Lenz in press, Swain/Johnson 1997) describes an approach where the foreign language becomes the medium of instruction and classroom communication in different content areas such as mathematics or geography (cf. Swain/Johnson 1997: 6-7, Wode 1995. In this way immersion, "along with many other forms of bilingual education, [is different] from contexts where the L2 is taught formally and only as a subject" (Swain/Johnson 1997: 6).
Immersion education and its effects on L2 development at regular schools has been investigated during the last decades. Compared to learners in traditional foreign language classes immersion learners develop a higher competence in the foreign language (cf. Wesche 2002). More intensive immersion programs lead to greater skills in the foreign language than less intensive immersion programs (ibd.). A collation of studies by Johnstone (2002) shows that even though immersion learners "become highly functional in two languages", they do not reach the same level of proficiency in the four communicative skills in the foreign language as native speakers, meaning that they do not reach the same proficiency in speaking and fall behind in terms of grammatical control and sociolinguistic competence (ibd.). One of the reasons for immersion students developing speaking skills below those of native speakers could be seen in Swain and Lapkin's (2002: 289) argument that present studies (cf. e.g. Allen/Swain/Harley/Cummins 1990) show that there are insufficient production opportunities for immersion students in the investigated contexts (cf. also Swain's hypotheses in Swain 1985: 249). These findings cannot be generalized to every bilingual environment. More studies are necessary. Swain (ibd.: 246), who conducted studies at immersive schools, specifies oral productive language use during immersion referring to personal observation by pointing out that peer-peer interaction not directed towards the teacher is usually conducted in the first language. Since students at Montessori elementary schools primarily work by themselves, with a partner or a group, and only rarely with the teacher, this would mean that they only produce a small amount of the foreign language orally. Observations undertaken by the author support this assumption. During bilingual Freiarbeit students used the foreign language very little orally, almost exclusively during teacher-student interaction and rarely in communication with peers. This first finding needs to be validated through further investigations (cf. Winnefeld in press).
With regard to the general findings in immersion as stated above and the preliminary findings in a bilingual Montessori context, one can assume that students in a bilingual Montessori elementary classroom do not produce a lot of oral output in the foreign language. Nevertheless, output plays a role in L2 learning, as described in the following section.

Output in language acquisition
Language production can promote L2 acquisition (cf. e.g. meta-analysis of studies in Keck/ Iberri-Shea/Tracy-Ventura/Wa-Mbaleka 2006 4 ). According to Swain's (1995: 125) Comprehensible Output Hypothesis "[...] producing language serves second language acquisition in several ways". Next to the function of output to improve a learner's fluency in the foreign language due to automatization taking place (cf. De Bot 1996, Swain 1995, output can serve the development of accuracy of the learner's interlanguage. Swain (1985: 248-249) states the value of "'contextualized' and 'pushed' language use", the latter meaning that learners are pushed to convey their message "precisely, coherently, and appropriately". She names three functions of output, namely the noticing/triggering function, hypothesis-testing function, and meta-linguistic function (cf. Swain 1995). Studies investigating the noticing function show that output opportunities have a positive effect on L2 learning (cf. Ghari/Moinzadeh 2011, Izumi 2002, Song/Suh 2008. Output is seen as facilitative for language acquisition, since "production makes the learner move from 'semantic processing' prevalent in comprehension to more 'syntactic processing' that is necessary for second language development" (Izumi 2003: 168). The production of language in an interactive context can lead to feedback from the interlocutor, and this can cause learners to modify their output (cf. Shehadeh 2001, Swain 1995. While different findings on the effect of such modified output on language learning exist (see Ogino 2008 for an overview), modified output is suggested by scholars as potentially fostering foreign language acquisition (cf. Swain/Lapkin 1995), as has also been shown in studies by Bitchener (2004) and McDonough (2005). The need for more research in this area remains.
With regard to theoretical claims and empirical findings on the facilitative role of output, it is assumed that output promotes language development and speaking in regard to fluency and accuracy. Thus, L2 learners could profit from greater opportunities for language production. According to Swain (1996: 97) this also applies to learners in an immersion classroom. Taskbased Language Learning is suggested to provide such opportunities for student interaction and meaningful output production.

6
Task-based Language Learning TBLL is an approach to foreign language learning which "seeks to develop L2 proficiency through communicating" (Ellis 2003: 9). The foreign language should be acquired through the use of the foreign language, which also means oral production.
A constitutive element of TBLL are tasks. There are multiple definitions of a task (for an overview see Bygate/Skehan/Swain 2001, Ellis 2003: 4-5, Van den Branden 2006. Bygate et al. (2001: 11) define a task as being an "activity which requires learners to use lan-guage, with emphasis on meaning, to attain an objective". Ellis' (2003: 4-5) definition of a task additionally entails that a task includes "real-world processes of language use" and that students use their "own linguistic resources" during task work. While a task can focus on any of the four language skills, namely listening comprehension, speaking, reading, and/or writing (ibd.), the term task in this paper will primarily refer to oral tasks in which students have to use their own linguistic resources through speaking.
Tasks are a means of eliciting more oral output from foreign language learners (cf. Shehadeh 2004: 4-6). Studies indicate that learners produce oral language when performing a task (cf. section 7.1.2). Thus, through the implementation of specially designed tasks, opportunities to learn can be created, in which children can 'naturally' interact with each other in the foreign language. During task work the learners produce oral output in the foreign language in order to 'get the task done', i.e. to attain the objective of the task which does not necessarily need to be of a linguistic nature (cf. Ellis 2003: 10, Pica/Kanagy/Falodun 1993: 10).

Task-based Language Learning at Montessori elementary schools
Tasks can be seen as a way of giving students opportunities to produce oral output in the classroom. Therefore, tasks could be a way to enhance foreign language production and therefore promote the development of speaking skills at Montessori elementary schools. By taking into consideration research findings, this section first examines the use of tasks to facilitate the development of speaking skills in general and particularly for children. Second, the question of whether tasks could be implemented at a Montessori elementary school without interrupting Freiarbeit according to Montessori will be explored.

TBLL: Promoting speaking skills through meaning-negotiation
Tasks are suggested to have a positive impact on language acquisition and the development of speaking skills for various reasons (cf. Eckerth 2008:15, Ellis 2003, Skehan 1998. Important factors that have been investigated in a number of studies are negotiation of meaning and task performance (cf. Eckerth 2008: 15). The latter will only be touched on in this paper.

Meaning-negotiation
According to Ellis and Barkhuisen (2005: 166) "the term 'negotiation of meaning' refers to the conversational exchanges that arise when interlocutors seek to prevent a communicative impasse occurring or to remedy an actual impasse that has arisen". Foster's (1998: 11) operationalization of negotiation for meaning implies comprehension checks, confirmation checks, and clarification requests. Following Gass, Mackey, and Ross-Feldmann (2005), these communication strategies will be defined according to Long (1983: 136-137). Long (ibd.) defines comprehension checks as followed: "Comprehension checks, such as Right?, OK?, and Do you understand?, […] [are] an effort to anticipate and prevent a breakdown in communication". In contrast to comprehension checks, confirmation checks and clarification requests are not uttered by the speaker himself but rather by the interlocutor listening to and trying to understand the speaker. Confirmation checks are "any expression […] immediately following an utterance by the interlocutor which are designed to elicit confirmation that the utterance has been correctly heard or understood by the speaker" (ibd.). Long (ibd.) further defines the third communication strategy, the strategy of requesting clarification "as any expression by a native speaker designed to elicit clarification of the interlocutor's preceding utterance(s)". Examples for these different ways of negotiation for meaning can be found in Gass et al. (2005: 585-587) and Ellis (2003: 71).
The use of these communication strategies are supposed to be facilitative to language acquisition. According to Long's (1996: 451-452) Interaction Hypothesis "negotiation of meaning, and especially negotiation work that triggers interactional adjustments by the […] [native speaker] or more competent interlocutor, facilitates acquisition because it connects input, internal learner capacities, particularly selective attention, and output in productive ways". It is also assumed that such negotiation for meaning influences language acquisition positively since it "[…] provides learners with feedback on their own L2 production […] it prompts learners to adjust, manipulate, and modify their output" (Eckerth 2008: 15).
Positive effects of negotiation have been shown in foreign language production (cf. e.g. greater amount of oral output in Van den Branden 1997 5 ), lexis (cf. e.g. de la Fuente 2002, Newton in press, Van den Branden 1997), and morpho-syntax (cf. e.g. Mackey 1999, Mackey/Oliver 2002. The studies by Bitchener (2004) and McDonough (2005) suggest that negotiation that results in modified output can have positive long-term effects. More studies on the effects of negotiation of meaning and modified output are necessary.
As stated, negotiation for meaning can lead to modified output. The Comprehensible Output Hypothesis indicates that tasks that focus on language production and interaction influence language acquisition positively. This is especially the case if negotiation of meaning -and also modified output -is elicited.

Task design and negotiation for meaning
On the basis of this assumption different conceptual features of tasks have been investigated in regard to the elicitation of negotiation of meaning. Findings seem to be clear on the effects of some task design variables, while contradicting results exist on the effects of others. In the following I will summarize definitions and research findings of three task design variables already mentioned by Ellis (2003): open/closed tasks, required/optional information exchange, and one-way/two-way tasks.
Open and closed tasks differ from each other as follows: "Open tasks are those where participants know there is no predetermined solution.
[…] Closed tasks are those that require students to reach a single, correct solution or one of a small finite set of solutions" (ibd.: 89). The 'open/closed' task design variable seems to lead to consistent findings. Research collated by Ellis (cf. Berwick 1990, Crookes/Rulon 1985, Manheimer 1995, Newton 1991, all cited in: Ellis 2003 shows that closed tasks bring about more negotiation for meaning than open tasks. Therefore, closed tasks can be said to be more facilitative to language learning.
When distinguishing required from optional information exchange tasks, Ellis (cf. 2003: 86) refers to the distinction between information gap and opinion gap tasks. In the latter task the information is shared which makes the exchange of information between the interlocutors optional (ibd.: 86, 346). By contrast, in an information gap task information is split between the participants causing the need to exchange information in order to reach an objective (ibd.: 86). Studies indicate a positive effect on the amount of negotiation of meaning in required information exchange tasks (cf. e.g. Doughty /Pica 1986, Gass et al. 2005, Newton 1991, cited in Ellis 2003: 87, Newton 1993. Tasks can also be classified in one-way and two-way tasks. "One-way tasks and two-way tasks are required information exchange tasks that are distinguished in terms of whether the information to be shared is split one-way, i.e. held by a single person, or between two or more people" (Ellis 2003: 88). Here the findings seem to be diverse. On one hand, Iwashita's (1999) study suggests that one-way tasks elicit more negotiation for meaning and modified output than two-way tasks. On the other hand, a collation of studies in Ellis (2003) and Iwashita (1999) indicates that there is support for both, one-way tasks or two-way tasks causing learners to negotiate more for meaning.
Different and contrary findings on the influence of task features on interlocutor interaction could have been caused by differing circumstances in studies, e.g. different operationalizations of meaning negotiation, different ages of learners, and different levels of language proficiency.
In a more detailed review of the studies mentioned above and a further review of other studies on the elicitation of communication strategies, the following (experimental) circumstances also have to be considered. It should be clarified whether or not other task design variables were kept constant when comparing the effect of one task design variable. The topic of a task, discourse mode (cf. Ellis 2003), the meaningfulness of the task to the learners (cf. Iwashita 1999), and perceived task difficulty and cognitive complexity (cf. Ellis 2003) could also influence interlocutor interaction. In general, it is not yet clear how task design features interact with each other (ibd.: 95).
A more in-depth analysis should give further insight into the design of tasks most likely to be beneficial for foreign language acquisition. Based on the preliminary summary of findings on the three task features presented above, it could be assumed that the implementation of tasks that are closed and require the exchange of information would lead to an improvement of the learners' interlanguage and speaking skills.
A study by Foster (1998) throws critical light on the assumption that tasks induce negotiation of meaning. Her results show that students in general only produced a small amount of meaning-negotiation and modified output in the classroom, and that there were great interindividual differences between learners. She suggests that language learners might not be prone to use negotiation strategies in situations of difficulty in understanding. Apart from negotiation of meaning there are other factors that can influence language learning positively. Peer assistance and co-construction (cf. Foster /Ohta 2005, Ohta 2001) occuring during learner interaction can also create beneficial learning opportunities. Furthermore, even though conversation tasks without required information exchange and closed outcome elicit less negotiation of meaning, they seem to offer greater opportunities for pushed output. Nakahama, Tyler, and Van Lier (2001) infer this from their finding that a conversation task led to longer and more complex language use than a task which requires information exchange and has a closed outcome (ibd.). For these reasons it seems advantageous not to restrict TBLL to these latter tasks, but to offer a variety of tasks that have different effects (cf. Winnefeld in press). However, the focus of this article remains on "negotiation of meaning" tasks.

Negotiation of meaning with children
Studies in the field of TBLL such as those mentioned above have primarily been conducted with adult foreign language learners. Transferring these findings to interactions of children might give rise to problems with external validity. This matter will be investigated in the following. In order to do this it seems feasible to look at some of the studies conducted with young learners.
There are only a few studies on the implementation of tasks with younger (beginning) learners (cf. Carless 2002, Keßler/Kohli 2006, Mackey/Oliver/Leeman 2003, Oliver 1998, Philp/Oliver/Mackey 2006, Pinter 2007. Pinter (2007: 191) argues that younger children are not sufficiently capable of attending to an interlocutor's needs yet, and therefore might have difficulty to work with a partner effectively. The ability to collaborate with an interlocutor and to take "full responsibility for ones' own utterances as well as one's understanding of the partners' utterances are skills gradually increasing with age" (ibd.). Oliver (1998: 377) also mentions children's "egocentricity" that decreases with age. If a child is still very focused on him-or herself while communicating, can it negotiate for meaning at all? Oliver (1998) examines this question in a study with children between eight and thirteen years. She comes to the conclusion that "like adults, primary school children can, and indeed do, negotiate for meaning with age-matched peers when working with communicative tasks" (ibd.: 379). It appears that children apply a variety of communication strategies just like adults but differ from them in their proportional use. Compared to adults children rarely use comprehension checks. Just as the negotiation for meaning of adults, children's negotiation of meaning seems to lead to the production of comprehensible modified output. Oliver concludes that tasks that elicit meaning negotiation can be used effectively for foreign language teaching with young learners (ibd.). The tasks applied in her studies are one-way and two-way tasks with a closed outcome and required information exchange (cf. Oliver 1998Oliver , 2009).
Based on Oliver's findings that children do negotiate for meaning and modify their output, the question is again whether findings on TBLL with older learners can be transferred to younger learners. Oliver (1998: 379) states that one needs to be careful when doing so and that the effects of negotiation for meaning of children differing in age need to be examined further. In a study (cf. Oliver 2009: 146) with younger children (five to seven years) she found that they "can and indeed do negotiate for meaning with their peers" as well. Building on Oliver's (1998Oliver's ( , 2009) and Mackey et al.'s (2003) findings that children are able to negotiate for meaning and to modify their output in peer-interaction 6 and due to the small amount of studies in the field of child TBLL, it will preliminarily be assumed that young L2 learners can profit from task-based interactions.

Further studies
There are further studies that elaborate on the effects of different task design on learners' interaction, such as teacher guidance, planning time, task complexity, and the repetition of (similar) tasks (cf. e.g. Oliver /Philp/Mackey 2008, Philp et al. 2006). However, an analysis of these designs is out of the scope of this article.

Performance-based task analysis
Tasks have not only been evaluated in regard to the elicitation of negotiation for meaning but studies have also investigated the influence of task design variables and task features on learner performance (cf. Ellis 2003 for an overview). According to Skehan (1996) tasks can be used with the aim of fostering fluency, accuracy or complexity of learners' foreign language utterances. Exemplary studies have been conducted by Rahimpour (n.d.) on the task design variable 'open/closed', by Robinson (2001) concerning task complexity, by Bygate (2001) on the effects of task repetition, and by Philp et al. (2006) on the influence of different planning time on children's oral task production.
In conclusion, research findings seem to show that the implementation of tasks fosters the learners' production of utterances in the L2 and results in improvement of different aspects of the learner language, also meaning speaking skills.

TBLL: Catering for individual needs in the Montessori environment
From the perspectives of negotiation for meaning and task performance, tasks seem to represent a beneficial approach to foster language acquisition, notably language production in school children. It could be expected that these positive effects might be a way to enhance student output in a Montessori classroom.
The implementation of foreign language learning in a Montessori environment is not trivial since Freiarbeit should not be disturbed, meaning that individual learning on the basis of interest, free choice, and auto-education also needs to take place regarding the L2. The question arises as to whether tasks could be successfully implemented in Freiarbeit in a Montessori classroom. TBLL appears to be implementable in the Montessori environment while at the same time making the Montessori way of learning possible. In this case tasks need to become part of the prepared environment. As any other material in different thematic areas, e.g. mathematics, tasks are additionally offered on specifically labeled ("English") shelves in the classroom or an assigned area on a shelf of a matching subject (e.g. biology). Kuty (1997) supports the view that different tasks, e.g. an information gap activity, could be used in the Montessori classroom.
The prerequisites for the implementation of tasks in Montessori Freiarbeit, notably that learning with tasks would be possible to take place on the basis of interest, personal freedom of choice, and auto-education will be discussed in the following.

Learning on the basis of interest using tasks
Montessori education tries to cater to the child's individual needs by giving him or her the chance to perform work that is of personal interest. Learning with tasks on the basis of interest seems to be possible if the creation of tasks takes into account the immediate environment of the children (cf. Cameron 2001: 30-31) and their respective interests. A variety of thematic areas could be offered using tasks (see below for examples).

Learning on the basis of free choice using tasks
In Freiarbeit children normally have the personal freedom of choice. Tasks appear to be suitable for Freiarbeit since it is possible to integrate tasks that cover different themes and have different formats in the prepared environment. Children could be given the chance to choose which tasks they would like to work with.
It might be necessary to teach some of the vocabulary before the actual performance of the task. Above all, however, there needs to be a demonstration of the work with the task, which can be supported by examples (cf. Oliver et al. 2008) or by "pre-task modelling" (cf. Kim/McDonough 2011). This seems to be parallel to the procedure of demonstrating how to work with different Montessori materials in the Montessori environment. Surely this also increases the amount of target language input that children receive in bilingual Montessori environments, and therefore needs to be taken into account when implementing tasks in Montessori elementary schools.

Learning on the basis of auto-education using tasks
Children in a Montessori environment are supposed to work independently once they know how to work with a certain kind of material of their choice. This seems to be possible with tasks. According to Nunan (2004: 36-37) the learner's active independent work and use of the foreign language which can be preceded by the teacher's explanation is a feature of TBLL. Pica et al. (1993) also state that "[…] participants take an active role in carrying out a task […]. In other words, […] a task is an activity which participants, themselves, must carry out".
A demonstration, or a pre-task phase, followed by the child's independent work by him-or herself, with a partner or in a small group would make TBLL during Freiarbeit possible. At the same time Freiarbeit would not be disturbed.
To conclude, the implementation of tasks could be a means of catering for the children's needs and of customizing foreign language learning. One could imagine a variety of tasks that are developed on the basis of current research findings and that cover different thematic areas and are of different difficulty.

Task examples
There are diverse task types which on the one hand require information exchange and on the other hand lead to a closed outcome said to foster language acquisition as stated above. In the following a selection of tasks containing these features will be presented.
Such tasks are picture difference tasks in different thematic areas as used in a variety of studies (cf. Duran/Ramaut 2006: 59, Gass et al. 2005: 608-609, Pinter 2007. Here two learners receive very similar pictures that only differ to a small extent. Without showing these pictures to each other the two learners try to find the differences through communicating to each other. These kinds of tasks require an exchange of information between the learners and have a closed outcome, i.e. a set amount of picture differences. Other tasks that also imply these exact task design variables (required information exchange; closed outcome) which positively influence language acquisition and oral language production are information gap activities as presented in Stoppe (2007: 19) and Klippel (2000: 274).
Here the learners each receive an incomplete picture that can only be completed by exchanging information with the interlocutor whose picture contains the missing information. A similar "jigsaw" task (for a task typology see Pica et al. 1993) is also used in Oliver (1998). It differs from the one described in the material used, in this case cut-outs and items that are to be placed in partly furnished kitchen outlines.
Besides these two-way tasks, there are one-way tasks hypothesized to foster language acquisition, such as the picture description tasks used in Iwashita (1999) and Oliver (1998). In these tasks a learner describes a picture to an interlocutor who then draws a picture according to the description. The closed outcome is the drawing that is supposed to correspond to the original picture.
Also imaginable is the following task: Building a farm. One learner has a picture of a farm which he describes to his interlocutor, who in turn creates the farm by placing elements such as plastic animals, a farm house, trees, etc. according to the description.
The different tasks described here cover a variety of themes in the context of classroom (cf. Duran/Ramaut 2006: 59), playground (cf. Gass et al. 2005: 608-609), home (cf. Oliver 1998, Pinter 2007: 206-207), zoo (cf. Stoppe 2007, and farm animals (cf. Klippel 2000: 274). Tasks could consider many other thematic areas and therefore cater to children's interests, while at the same time containing the described task design features of required information exchange and closed output, thus creating promising opportunities for language acquisition.

Conclusion
Montessori elementary schools try to cater to the individual needs of children through the concept of Freiarbeit, which is the prevalent way of learning in these environments. Those Montessori schools offering foreign language learning by bilingual education as part of the regular Freiarbeit are assumed to lack sufficient opportunities for children's oral use of the foreign language. Most classroom time in Montessori education is spent in student-student interaction, which is most likely to occur in the mother tongue. As demonstrated above, L2 learners can profit from opportunities for oral production in the foreign language, due to the positive impact of language production and interaction on language learning. Task-based Language Learning is seen as a way to create such opportunities for oral interaction and L2 production. The consideration of research findings on the effects of task design and implementation variables can optimize these learning opportunities. It is suggested that this also applies to the use of tasks with young learners. The implementation of TBLL in a Montessori elementary classroom thus appears to be beneficial. TBLL in a Montessori environment could be imagined to take place without interrupting Freiarbeit, since a variety of tasks can become part of the prepared environment and task-based work can happen on the basis of interest, individual choice, and auto-education, i.e. the Montessori way of learning.
The implementation of tasks in a bilingual Montessori environment is thus hypothesized to be possible regarding individual learning according to Montessori and to cause learners, even young beginner foreign language learners, to produce more output, negotiate for meaning, modify their output, and therefore improve their speaking skills. This hypothesis certainly has to be tested in the field. A qualitative study on task-based interactions, e.g. negotiation of meaning and amount and nature of output, in a bilingual Montessori elementary school is currently in preparation (cf. Winnefeld in press). Other studies on the effectiveness of tasks indicate that the implementation of tasks can enhance output opportunities in bilingual Montessori elementary schools.