THE ROLE OF INTERNAL CONSTRAINTS AND FANTASY ELEMENT IN CREATIVE LANGUAGE TASKS FOR FOREIGN LANGUAGE LEARNERS

For the past few years, language teaching, especially for foreign learners, has become more and more creative. Creativity has been deemed paramount in the history of language use by human beings, including day-to-day language play and literary works. It is also suggested that creative language tasks are helpful for learners when it comes to writing. This small-scale study involving two Indonesian university students in Auckland was conducted to find out how adult learners, at any rate, take advantage of imagination in language learning. It examines the role of internal constraints and fantasy element in helping adult learners create meanings in doing creative language tasks. The results show that these two components in creative language tasks could help adult learners play with language. By combining prior knowledge and existing linguistic resources, the participants came up with new meanings from previously known concepts which did not seem to relate to each other. Specifically, input and outcome constraints are useful, while external constraints should be left out. Fantasy element is also beneficial as it requires an act of imagining. For recommendation, to stimulate students to experiment with language, teachers need to design creative tasks which have appealing topics and require them to collaborate.


INTRODUCTION
Even though many scholars have differing definitions for the term "creativity", in general, creativity is often associated with something out-of-the-box, resulting in a new or an original outcome. Elias (2012) points out that creativity is human beings' innate ability "to take existing elements of their environment and modify them to meet their needs" (p. 4), which is a distinct feature from animals and, from an archaeological point of view, has been evident in the material culture, such as stone-tool making and other early technologies. In other words, creativity is human beings' natural response to adapt to their environment which touches almost all aspects of life.
In a narrower sense, some people relate creativity to more novel and concrete products, for instance, music, arts, architecture, and language. Particularly concerning linguistic creativity, a subtle line can be drawn between creative language use in everyday life and literary creativity. The former type of linguistic creativity may originate from the notion that human beings are homo ludens who, as mentioned by Carter (2004, p. 54), find amusement in the act of playing, including playing with language. Experimenting with puns and proverbs, for instance, is a common activity in conversations.
While everyday language play deals with the process, literary creativity is somewhat more connected to literary products. These products have at least three characteristics, each of which is linked to the three approaches to literary creativity in Goodman and O'Halloran (2006). First, the inherency approach regards literary creativity as a linguistic violation by deviation or parallelism. Second, the sociocultural approach sees literature serving a bigger purpose, which is to present social and ideological views. Third, the cognitive approach perceives literature to be schema-refreshing (Goodman & O'Halloran, 2006, pp. 6-23). Thus, it is not an overstatement to say that human beings tend to have a positive attitude towards language play in order to gain pleasure by producing verbal arts or consuming literary works.
When it comes to language teaching and learning, creativity is also argued to benefit learners. An early study conducted by Ottó (1998) revealed that students' scores on creativity have substantial relations to their English grades, which were designated as an indication of success in second language learning. It is suggested that the higher the level of creativity of a student was, the higher his/her English grades were. In the study, the scores on creativity incorporate four aspects, namely fluency, flexibility, originality, and elaboration, and each of them was reported to contribute to L2 learning. For example, the author postulates that the score for ideational fluency, which is defined by Carroll (1993in Ottó, 1998 as "the ability to think of different verbal responses falling into a specified class", helped students to come up with a wide range of ideas when doing tasks in the English tests. Although this study was smallscale and involved only Hungarian students learning English, the results can shed light on the positive relations between creativity and English scores. It should be noted, however, that the relations had not been proven to be causative and were still correlative. Furthermore, albeit limited, empirical research into the observed effects and implementation of creative language tasks in English as a Foreign Language (EFL) settings was conducted. First, with regard to the divide between language form and meaning in language teaching, Tin (2011) examined 23 Indonesian university students in their process of doing one acrostic and one simile task. She concluded that acrostics, as an example of form-oriented play task, afford opportunities for students to attend form and meaning simultaneously. Meanwhile, the simile task, lacking formal constraints compared to acrostics, gave room for students to generate meaning in L1 first and then find the equivalents in L2.
Second, more recently, Chow, Hui, and Chui (2018) conducted an experimental study with 90 elementary students in Hong Kong over ten weeks. Students who received a total of six creative literacy activities as part of their compulsory literacy instruction at school were reported to exhibit increased positive reading attitude, as opposed to the decline in reading attitude of students in the control group, during the intervention programme. Third, with a similar age group of participants, Liao, Chen, Chen, and Chang (2018) carried out an eight-week intervention to an English curricular program with 256 pupils in Taipei. The treatment given to the experimental group was brainstorming strategies, in addition to the same teaching materials used by the control group. The results show that the primary students receiving the treatment improved significantly in their English vocabulary, creative-thinking skills, and, to a lesser degree, learning motivation test results. These were attributed to the extent of autonomy granted to students when doing the classroom tasks, afforded by the creativepedagogy approach.
Those studies confirm some benefits creativity may bring to language learning and teaching. The next question, though, would be how creativity can be applied in second/foreign language learning and teaching. In terms of materials development, examples were offered by Maley (1998) through his twelve generalizable procedures to adapt texts and make them creative materials, which are expansion, reduction, media transfer, matching, selection/ranking, comparison/contrast, reconstruction, reformulation, interpretation, creating text, analysis, and project work, while the types of text that can be used range from haiku to short newspaper articles (pp. 288-291). In terms of task design, Dörnyei (2001) proposed the inclusion of challenge, interesting content, a novelty element (i. e. the activity is unfamiliar or unexpected), an intriguing element, an exotic element, a fantasy element, a personal element (i.e. students can relate the activity to their lives), competition, tangible outcome, or humor (pp. 76-77) to make tasks more exciting and motivating for students. Apart from the above discussion, it turns out that creativity is not a factor that stands alone without any stimuli. A trigger may be needed to evoke one's creativity, and one of them is imagination. Imagination and creativity are often mistaken as one single entity, while they are actually not. As mentioned by Mellou (1995, p. 97), they are "closely related and involve complex internal processes rather than single measurable and observable behaviours." Vygotsky (2004, p. 9) did not use those terms interchangeably and stated imagination to be the basis of any creative activities. Still, although they are different by definition, they both operate in the same principle of reproduction of known ideas into new meanings, which was referred to as the 'third order' of imaging (Ainsworth-Land, as cited in Mellou, 1995, p. 104). This makes imagination central in language learning. In fact, to completely understand the meaning of a word from another language and be able to use it for communicative purposes, adolescents or adults need imagination to relate the meaning of the word to their reality (Sawyer & Ebrary, 2003, p. 71).
That said, this study proposes that incorporating imagination to stimulate creativity in language learning can be done by designing creative language tasks which comprise at least two components, namely fantasy element and constraints. Regarding the former, in his theory of creativity, Vygotsky (in Sawyer & Ebrary, 2003, pp. 69-70) pointed out that an individual's creative imagination starts from childhood play then continues to adolescent fantasy, resulting in subjective (which is associated to wish fulfillment) and objective (which is associated to cultural transformation) fantasy which are gradually balanced in adulthood. Similarly, Dörnyei (2001, p. 76) wrote that "everybody, children and adults alike, enjoy using their imagination for creating make-believe stories, identifying with fictional characters or acting out pretend play", so it is assumed that giving tasks that require students to fantasize may arouse their willingness to practice with language.
As for the latter, although it seems that imagination is linked to giving freedom to learners so that they can explore the use of words in L2, too much freedom can restrain them from language play. Instead, constraints will support creative language learning and teaching "by both limiting and directing the creative process and search" (Tin, 2012, p. 179). To clarify, "constraints" here refer to certain conditions set by the teacher so that the tasks can optimize students' linguistic potentials. Joyce (2009( , as cited in Tin, 2012 differentiates between external constraints (environmental, social, and other external conditions surrounding the performance of a task, such as time limit and supervision) and internal constraints (constraints relating to the design of a task). Tin (2012, p. 181) explained three ways internal constraints can be manipulated which make creative language tasks distinctive from communicative tasks, as follows: 1) goal constraints (the general purpose of the task is presented in an ill-defined manner, rather than in a well-defined manner); 2) input constraints (access to input and resource is restricted, and any input given has to be usedinput requirement); and, 3) outcome constraints (the requirements imposed on the product of a task).
While creativity in childhood and adolescents has been studied (e. g. Vygotsky, 2004), little is relatively known about how creativity develops in adults' life, mainly in its relation to foreign language learning. The empirical studies reported previously address creative language teaching and learning for children (Chow, Hui, & Chui, 2018;Liao, Chen, Chen, & Chang, 2018) and university students (Tin, 2011). Differently, this study aims to find out how adult foreign language learners, at any rate, take advantage of imagination in language learning. In particular, this study seeks to examine the role of fantasy element and constraints in helping adult learners create new meanings from known forms in doing creative language tasks.

METHOD
Instead of doing experimental research or using a quantitative approach in a confirmatory fashion, this study follows Tin's (2011) qualitative approach intended to get a closer look at both the product and process of creative language tasks. Adapting from Task-Based Language Teaching (TBLT) evaluations (Ellis, 2003, pp. 323-324), this study reports the results of a micro-evaluation which focuses on single tasks rather than a whole task-based course or macroevaluation. Two types of evaluation were included. First and foremost, a response-based evaluation examined the process and the products of the task implementation to see whether the actual outcomes were in line with the predicted outcomes. For the present study, internal constraints and fantasy element were projected to help adult learners play with language. Second, a student-based evaluation identified students' attitudes towards and perceptions about the tasks.

Data Collection
The participants were two female Indonesians who were adult learners of English. Both of them were international students at a university in Auckland, New Zealand, each taking Master of Commerce in Accounting (participant A) and Master of Arts in Psychology (participant B). At the time of the data collection, participant A had been studying in Auckland for 7 months, while participant B had been studying longer for 14 months. As both participants were international students, they had satisfied the postgraduate English language requirements from the university which can be equated to at least B2 level as independent users according to the Common European Framework of Reference for Languages (Council of Europe, & Council for Cultural Co-operation, 2001, p. 24).
The data collection procedure consists of two parts. The first one was teaching a pair of creative language task to the participants who acted as learners. The instruments used for the first data collection included Task 1 worksheets, Task 2 worksheets, observation notes, and audio recordings. The observation notes were taken by the researcher during the lesson. The audio recordings were taken during the lesson and were then transcribed. The lesson was taught in a non-formal instructional setting at one of the participants' flat to create a relaxed learning environment. Nevertheless, English was used as the language of instruction to resemble a typical EFL classroom situation. As such, the participants were thinking out loud in English while doing the tasks although sometimes they used Indonesian when searching for words in the L2 as they were allowed to have a discussion in either English or Indonesian. While they were working on the tasks, the researcher audio-recorded their conversation using a recorder application and took notes of observation. Task 1 was finished in 10 minutes 47 seconds, while Task 2 was finished in 12 minutes and 28 seconds.
The second one stage of the data collection was interviewing the participants on the same day, not long after the tasks were finished. The instruments used in this stage were structured interview questions and audio recordings. The interview was a guided interview (Bell, 2005, p. 161) without any prepared checklists, but it had a focus of discussion. There were originally eleven questions devised as prompts, but there were two added follow-up questions. These initial questions addressed the learners' perceptions of the creative language tasks (whether or not they enjoyed doing them and which one they enjoyed the most and they found to be the most challenging), recall of the writing process, and their preference for working independently versus collaborating with peers. The interview was done with both participants to enable them to complement each other's answers. Their answers were audio-recorded, and notes were taken as well. In addition, a short follow-up interview, which was also audio-recorded, was done on another arrangement with participant B to get further information. With the approval of the participants, all of the interviews were carried out in English as the participants were capable of doing so.

The Tasks
The creative tasks themselves consisted of two writing tasks. Writing tasks are considered to generate tangible outcomes (Dörnyei, 2001, p. 76), so presumptively, they would appear motivating to the participants. Moreover, the choice of writing skill over other skills, especially the receptive ones, was thought to be feasible to be done as both participants had been studying in an English-speaking environment and had been accustomed to writing postgraduate assignments.
In Task 1, the participants were given four pictures and asked to choose one of them randomly but without knowing what to do with the picture selected. This presented an ill-defined goal (goal constraint) at the beginning of the task as well as an input requirement (input constraint). Since the pictures were put upside down, they could not see and, thus, could not deliberately pick the picture they wanted. The reason behind the use of images lies on the fact that "image" and "imagination" are semantically related. To be able to imagine, one supposedly needs a mental image in his/her mind. This is in line with Ribot and Baron (1906, pp. 4-5) who said that creative imagination arises from preceding perceptions which come from stored images of past experiences.
Hence, as the participants did not know what kinds of pictures they would see, the researcher chose pictures of things about which the participants had background knowledge. The pictures represent Indonesian local cultures which were images of a girl wearing kebaya (Indonesian ethnic women's outfit), a plate of gado-gado (Indonesian side dish), a traditional house named gadang, and Borobudur (a historical site of statutes in Central Java). The participants were then required to write ten sentences describing the picture, which would serve as pre-inventive forms, referring to the forms generated "without a clear knowledge of the actual meaning and function they will serve" (Tin, 2012, p. 181), as opposed to pre-specified language forms in usual controlled practices. To do Task 1, a pen and a piece of task paper were distributed to them. They were also told that they were expected to work together and not limited by time in doing the task. This was meant to eliminate external constraints.
In Task 2, they were asked to choose ten content words from the description that they wrote in Task 1. Thus, an outcome constraint was presented. They were permitted to choose a noun, an adjective, an adverb, and a verb, but not a pronoun, a preposition, or a connector. Those ten words were later used to write a 10-sentence princess story. The idea of asking them to write a princess story, instead of other types of story, was because a princess story typically has a fantasy element in it. The ten words did not need to be put right after one another but all of them were to be used and could be repeated. An additional instruction was given to the participants that the story should be as good as the one that would appear in a story book.

Data Analysis
In principles, the results from the first data collection were used for a response-based evaluation, in which the data collected were quantitatively analyzed through the process of coding, determining themes, constructing an argument, and reassessing the data (Holliday, 2005, pp. 72-73). Then the actual outcomes were be compared to the predicted outcomes (Ellis, 2003, p. 324). Meanwhile, the results from the interview stage were used for a student-based evaluation aimed at investigating learners' perceptions and attitudes.
In practice, the three audio-recordings were transliterated. Combined with notes from direct observation, the audio-transcripts were subsequently analyzed. Comments were given to both the products and the process in each task. Finally, significant findings were jotted down and categorized into three sub-topics under the discussion section, namely "Imagination and Idea Generation", "Imagination and Idea Exploration", and "Individual Differences and Task Preference". The terms "idea generation" and "idea exploration" in this article refer to the Geneplore model by Finke, Smith, Ward & MITCogNet (1996, p. 17) regarding the cognitive process of creativity which consists of the generative phase and the exploratory phase. Meanwhile, the sub-section "Individual Differences and Task Preference" grouped together emergent themes in the interview data and observation notes.

FINDINGS
The results of this study focus on two areas. First, as part of the response-based evaluation, the products of the task implementation and the process that came along with it are reported based on the participants' writings, the transcripts of the audio-recordings, and the researcher's observation notes. Second, for the student-based evaluation, the interview data are reported. In addition, for the purpose of discussion, participant A will be called Vita, while participant B will be called Zahra.

Response-based Evaluation: Task 1
The picture picked up by the participants was the one with the girl in kebaya (Indonesian ethnic women's outfit). On the left column are the sentences they wrote to describe the picture. On the right column are the observation notes about the process. Zahra directed attention towards the hair of the woman in the picture, but she did not know the word for a meaning in L1. Vita then came up with the correct term. 6 The kebaya consists of two parts, the upper part made of brocade material, and the skirt usually made of batik.
Vita proposed to discuss the shoes, the length, or the color of the outfit, while Zahra talked about the parts of the outfit. Vita followed on from Zahra's idea and, again, helped finding the English equivalents for the Indonesian words that Zahra mentioned, for example, "upper part" and "skirt", except for "brocade" which they looked up to the online dictionary. 7 The picture shows the modern style of kebaya.
Zahra paid attention to the woman's pose, but this time, Vita raised that the outfit is a modern version of a traditional dress. Although initially Zahra questioned Vita's opinion, she finally agreed. 8 When you wear kebaya, you will feel more confident and beautiful.
Vita then thought that the dress would fit the shape of one's body and boost confidence, and based on this account, Zahra wrote the full sentence for number 8. 9 Kebaya can be modified into different colours, styles and materials.
Zahra was considering the color and the make-up before pointed out to the possibility of modifying the outfit. Vita only helped elaborate what Zahra said, but it was Zahra who wrote the full sentence for number 9. 10 It is unusual if you wear kebaya in daily activities.
Zahra contributed fully for sentence number 10.
From the audio-transcripts, it was found that they focused on form once only during the writing process, specifically of sentence number 3. In comparison, they focused on meaning in four occurrences, particularly when searching for the L1 equivalents for words in L1 for sentences number 5 and 6.

Response-based Evaluation: Task 2
For Task 2, the words they picked were "kebaya", "wear", "wedding", "ceremony", "hair", "batik", "confident", "beautiful", "colours", and "styles". On the left column are the sentences they wrote for the princess story. On the right column are the observation notes about the process. Vita started with the phrase "once upon a time" and decided that they needed to introduce a female character. Both were discussing whether to use present or past simple tense until Zahra suggested the use of an apostrophe, which can represent both tenses. Vita then mentioned the name "Isabella", which was deemed too common by Zahra. Zahra recommended the name "Kokom", and Vita did not object to this. 2 She dreamed about meet and marry her prince charming.
Zahra initiated with "she dreamed about marrying a prince". They were then discussing how a typical plot of a princess would be before finalizing the sentence. Vita reminded Zahra that they were required to use all the 10 words they had selected. 3 On the day when she wears batik dress, a prince Zahra threw the idea of a prince coming to a village and looking for a girl wearing a batik dress, leading him to fall in love with her. "Batik" was one of the chosen words. Zahra then added that from Tanjung Sari Kingdom visit her village.
the prince came from far away, and Vita mentioned "Tanjungsari". Zahra convinced her that they needed to say it as a kingdom. Vita wanted to elaborate with the event of the prince falling in love, but Zahra prevented this, probably saving this event for a later sentence. 4 When the prince saw her, he fell in love of the first sight.
Zahra continued from Vita's previous idea. Vita helped correct the grammar.

5
The prince ask her father for marriage and the father said yes.
There was quite a long discussion on the plot, but Vita's idea was finally used.

6
The whole kingdom start to busy in preparing the wedding.
Vita took the lead and wrote the full sentence.
7 They need to decide what colour theme and style they want to use for the wedding.
Vita took the lead again and wrote the full sentence to include the words "colour" and "style". Up to this point, Vita realized that they had not used all of the chosen words. The words left were "ceremony", "hair", "confident", and "beautiful". 8 While Kokom also busy in styling her hair so she will look beautiful on the day.
Vita was brainstorming by herself and wrote the full sentence to include the words "hair" and "beautiful".

9
The day has come, and with confident, she wear her dream kebaya to see and marry her prince charming in the ceremony.
Both participants contributed equally to this sentence and managed to use the remaining vocabulary.
10 And they live happily ever after.
Zahra came up with the full sentence.
Similar to the implementation of Task 1, the occurrence of focusing on form was rare, precisely twice only. Meanwhile, they no longer searched for words in the L2. Instead, they were engaged more in discussions on the plot of their story.
Zahra : This one we can use before the wed... Vita : Iya, ya, ya. The wedding ceremony... Zahra : Because the girl should be style and colourful she should wear at their wedding. Ya. Vita : Ya, ya, ya. But do you think he needs to propose, propose first before the marriage?
There were also long conversations on their strategies to comply with the input requirement and output constraint.

Student-based Evaluation
Based on the interview results, both participants enjoyed doing the tasks but for different reasons. For example, regarding Task 1, the following were their responses.
Vita : It is enjoyable because we explain something that we know about the picture. So, and then, we love to give, especially because the picture is about Indonesian traditional dress, so we love to share the story about the picture. When asked about their preference for the task, Vita favoured Task 1 over Task 2 since the first task provided her with facts in the picture, unlike Task 2. She also found Task 2 more challenging because she needed to make up a story.
Vita : Because we need to make up a story. Mm, I mean there is no fact […] in the story.
And the first task we speci, we just described from the fact that we see.
On the other hand, participant B enjoyed Task 2 more than Task 1 as she herself was interested in princess stories.
Zahra : Because I more, I like the princess story more than the first one. Zahra : It's like more likeable? Ya.
In terms of the task arrangement, Vita preferred doing Task 1 on her own because she could just describe the picture by herself, while she needed help to do Task 2. As opposed to her, due to personal preference, Zahra preferred doing both tasks in a pair group.
Vita : Oh, we're different then because like I enjoy for the first task to be by myself but the second one is I'd like to do it in pair. Because I need an imagination for the second one. Researcher : And why in the first task you wanna do it alone? Vita : Because you can just explain from the picture. Zahra : You don't need help. Vita : Yeah, I mean like you, you can just use whatever you see to make a sentence but… Zahra : You don't ask me the reason.

DISCUSSION
The present study aims to look at the role of fantasy element and constraints in helping adult learners experiment with L2 in doing creative language tasks. The findings previously reported will be elaborated under three major themes.

Imagination and Idea Generation: Task 1
The sentences produced by the participants are relatively short and simple in structure although two sentences are complex sentences (number 8 and 10) and another one is a compound sentence (number 6 despite its grammatical error). A fragment is also found in number 4. During the writing process, the participants attempted at correcting a grammatical mistake, but errors were still found.
In regards to imagination, only five sentences give description based on the picture (number 1, 5, 6, 7, and 9). The rest of the sentences, instead of coming from the physical features in the picture, contain additional information about the outfit, which was largely based on the participants' prior knowledge about the object in the picture. It will be overstating to say that this is an example of imagination. The most sensible explanation can be that the participants simply recalled their existing knowledge or previous experiences related to kebaya.
However, there was an interesting finding regarding sentence number 2 when Zahra suddenly came up with a female name.
Zahra : Err, must call, introducing herself first. When writing it down, Zahra changed the name to "Isyana". At first, this indicated a product of imagination because it has nothing to do with the general idea of kebaya and she herself admitted that she made it up. In fact, in the follow-up interview, Zahra said that when doing Task 1, she remembered the name of an Indonesian female singer who was rising to fame named Isyana Sarasvati. This suggests that what seems to be imaginative was actually based on existing information that she knew in real life.
During the process, the participants had different shares of contribution although they took turn writing the sentences. In this task, after writing the first sentence, Vita handed the paper and the pen to Zahra. This was when she created the system of turn-taking in writing. Vita wrote three out of ten sentences and provided the correct vocabulary when Zahra was searching for words in English. Meanwhile, Zahra's roles included writing seven sentences and adding ideas whenever Vita ran out of ideas. While this sounds collaborative and affording negotiation of meaning, linguistic experiments in L2 were limited. This also happened to the simile task in Tin (2011, p. 230) when meaning was conceived in L1, while L2 was used to deliver the prefixed meaning. It is important to note, however, that the purpose of Task 1 in this study was for the participants to generate pre-inventive forms to be used in Task 2.

Imagination and Idea Exploration: Task 2
It was expected that the participants would face challenges in the second task as the topic of Task 1, which was kebaya, is not directly associated with the topic of Task 2, which was a princess story. However, the words they produced in Task 1 seem to be relevant, except for 'kebaya' and 'batik'. It is not to mention that they purposely chose words they thought were easy to use in writing a story although during the selection, they had not been told that they needed to write a princess story. This was confirmed in the interview.
Vita : Because to the truth that we need to make a story from the ten words, so I think the ten words will be easier for us to make a story. Zahra : Ya, we already imagine what kind of the story we will make. So we think we're gonna use, we're sure that we will use that words.
The story revolves around a girl character, a prince, a kingdom, and the romantic life between the girl and the prince. This plot may remind most readers of Cinderella, Sleeping Beauty, and other fairy tales having similar plots. In the writing process, it was found that Zahra gave the idea of a girl who longed for marrying a prince. In the interview, she commented that she came up with that story idea "because that at some point she [was] going to be married" and "the only one option [so] that she will be a princess [is] that she need[s] to be married [to] a prince because she [was] only a girl." When asked about whether there were stories that inspired them to write their story, both agreed that Disney princess stories did. Vita mentioned Cinderella, while Zahra mentioned Sleeping Beauty and Princess Viona in Shrek.
Furthermore, the story contains the formulaic expressions of "once upon a time" (number 1) and "and they live happily ever after" (number 10). In oral communication, formulaic expressions are asserted to be beneficial for both speakers and listeners as a discourse signal (Wray, 2000, p. 478). It is assumed that the same principle applies in the context of writing. Nearly all fairy tales begin with formulaic phrases referring to time, such as "once" and "one day", and are concluded with a happy ending. Nevertheless, the way the participants' story was ended was also due to Zahra's personal preference.
Researcher : And why is the ending of the story like that? It's a happy ending. Zahra : Ah, it's a normal ending. I don't like if it's sad ending.
There might be two reasons causing them to write a princess story which has a rather typical plot and prescribed expressions. First, the words generated from Task 1, for instance, "wedding" and "beautiful" are somehow related to and enable them to write a girl's love story. Second, it can be due to conformity effect, which is defined as the tendency to "duplicate the features of examples" (Smith, Ward & Schumacher, 1993, p. 839). In this case, Disney princess stories function as examples as well as models in terms of plot and opening and closing phrases. In the process, the participants were trying to follow the conventions in princess stories in terms of story plots, as noted in the following transcript: One distinctive feature that is apparent in their story is its local touch. While most fairy tales are set in Western or unknown faraway lands, their story is located in Tanjung Sari, a factual district in West Java, which is the hometown of Zahra. The girl's name is Kokom, which is a local female name. "Kebaya" and "batik" are mentioned as well as the girl's important outfits. It was participant A who initiated to use local elements in the story. For example, at first, participant B proposed the name "Isabella", but Vita said it was too mainstream and proposed the name "Kokom". Thus, although the story plot is similar to classic tales, the ethnic elements can be considered a product of creativity in regards to its "newness" (Mellou, 1995, p. 102). It is important to note that princess stories, in which a girl meets and eventually marries a lover of royalty, are not widely common in Indonesia's storytelling culture. Hence, the story of a home-grown girl turning into a princess can be regarded new and unusual. Looking back at the writing process, agreeing with Ribot and Baron (1906), the creativity resulted in Task 2 was, again, sprung from the participant's prior knowledge.
Concerning the interaction between the participants, since the beginning of Task 2, they had been more communicatively involved with each other, especially when deciding the storyline as to which event happened before another. They did not focus on form, and word equivalents in L2 were not recalled. For the latter happening, this may mean that they used all the vocabulary that was readily available in their working memory. However, they were actually negotiating meaning as they were trying to use all of the content words they chose from Task 1. This attempt was particularly apparent when they were writing sentences number 7-9. For example, to write sentence number 7, Vita started her brainstorming about a wedding and related with three predetermined words. The above process could be facilitated due to the input requirement of the second task. Vita was trying to fulfill the semantic constraint (Tin, 2011) to fit the meaning of those words into the storyline. A similar process took place when they were writing sentence number 9 and integrate the words "confident", "kebaya", and "ceremony" altogether. Zahra suggested the use of "confident", which Vita related to the way the princess dressed in the story. For "kebaya", they had to modify the concept of wedding dresses in typical princess stories. At first, Vita raised a dream dress and, in a later thinking, changed the word "dress" into "kebaya" although it is uncommon in other princess stories. Meanwhile, the word "ceremony" fitted the story right away.

Individual Differences and Task Preference
Another important finding which needs to be taken into account is that each participant has different personalities and preferences for task features. For example, Zahra was motivated by both tasks due to the freedom in doing the tasks. Nevertheless, the design of both tasks actually presented goal, input, and outcome constraints. Presumably, Zahra did not recognize this because the constraints were inherent in the tasks, and no pre-specified form was required to be used. Her impression on the freedom granted in doing the tasks might also come from the nature of writing itself. Unlike in speaking, written work is generated in reflection, rather than in a sudden moment, allowing a writer to "make choices about vocabulary and style" (McCallum, 2012, p. 103). To compare, in a grammar drilling activity, a learner needs to choose one out of a few pre-specified forms provided, so her/his choice is limited relative to Task 2 in this study, in which the participants had an infinite number of options in arranging the story elements and pre-inventive forms. Finally, the perceived freedom may be attributable to the absence of external constraints (Tin, 2011), such as time limit and the language they could use when completing the creative tasks.
In comparison, Vita seemed to prefer having restricted alternatives in writing as she enjoyed the describing task more than the story writing task. She noted in the interview that Task 2 required imagination, and, thus, she needed help from Zahra. Borrowing from Jung's personality theory, certain learners tend to be sensing and perceive the world based on facts, while some others tend to be intuitive and are more driven by imagination (Wilde & SpringerLink, 2011, pp. 8-9). Corresponding to this, Erhman (1989( in Brown, 2007 lays out assets and liabilities associated with Myers-Briggs character types. For instance, those with the sensing preference can pay attention to detail and give a close observation on something, but they can be hindered by lack of structure in language. Differently, those having the intuition preference may be good at conceptualizing but may miss important details. Nonetheless, Brown (2007) reminds that despite those "natural" tendencies, learners can be successful in learning if they understand their own strengths and weaknesses to optimize the former and make up for the latter.
The implication on creative language teaching is that teachers need to be aware of such differences in designing tasks. The tasks should be stimulating for learners to collaborate with each other. As suggested by Vygotsky (in Sawyer & Ebrary, 2003, p. 82), collaboration can help spur creativity although a successful collaboration can be achieved if, according to John-Steiner (in Sawyer & Ebrary, 2003, p. 82), "long-term engagement, voluntary connection, trust, negotiation, and jointly chosen projects" are present. Moreover, it can be helpful for teachers to present a topic to which learners can personally and/or culturally relate. On Vita's account in the interview, the topic on Indonesian cultures had made her enjoy doing Task 1. This way, hopefully the interaction between the task and the participants can increase, hence allowing more creative and intricate language use (Tin, 2011, p. 227).

CONCLUSION
This study managed to integrate internal constraints and fantasy element into a set of creative language tasks. In keeping with previous studies on creative language teaching, it can be concluded that for a task to be able to foster creativity, it should contain internal constraints.
First, an ill-defined goal will invite learners to finish each step of the task. Second, input requirement, as part of input constraints, will help direct learners towards a particular set of linguistic structures without providing a pre-specified form, which may be seen as restricting learners' autonomy in exploring the L2. Third, outcome constraints can focus learners more on meaning due to the required fulfilment of semantic constraints in the task.
Regarding fantasy, this element inevitably leads learners to imagine at least a fictional character and a fictional setting. However, it should be noted that not all learners will equally find a creative task with a fantasy element stimulating. Regardless of individual differences, though, when learners are presented with a fantasy element, they can start employing imagination to create a make-believe story with a certain degree of originality. In this study, localization of typical princess stories emerged as the product of their creativity. Still, the extent of novelty in their written work will depend on the learners' existing knowledge, which consists of past experiences and memories.
The present study is limited in its number of participants, so the rule of generalization may not always apply to other learners. As such, the factor of individual differences could not be explored further that what was found in the results. Hence, this issue should be resolved in future research. Future studies which would like to replicate the design of this research may also want to provide input requirements which are entirely irrelevant to the topic of the second task in order to see how students use imagination in linking seemingly unrelated ideas while using the L2 in foreign language learning.

ACKNOWLEDGEMENT
This study was based on the author's coursework assignment during her Master's study at The University of Auckland.

AUTHOR'S BIOGRAPHICAL STATEMENT
Lavinia D. W. Araminta obtained her Master of Arts degree from the University of Auckland with a major in Applied Linguistics. She is currently teaching academic English at Universitas Indonesia. She continues working on research projects on the themes of language needs analysis, linguistic self-confidence, and humanities-related topics.