The effect of task-supported and task-based teaching on the use of Russian verbs of motion

: The goal of this investigation is to determine to what extent task-supported (TS) and task-based (TB) language teaching approaches are effective for learning Russian verbs of motion. To this aim, an experimental study was carried out under real classroom conditions as a part of Russian language courses at the University of Barcelona. Participants were organised into two groups that received a TS and a TB treatment, respectively. The implementation of the didactic strategies subsumed in the design of the TB treatment by means of a Learning Management System (Moodle), the benefits and key features of which facilitate task-based instruction, were analysed. The study followed the pre-test, post-test, and delayed post-test design. In light of the results obtained, it is safe to say that the two approaches are beneficial for learning verbs of motion since both groups showed a significant improvement in all the measured aspects, that is, the correct use of verbs both in the oral and written production and in the grammar tests (fill in the blanks, multiple choice, and grammaticality judgment tests). Nevertheless, the TB group performed significantly better than the TS group in the oral tasks and in most delayed post-tests.

Los enfoques por tareas y con tareas en los verbos de movimiento del ruso RESUMEN: El objetivo del presente estudio es determinar en qué medida los enfoques de enseñanza de lenguas con tareas (TS) y por tareas (TB) son efectivos en el aprendizaje de los verbos de movimiento del ruso.Con este fin se realizó un estudio experimental en condiciones de clase reales como parte de los cursos de lengua rusa de la Universidad de Barcelona.Los participantes se dividieron en dos grupos, uno de los cuales recibió instrucción por tareas y el otro con tareas y se diseñó un estudio según el esquema de preprueba, postprueba y postprueba diferida.Las estrategias didácticas se implementaron mediante un sistema de gestión del aprendizaje (LMS), en este caso, Moodle.En el estudio se analiza la idoneidad y necesidad de los LMS en la enseñanza de lenguas actual.Los resultados muestran una mejora significativa tanto en el volumen como en la corrección de los verbos producidos.Esta mejora se da en la producción oral y escrita y en las pruebas de gramática (rellenar espacios en blanco, opción múltiple y juicio de gramaticalidad).Sin embargo, el grupo TB obtuvo resultados significativamente mejores que el grupo TS en la tarea oral y en la mayoría de postpruebas diferidas.

IntroductIon
In the last few decades, task-based language teaching (TBLT) has received considerable attention in second language acquisition research (see Ahmadian & Long, 2020 for references).However, TBLT is not the only approach that has adopted tasks to enhance language learning.Many methodologists and teachers prior to and in parallel to the appearance of TBLT have simply incorporated tasks into traditional language-focused approaches to teaching, using tasks as a possible communicative activity to provide learners with opportunities for oral practice and "free" production, the third P in PPP approaches (Presentation -Practice -Production).One of the most communicative implementations of the PPP teaching approach has been referred to as task-supported language teaching (TSLT).There is a vast amount of studies that support each approach (Ellis, 2003;Samuda & Bygate, 2008, among others).However, most information about the effectiveness of these approaches refers to English language teaching, and the picture is far less complete when it comes to languages with a rich inflectional system such as Russian.Some researchers and pedagogues even suggest that purely task-based courses for elementary Russian without explicit grammar explanations are not feasible (Comer, 2007).
In essence, the aim of this paper is to determine to what extent TBLT and TSLT approaches are effective for learning new complex lexical items such as the Russian verbs of motion (VM) and to investigate whether and how one approach is more appropriate than the other for developing a given skill and promoting the correct usage of the target forms.
Verbs of motion as a group of target forms are an important issue in teaching Russian as a foreign language.This is because they present special features related to form, function, and meaning.More specifically, they encompass all the possible semantic elements of lexical meaning and grammatical aspect of the Russian verb1 .Moreover, VM include an important proportion of irregular, opaque forms.

TBLT versus TSLT
The concept of the task is at the core of TBLT.Skehan (1998, p.95) identified a series of traits most researchers would agree on when conceptualising a task.A task is an activity in which: • meaning is primary; • there is a communication problem to solve; • there is some sort of relationship to comparable real-world activities; • task completion has some priority; • the assessment of the task is in terms of outcome.
Task is the central unit of analysis in all the stages of task-based syllabus designfrom needs analysis to student assessment.As Long (2015) pointed out, "tasks themselves constitute syllabus content, and lessons are built around them.If they appear in the syllabus, it is for a reason, and there is no covert grammatical syllabus" (p.305).
In TSLT, tasks are used to realise a range of curricular objectives: to develop fluency, to raise awareness of specific linguistic features, to assess progress, to provide practice for specific grammar points or to activate prior knowledge.Tasks are usually used in conjunction with different types of pedagogic activity (exercises, rule explanation, focused practice, among others).As said above, TSLT usually, but not exclusively, employs a methodological procedure consisting of PPP.A language item is first presented to the learners with or without an explanation.This item is then practised in a controlled manner through exercises.Finally, opportunities for using the item in free production are provided.It is in this "production" stage that tasks may be employed.However, as Ellis (2003) notes, it would be wrong to characterise TSLT entirely in terms of PPP, since it can take other forms.For example, the sequence can start with the production stage in which tasks would play a diagnostic role.

The evaluation of TBLT
Over the past decade, task-based programmes have been implemented at the national or regional level worldwide, such as Hong Kong, Malaysia, Thailand, Mainland China, Japan, and Belgium (Van den Branden, 2006).While traditional grammar-focused language teaching programmes are still dominant, there is a slow shift towards TBLT programmes.In this changing context, it seems particularly necessary to measure empirically the effectiveness of TBLT as compared with the traditional PPP approach.It is also relevant to pinpoint the challenges practitioners might encounter when implementing TBLT in their classrooms.
The overview of task-based programme evaluations presented in Markina (2019) showed that TBLT has been implemented for teaching different languages (e.g.English, Spanish, Dutch or Chinese) in a variety of foreign language contexts with very young learners (Shintani, 2011), high school learners (Nielson, 2014), andadults (González-Lloret &Nielson, 2015) in face-to-face and online contexts.In all these studies, after the task-based treatment, the learners' progress from the pre-test to the post-test was statistically significant, and they were also capable of successfully performing most of the target tasks.
One of the examples of a detailed evaluation of a full-fledged TBLT program for adults was reported by González-Lloret and Nielson (2015) who assessed a task-based Spanish course for the students of the US Border Patrol Academy.González-Lloret and Nielson found that students in the task-based group performed significantly better than those in the grammar-based group on fluency and structural complexity and performed very similarly in terms of lexical complexity.There are studies that focus on the acquisition of specific target forms or general accuracy in Russian, but only within the TSLT teaching approach (see Gor & Jackson, 2013;Quero Gervilla, 2004), and few attempts were made to apply task-based methodology as a classroom procedure (Comer, 2007) and for evaluation purposes (Long et al., 2012).
A partial evaluation of TBLT in Russian is provided by Comer (2007).Participants in his study were college-level learners of Russian in the United States.The author's conclusions were based on his analysis and interpretation of the lessons which were recorded.In terms of student engagement with the language, Comer evaluated the class as moderately successful.However, the quality of their oral production was lower than expected, because participants had difficulties with matching subjects with correct verb forms.Comer concluded that although TBLT has great potential in getting students actively engaged in using the language in the classroom, there were unsolved problems, such as the lack of needs analysis for diverse groups of Russian learners and the lack of knowledge of the actual developmental stages of learners' interlanguage.Comer's position is shared by a great number of Russian teachers, and it is commonplace in Russian as Foreign Language (RFL) teaching that communicative approaches do not provide good results, especially with regard to accuracy, since they result in the fossilization of errors.Long et al. (2012) conducted another relevant TBLT study on L2 Russian, with the aim of determining whether there is a relationship between the control of linguistic features and the Interagency Linguistic Roundtable proficiency levels (USA).The tasks sampled subjects' control of Russian phonology, morphology, syntax, lexis, and collocations.Statistical analyses showed that 32 of the 33 tasks in the data-collection battery significantly differentiated ILR proficiency levels 2 and 3.The results of this study reveal their potential to discriminate levels.
Another aspect related to our experimental procedure that has been an object of study in the last few decades is the benefits and risks of explicit versus implicit/incidental training in the pre-task.The comparison between explicit and implicit learning, as the extensive literature shows, cannot be unequivocally solved in favour of one or another, rather they lead to inconclusive results.Ellis (2003) and Ellis et al. (2019) claim that prior explicit instruction can interfere with task performance causing learners to prioritise the target structure at the sacrifice of the overall performance.Robinson (1996) and Andringa and Curcic (2015) reported that learners under explicit instruction outperformed learners under implicit and incidental conditions in grammaticality judgment (GJ).In Van de Guchte et al. (2019), focus on language seemed to generate a more accurate use of the targeted grammatical structure, whereas focus-on-meaning promoted a more complex task performance.Denhovska and Serratrice (2017), who examined the acquisition of the gender agreement in Russian, show that participants in the incidental and explicit condition performed similarly in the GJ, but the explicit condition learners performed significantly better in the fill-in-the-blanks (FB).
As seen, the question about the effectiveness of the task-based approach for learning Russian as compared to widely spread TSLT has not been discussed in the reviewed studies, and this also constitutes a goal of this study.

The teaching context
During the last few years, courses within the TBLT approach have been introduced to the teaching practice at the University of Barcelona.The current study was carried out under real classroom conditions as a part of the course "Russian Language II".At the time the experimental treatment began, learners had received approximately 120 hours of instruction.The overall level of learners' language proficiency was between A1 and A2 as measured by internal university proficiency tests.One of the objectives of the project was to empirically support the decision to definitively include or discard TBLT in the course program and promote or not the design of new courses.

Participants
Participants (n=34, 28 females and 6 males) were adult learners of RFL.Their ages ranged between 18 and 25 years old.Most of the participants were Spanish and Catalan speakers.Seven learners spoke only Spanish as their L1, and two learners had L1 Italian and French, respectively.

Research design
Two groups were organised according to two conditions: the TS group (n=15) and the TB group (n=19).First, all the participants performed a pre-test, which was followed by eight treatment sessions, designed for each condition.The general length of the treatment for each group/condition was 14 hours.At the end of each treatment, an immediate post-test was carried out, and a delayed post-test was performed one month later.A control group was not included in the experimental design, as in the language learning context in Spain, the input received by learners outside of class is irrelevant, as shown by the results of the pre-tests.Linguistic skills are mostly developed through exposure to the language in the classroom.
The implementation of the two approaches was possible because data were collected during two different academic years.The TSLT treatment was carried out in the first year, and the TBLT treatment in the following one.In both cases, the instructors were the authors of this article.

Technology-mediated TBLT course
The PPP approach presents perfectly ordered phases as the teacher can control the timing of the processes that learners go through.The teacher decides whether learners are ready to move on to the next stage.Since communicative approaches rely on the transition from input to intake without specific actions by the instructor to explicitly promote consciousness, the amount and quality of input from which the learner must elicit the necessary linguistic information and interaction activities that must facilitate the intake must be greater.This is where Learning Management Systems (LMS) and the information that the internet can provide play a crucial role in the teaching strategy as they facilitate the organization of the didactic units; the management of autonomous activities; the provision of (individualised) feedback for students and the possibility to access an unlimited number of authentic materials.
One of the current tendencies in TBLT is a shift towards technology-mediated language teaching and learning.The meteoric rise of technology has brought some substantial changes to the process of task selection and design.González-Lloret (2015) presents how different technologies ranging from more traditional computer-mediated communication tools such as e-mail, forums, and chat to more cutting-edge technology and applications such as wikis, virtual synthetic environments, or multiplayer online games can be integrated into the task-based classroom.She concludes that "these technologies fit perfectly within TBLT principles of learning by doing, task authenticity, and meaning and goal orientation" (p. 6).
As technological innovations increase efficiency, it becomes more and more difficult to do without them both in the syllabus design and in the actual implementation of courses.Our didactic proposal involves the necessary use of Moodle, one of the most popular LMSs, and some of the teaching resources included in it.This online system meets the needs of organization and successful implementation of TBLT syllabus by offering several options and settings that facilitate exposure to input, design of assessment activities and submission of different types of student output.

Task-based condition
In TBLT, tasks have a specific structure that facilitates the solution to a communicative problem and the achievement of a communicative goal.Before the task itself, it is necessary to ensure that the students can complete it.Language elements necessary for successful task performance are introduced in the pre-task phase.
Although TBLT is a radical implementation of the communicative approach, and, hence, the primary focus is on meaning, and no explicit explanations are provided, it also allows to focus-on-form in a contextualised way.To draw learners' attention to the target forms, different focus-on-form techniques were used.In our course, the texts used in the pre-task included input flooding, input enhancement, and input elaboration.Input flooding is a higher presence of target items without making the text artificial or contrived.Input enhancement consists of flagging target items in different ways, for example, through manipulation of typography and the use of typographic cues (italic, bold face, etc.).Finally, in input elaboration, comprehension of difficult items is facilitated as compared to the authentic, unmodified text.In this way, learners are exposed to target vocabulary and structures while still understanding the text.
In addition, all interaction activities and the tasks themselves were designed in such a way that some target forms were task-useful or task-essential (Loschky & Bley-Vroman, 1993).In other words, there was linguistic material that not only favoured task performance, but also was necessary to complete the task.The length of the pre-task varied from one to several lessons, depending on the scope and requirements of the task.This stage usually includes a lot of textual, audio and video materials, links to online resources, quizzes, etc.
In our course, pre-task materials were divided into three sections: • Reading.The texts provide enough examples so that students can extract the language elements (vocabulary, grammar forms, and structures) necessary to complete the task.• Listening.In all tasks, audio files are flooded with the necessary language elements.
Not only does input flooding help to identify vocabulary and structures important for the task at hand, but it also makes it easier to understand the audio text, since the vocabulary and structures encountered are repeated many times.
Interaction.This part offers practical activities and engages learners in communicative situations that require knowledge of content that is useful for accomplishing the task since the activities that are developed at this point are often part of the communicative requirements of the task itself.
Four tasks were used in the task-based treatment.The instructions given to students and the language structures necessary for the task completion are described below.
You are writing a blog in Russian about tourism in Barcelona (or another place).On your website, you need to write about how to get to the places you recommend visiting.
Task 2. How did you get to...?
You were given instructions on how to get to a certain place.Explain how you got to your destination.
The linguistic resources that learners must operate with are very similar to those posed in the previous assignment.To complete the task, students need to use the same verbs in the same aspectual form (perfective), but in this case, in the past tense.

Task 3. Weekly schedule
Write your schedule for the upcoming Monday, Tuesday, and Wednesday.Your schedule should include all your regular activities for those days, as well as any errands, tasks, or unplanned activities that might arise.Make sure to be as specific as possible.The task will consist of communicating your schedule to your partners and allow for some flexibility since you will need to find space to do some unexpected things together.
In this task, learners must be able to talk about daily repeated actions.This entails the use of the vocabulary related to everyday life actions, the expression of iteration by means of verb forms and adverbs, the expression of time, the expression of usual actions with the present tense and the use of the future perfective for specific actions in the future.Finally, the use of imperfective VM with prefixes is necessary to express multidirectionality as well as the conjugation of the corresponding aspectual pairs, including prefixes: при-/у-/в-/вы-ходить / -езжать, 'arrive/leave/enter/go out'; при-/у-/в-/вы-/по-йти / -ехать 'arrive/leave/go out/go').

Task 4. (I know) what you did last summer
Describe what you did last summer: where you went, when you arrived and departed from each location you visited.What places did you visit in your trip?What did you do in those places?
This task requires the use of perfective verbs when there is a defined time for the completion of a movement, such as 'он пришёл в восемь' ('he arrived at eight'), and imperfective verbs expressing the concept of 'being in / visiting a place': Летом я ездил в Москву, 'In the summer, I travelled to Moscow').
After completing the tasks in pairs or small groups, learners were given post-task activities to work on important aspects.Moodle was used to upload various types of work, such as oral monologues, essays, or interactive exercises related to the task.In this final stage, the teacher when asked to focus on language units (grammatical forms, structures, expressions, and individual words) that were difficult or used incorrectly by students during the task.

Task-supported condition
The TS treatment (eight sessions in total) followed the PPP sequence with two tasks (Task 1 and Task 3) used as the communicative practice activity.The lessons were mostly teacher-centred.Sets of grammar forms were the main units of the classroom activities, and they were sequenced according to their presentation in the course book.The instruction involved what Long (1991) called focus-on-formS.New language material was first explicitly presented and explained by the teacher.Then target forms were practised by means of various exercises (FB, transformations, asking and answering questions following a model, etc.).The decontextualised exercises were taken from Castellví et al. (2001).Finally, participants were expected to produce these forms while performing oral tasks (only the task-cycle phase) which were used to provide learners with the opportunity to practise the "forms of the lesson" -the VM with corresponding prepositions -in a meaningful context.

Instruments
In the pre-test, post-test, and delayed post-test of the two conditions, participants were asked to complete the same written task, oral task, and grammar tests.
The written task consisted of writing an email to invite a Russian friend to a birthday party.Students had to explain how to get home from the nearest bus stop or metro station.The oral task was a map task.Learners were asked to give street directions by looking at the map.Three types of grammar tests were selected in order to gather information about the learners' linguistic knowledge according to their capacity to recognise the right form (multiple choice (MC), 20 sentences), identify a grammar error and solve it (grammaticality judgment (GJ), 20 sentences), and retrieve linguistic information in the given context (fill in the blanks (FB), 30 sentences).

Data collection procedures
Written data were collected in the lesson time (pre-and post-test) and at the end of the term during the final exam (delayed post-test).Participants had 90 minutes to complete the written task and tests and were not allowed to use any learning materials.
Oral data were collected individually.Instructions to perform the task were given in Russian or Spanish/Catalan if needed.Participants were given as much time as they needed to prepare, and the overall duration of the oral performance was 2-4 minutes.

Data analysis
In our project, we analysed lexical and syntactic complexity, accuracy, and fluency (Markina, 2019), but in this paper we will focus on accuracy because of space constraints and because one of the aims of this article is to confirm or refute that the complex set of features of Russian VM can be learned within both approaches.
For each of the grammar tests, FB, MC, and GJ, the number of correct answers was counted.One point was given to each correct answer, and zero points were assigned to incorrectly answered or unanswered items.The percentage of the correct answers was calculated for each test by dividing the number of correct answers by the total number of items and multiplied by 100.
To analyse learners' oral and written production, the total number of VM produced and the number of accurately produced verbs were calculated.

Statistical procedures
In order to account for the effectiveness of TS and TB treatment, a number of statistical analyses have been performed.First, during preliminary analyses, the data were screened for outliers and the normality of distribution (Shapiro-Wilks test); the assumptions of sphericity (Mauchly's sphericity test) and of homogeneity of variances (Levene's test) were checked.
When data were normally distributed and met the two assumptions, a mixed between-within subjects ANOVA was performed.When one of the assumptions was violated but the scores followed a normal distribution, parametric tests (repeated measures ANOVA, multivariate tests, and paired samples t-tests to measure within group differences and independent samples t-tests for between-group differences) were used.The pairwise comparisons within each group were adjusted for multiple comparisons (Bonferroni correction).
The data which were not normally distributed was analyzed by means of nonparametric tests (Wilcoxon Signed-Ranks test within groups and Mann-Whitney test between the two groups).Following Field (2009), the effect sizes were reported as partial ŋ2 (partial eta squared) for ANOVA and as r (Pearson's correlation coefficient) for other tests 2 .

results
The results are organised as follows: the total amount of produced VM, the target-like use of VM, and, finally, grammar tests.
The mean scores in all the pre-test and delayed post-tests deserve a preliminary comment.As expected, both groups scored low in the pre-tests, since in the Spanish learning context, students rarely have access to any input in Russian outside their classes.In the delayed post-test, instead, scores are surprisingly high, often higher than in the immediate post-test.This is due to the study being conducted as part of a regular course at the University of Barcelona.The delayed post-test was actually the final exam, worth 35% of the course grade, and, for this reason, the information that follows from the delayed post-tests does not inform about the competence or knowledge that remains after a certain period, but to what learners retain going through what they have been working on over the course.

Total amount of verbs of motion in written and oral production
The results of descriptive statistics showed that the use of VM by learners in both their oral and written production improved.Figures 1 and 2 illustrate that the growth in the number of target verbs produced by participants of both task-supported and task-based treatments in the post-tests, and that TB learners produced more VM than learners in the TS group both at the immediate and delayed post-tests.In the written production, paired samples t-tests showed that there was a significant difference between pre-test and immediate post-test scores both in the TS (p=.001, r=.67) and in the TB group (p=.000, r=.78), and non-significant difference between the immediate post-test and delayed post-test.
In the oral production, learners significantly increased the use of VM from the pre-test to the immediate post-test in both treatments (TS: p=.003, r=.82; TB: p=.000, r=.88), an there was a statistically significant difference between the immediate post-test and delayed post-test scores in the TS group (p=.009, r=.72), but not in the TB group (Wilcoxon signedranks tests).
As for the differences between approaches, there were no statistically significant differences between the TS and TB groups at the pre-test, but in immediate post-test and delayed post-test the TB group significantly produced more target forms than the TS group (Independent samples t-test, p=.000, r=.64, in the immediate post-test, and p=.012, r=.45 in the delayed post-test; Mann-Whitney, p=.088).

Target-like use of verbs of motion in written and oral production
Results obtained for the accurate use of VM in the written and oral tasks are similar to those reported in the previous section.There was an improvement in scores from the pre-test to the post-test and delayed post-tests in both treatments.While both groups improved their use of verbs of motion, the TB group scored higher than the TS group both in written and oral immediate and delayed post-tests (Figures 3 and 4 for written and oral production, respectively).

Figure 3. TLU of VM in written production Figure 4. TLU of VM in oral production
Wilcoxon signed-rank tests showed that in both the written and oral tests there was a significant difference between the pre-test and the immediate post-test scores within both the TS group (p=.018, r=.63 in the written test, and p=.005, r=.78 in the oral test) and the TB group (p=.000, r=.86 in the written test, and p=.000, r=.88 in the oral test).
The between-group comparisons scores obtained in the written task showed that there were statistically non-significant differences between the TS and TB groups in the pre-test and in the written immediate post-test, large significant differences in the oral immediate post-test (p=.044, r=.51), and moderate significant differences between groups in the delayed post-test (p=.022, r=.41).
In summary, both task-supported and task-based treatment led to statistically significant improvement in learners' use of VM in their oral and written production.However, the task-based treatment proved to be more effective for learning verbs of motion than the task-supported treatment.

Verbs of motion in grammar tests
Students performed three grammar tests: FB, MC, and GJ.The results of descriptive statistics indicate that both the TS and TB group showed constant improvement of their results from the pre-test to delayed post-test.The dynamics of the two groups over time are illustrated in Figures 5, 6, and 7.

. Fill in the blanks
The mixed between-within-groups ANOVA showed that there was a significant large main effect of Time (p=.000, partial ŋ 2 =.741), which means that there was significant progress over time, but the analysis of between-group effects showed a statistically non-significant effect p=.810, ŋ 2 =.002, which suggests that task-supported and task-based treatments did not significantly differ in their effect on learners' scores.
The changes in scores within each group over time were similar in the two groups as proven by paired samples t-tests.There was a statistically significant difference between pre-test and immediate post-test scores (TS: p=.000, r=.72; TB: p=.000, r=.77) and between pre-test and delayed post-test scores (TS: p=.000, r=.85; TB: p=.000, r=.83).In the TB group, there was also a significant difference between the immediate post-test and delayed post-test (p=.014, r=.36), but not in the TS group (p=.207).
Taken together, statistical analysis of FB scores showed that both the TS and TB groups significantly improved their use of verbs of motion over time.The treatment type did not significantly differ in its effect on learners' performance.

Multiple choice
The results for the MC tests suggest that there was a significant change in MC scores over time in both groups of participants with a significant main effect in time.
There was a significant, long-lasting effect of the two types of treatment on learners' MC scores as indicated by the large significant difference between the pre-test and the two post-tests both in the TS group r=.47;r=.61)and in the TB group r=.92;r=.92).
The differences in MC scores between the two groups were significant in the pre-test (p=.003, r=.47) and delayed post-test (p=.013, r=.45), but not in the immediate post-test.
To sum up, participants in both groups significantly improved their MC scores.The between-group comparison showed that learners in the TB group obtained significantly bet-ter scores in the delayed post-test when compared to the scores obtained by the TS group although the TB showed significantly worse results in the pre-test.

Grammaticality judgment
As the results of ANOVA and multivariate tests show, there was significant progress for both the TS (p=.000, ŋ 2 =.507) and the TB group (p=.000, ŋ 2 =.822).
As well as in other grammar tests, in the pairwise comparisons, the TS and TB groups showed significantly better results as compared to the pre-test, in the immediate post-test (TS: p=.005, r=.54; TB: p=.000, r=.67) and the delayed post-test (TS: p=.003,r=.57;TB: p=.000,r=.79).
Independent samples t-tests showed that there were no significant differences between the two groups' scores in the pre-test and immediate post-test.However, differences in the delayed post-test between the groups were significant (p=.021, r=.41).
Taken together, learners' results in both groups are significantly better at the immediate and delayed post-test than at the pre-test, and the TB group significantly outperformed the TS group in the delayed post-test.

Summary of results
As we have seen, learners in both groups showed a significant progress in the total amount and in the accuracy of VM produced, both in the oral and written performance, and in the three grammar tests.
Differences between the TS and TB groups depend on the measure, as summarised in Table 1.In this table, the equals sign = is used when statistical tests showed no significant difference between the groups, while the greater-than sign > corresponds to significant differences.

dIscussIon
The results reported in the present study confirm the idea that both the TS and TB approaches create the conditions for learners to use more verbs of motion and in a more target-like way.The improvement after the TS treatment was expected as PPP/TSLT is the predominant approach in Russian teaching and has been constantly refined over the last five decades.Despite differences in how the approaches expose learners to new knowledge, both treatments resulted in significant and consistent improvements in all tests, both immediate and delayed.
As far as the comparison of treatments is concerned, in the oral task, the TB group produced significantly more verbs of motion than the TS group did in the immediate and delayed post-tests.The analysis of the written task, contrastingly, showed similar results for the amount of VM produced in both treatments with no significant differences, though the TB group tended to produce more verbs of motion than the TS group.This increase in the amount of VM produced over the instruction period is accompanied by the increase in the amount of accurately used target forms in both approaches.This is not in line with Ellis et al. (2019) who claim that pre-task explicit instruction can lead to more frequent use of the target structures, on the one hand, and to the detriment of global complexity, accuracy, and fluency, on the other hand.
The grammar tests deserve special mention.In line with Long (2015), some scholars advocate that, in TBLT, assessment must be carried out on the basis of task resolution.However, we decided to assess not only the outcomes in learners' performance in oral and written tasks, but also grammar tests (FB, MC, and GJ) after both instructional treatments, as we considered it unfair to evaluate the progress of TBLT learners solely based on traditional grammar tests, just as it would be unjust to assess the progress of PPP/TSLT students exclusively on their task performance.Our research revealed significant positive results in the grammar tests, with both treatments yielding a similar level of grammar competence, irrespective of the assessment method employed.In the immediate post-tests, there were no significant differences between the two groups.As for the delayed post-tests, the TB group significantly outperformed the TS one in the MC and GJ tests.The results regarding the MC tests are especially interesting, since the TB group performed significantly worse than the TS group in the pre-test, while in the immediate post-test, both groups had similar results, and in the delayed post-test the TB group significantly outperformed the TS group.
In the TB pre-task, participants incidentally encountered the VM that were necessary to complete the tasks, while the TS group received an explicit explanation of those necessary VM.In our study, both explicit and implicit/incidental treatments yield similar results with a certain advantage for the latter, which is not completely consistent with the studies on explicit versus implicit/incidental training in the pre-task carried out in the last few decades.On the one hand, our results are only partially consistent with Ellis (2003) and Ellis et al. (2019) since prior explicit instruction did not lead learners to prioritise target structures at the sacrifice of the overall performance.On the other hand, based on Robinson (1996), Andringa and Curcic (2015), Van de Guchte et al. (2019) or Denhovska and Serratrice (2017) we could expect that learners would perform similarly or better under explicit instruction in grammar tests.
One possible explanation for the difference between our results and those reported in other studies may lie in the duration of the instructional treatment.Our results are in line with Doughty (2003) who claimed that short-term instruction periods could result in a bias.In the above-mentioned studies, the length of the training period was rather short: from 15 minutes of the instruction time in Denhovska and Serratrice (2017) and 30 minutes in Van de Guchte et al. (2019) to a 2-hour instruction treatment in Ellis et al. (2019).
We used the pre-task, task-cycle, and post-task framework for our TBLT treatment, without explicit explanations during the pre-task or task-cycle phases.Instead, we focused on providing enough comprehensible input and exposure to new information essential for task completion to facilitate attention-drawing and elicitation of this relevant linguistic information.This required careful task design and the use of focus-on-form techniques (see Gilabert & Castellví, 2019).Tasks typically lasted 3-6 classroom hours, with at least twothirds of that time devoted to the pre-task phase.The task-cycle phase was usually less than a 2-hour session, and most post-tasks were given as homework.Our task design process involved analysing the necessary language, comparing it with learners' existing knowledge, and providing enough input to elicit and notice the new language constructions and lexical units.In the pre-task, apart from the focus-on-form techniques, short activities implying the use of the target language were included with the aim to ensure not only the comprehension, but also the capacity to use the target language.Consequently, pre-tasks took on average twice the time of the task-cycle phase.

conclusIon
In this paper, we have presented the results corresponding to communicative situations demanding the use of verbs of motion.We have observed that both TB and TS approaches are appropriate for teaching Russian at the beginner/waystage level, and that the TB group outperformed the TS group in several aspects.Therefore, our results undermine the idea that communicative, meaning-based approaches are less suitable than "focus-on-formS" approaches and that difficult aspects of morphologically complex languages cannot be learned unless explicitly.Learners engage in the resolution of a task (focus-on-meaning) while paying attention to new forms (focus-on-form).
Nevertheless, one could expect better results in oral production for the TBLT approach, while the TSLT approach should provide better results in the target-like use of language in written production and grammar tests since students were trained to specifically target predetermined linguistic features.This is confirmed in the oral task, partially confirmed in the written task and refuted in the grammar tests.TS learners were expected to have an advantage in grammar tests as this type of exercises was used in their instruction, but in general terms, the TB group outperformed the TS group.Differences between both groups were significant in the GJ and MC test, but not in the FB test.In other words, expectations are not completely met, which prompts numerous reflections and pedagogical implications, especially for TBLT.For example, that learners can obtain declarative knowledge through noticing and consciousness-raising activities or that TBLT courses can train learners for standardised tests as any other language course.Moreover, if we intend to implement TBLT in regulated courses forming part of larger programmes that include other PPP/TSLT courses, the level of competence must be ensured in every course, and students must be prepared for official assessments regardless of the teaching approach.International standardised certification exams usually do not align with TBLT, so it is fair and even necessary to use the same assessment method for TBLT students as other courses.
In addition, our research shows that the formal complexity of the language taught should not impede applying the TBLT approach, but it does affect task design.Furthermore, while important advances in studying linguistic complexity have been made, we believe that further research should be conducted, and linguistic difficulty should be included in the model of task complexity in order to develop more formal tools to design tasks for languages with a rich derivational system. 3

Figure 1 .
Figure 1.Total VM in written production Figure 2. Total VM in oral production