Task and L1 effects Dutch students acquiring the Spanish past tenses

Research on tense-aspect phenomena has shown that the type of experimental task can affect the performance of L2 learners. This pilot study on the understudied language combination Dutch-Spanish investigates this issue by focusing on the interaction between known affecting variables (inherent aspect; L1 effects) and different tasks (multiple-choice, fill-in-theblanks, free production). First findings show that, indeed, both task type and L1 have an influence on the outcome. Generally, Dutch learners seem to prefer the Imperfect over the Preterit. This stands in contrast to previous research but can be explained by the imperfective features of the Dutch Simple Past with which the learners associate the L2 forms. Whereas this L1 effect is not visible in the multiple-choice task where the choice is forced, it manifests itself in tasks where students can choose freely between forms they know. Especially in the free production task, the L1 effect interacts with a high individual variability.


Introduction & background
The acquisition of the aspectual contrast of the Spanish past tenses is known to represent a severe learning challenge for Second Language (L2) learners if their First Language (L1) system does not possess a similar feature (Comajoan, 2014). The main issue concerns differentiating between perfective and imperfective contexts corresponding to Imperfect and Preterit forms (Zagona, 2007). The success of acquiring this property depends significantly on the learners' L1 (see e.g. González & Quintana Hernández, 2018, and González & Diaubalick, in press, for empirical proof ). Especially for speakers of a Germanic language, in which this contrast is not grammaticalized, the acquisition is broadly studied -however, still without a generalization. This may be partially due to a dependence of the findings on the used experimental task. For instance, Salaberry (2003) and Domínguez, Tracy-Ventura, Arche, Michell and Myles (2013) have highlighted the influential nature of the type of task. In this paper, we will present findings of a pilot study among Dutch students, showing interactions between the type of task in the data collection and the learners' L1 in order to understand how these variables affect the use and choice of the verb form in Spanish L2. This way, we can discuss data from a different L1 than in previous studies (mainly English) and both L1 factor and task effects are treated in one study.
Four perspectives on L2 aspect acquisition are considered for the theoretical framework: the Lexical Aspect Hypothesis (Andersen, 1986;Shirai & Andersen, 1995), the Default Past Tense Hypothesis (Salaberry, 2003), the Predication-Effect Hypothesis (González, 2003) and the role of dynamic contrasts (Domínguez et al., 2013). First data indicate that, indeed, both the task type and the participants' L1 play a role. This conclusion encourages future research to combine different tasks and different types of participants (of varied proficiency) in order to obtain a bigger picture on how aspectual interlanguage is represented.

Aspectual contrasts in Spanish and Dutch
2.1 Spanish tense-aspect system 1 There have been different accounts for the semantics of grammatical aspect (Domínguez, Arche, & Myles, 2017;González & Verkuyl, 2017 to name a few in the context of SLA research), and not all languages mark it by the same means. One common approach distinguishes between the notions of perfective and imperfective (Comrie, 1976), which, in the Spanish past tenses, are required to be marked (Zagona, 2007;González, 2003). The three main past tense forms are: Present Perfect, Preterit and Imperfect. The perfectivity/imperfectivitycontrast is found between the Preterit and Imperfect. However, the Present Perfect also plays a role in aspectual distinctions, as exemplified below.
Importantly, in the Spanish past tenses, the marking of grammatical aspect is independent of the lexical meaning and, likewise, of the inherent aspect of the verb predication. Whereas the first term refers to the idea that lexical entries of verbs contain information on their aspectual features (e.g. dynamicity or telicity; see Comrie, 1976;Montrul, 2004, p. 94 for details), the second term highlights the observation that objects of a tenseless predicate can interfere with these notions. A well-treated classification of lexical aspect follows four classes defined by Vendler (1957): states (non-dynamic verbs), activities (verbs without inherent end-points), accomplishments (verbs with such an end-point) and achievements (so-called punctual verbs 2 without a notable duration).
In the following, however, we will focus on the whole predication and adapt the bipartition proposed by Verkuyl (1993). Verbal predicates can be divided into two subclasses (terminative/durative) determined by a compositional function of the properties of the verbs and its arguments (see González, 2013 for details).
In Spanish, every verb has a full paradigm: it is possible to conjugate each verb in all past tense forms. Although especially L2 learners have been observed to establish correlations between the two aspectual levels during the acquisition process (see below), such correlations are not made by Spanish speakers. Instead, the tenses are characterized by the following parameters:

Present Perfect:
(1) He I-have escrito write.participle una one carta. letter 'I have written a letter. '

Imperfect:
(3) Escribía I-write.imp una one carta. letter. 'I was writing/wrote/used to write a letter. ' The Present Perfect in (1) is mostly used in hodiernal contexts, where it expresses anteriority with respect to the present, and focuses on the result of the event. It is more common in Peninsular than in American Spanish (see Verkuyl, 2017 andGonzález, Jara Yupanqui &Kleinherenbrink, 2019, for a description of different varieties). The Preterit in (2) presents an event as a discrete whole at some specific moment in the past (perfective aspect) and is not used in hodiernal contexts such as *Comí hoy (I ate today) in European Spanish. 3 Finally, the Imperfect (Example 3) presents an event in process, unspecified as to its completion. Due to this property, the Imperfect can express several notions: progressive, habitual and continuous aspect (González, 2003;Domínguez et al., 2017). 4 Thus, the concrete interpretation in a given situation does not only depend on the verb form, but also on the sentence context.

2.2
Dutch tense-aspect system 5 As a Germanic language, Dutch is strongly "tense-oriented" (Broekhuis, Corver, & Vos, 2015). This means that (except for the imperative forms), Dutch does not mark mood within the verb, and the progressive can be expressed with the Simple Present or Simple Past. Moreover, futurity does not need to be marked explicitly either. 6 It has, however, two forms of referring to past events: Present Perfect and Simple Past. 7 The distinction between these forms could be understood as aspectual (Borik, González, & Verkuyl, 2004), as the Simple Past is imperfective in nature and the Present Perfect acts as both perfect and perfective, depending on the context. In (4), there is a Simple Past with a habitual (hence imperfective) reading and in (5) the perfect form is used with a perfective meaning 8 (Van Hout, 2005). In 6, however, the perfect has a perfect reading (hodiernal).   González & Verkuyl (2017) defend the idea that the progressive is not an imperfective reading. Yet, for the purposes of this paper we adhere to the more traditional understanding. González & Verkuyl (2017), González & Diaubalick (in press.)

7.
There is also a Past Perfect, which, like the past perfect paradigm of Spanish, does not play a role in this study.

8.
It is noteworthy that, as in other European languages such as French, Italian (Romance) or German (Germanic), adverbial phrases such as gisteren 'yesterday' , referring to temporal intervals preceding speech time (i.e. prehodiernal), are used in Present Perfect constructions. Additionally, Dutch also has a progressive construction, as shown in (7). This periphrasis refers "to some eventuality during speech time" (Broekhuis et al., 2015, p. 151). Although this definition is based on the progressive construction with a present tense auxiliary, it can also be applied to its past tense counterpart.

State of the art
Both the effect of task type and of L1 transfer have been the topic of extensive research in the literature on L2 Acquisition. Consequently, there are numerous studies contributing to the shaping of this paper. However, the focus on the specific language combination Dutch-Spanish in combination with these methodological questions has not been addressed yet. Given the imperfective property of the Dutch Simple Past that sets it apart from English, it is empirically valuable to consider this constellation.
Practically all studies on L2 aspect acquisition refer to the Lexical Aspect Hypothesis (LAH) by Andersen (1986). Following Vendler's (1957) classification of verbs (see Section 2.1), Andersen (1986) claimed that L2 learners first associate the Preterit with achievements and accomplishments, whereas the Imperfect emerges later in combinations with states and activities. Non-prototypical combinations (Preterit with activities/states; Imperfect with achievements/accomplishments) do not appear before advanced learning stages. Even though some subsequent studies questioned the specific nature of the postulated stages, many authors agree that the inherent aspectual information indeed biases the use of past morphology (Andersen & Shirai, 1996;Slabakova, 2001;Montrul, 2004).
Interestingly, most research papers concentrate on L1 English learners (see Comajoan, 2014 for an overview). Obviously, other languages need to be studied as well, if we want to have a complete picture of the interlanguage characteristics. For instance, González (2003) collected free production compositions by Dutch learners of Spanish and showed that they pay attention to phrasal aspectual information (following Verkuyl, 1993). It is thus the predicational (not lexical) aspect which biases their use of past tenses. More recent studies within the LAH framework claim that some consideration should be paid to other single aspectual features such as dynamicity, i.e. input of energy (Domínguez et al., 2013;Kihlstedt, 2015). Salaberry (2003)'s Default Past Tense Hypothesis (DPTH) claims that L2 beginners only use one past tense (usually the Preterit) for all past events regardless of their aspectual features. Its formulation is based on data from English speakers learning Spanish. However, there is still debate on whether the DPTH is universal or L1-based. In fact, it seems plausible that this hypothesis relies on the properties of English (González & Diaubalick, in press). According to Salaberry (2008, p. 120), this is "mostly a consequence of the fact that in English the Simple Past marks only tense, but not aspect, coupled with the claim that learners are relying on the use of general (non-modular) cognitive processes to learn the L2. " Following McManus (2015, p. 159), the hypothesis is in fact an "operationalization of perceived L1/L2 similarity/difference" between the English and the Spanish tense-aspect systems. More research with other L1s is therefore required to see whether the DPTH can be generalised or if it only works with English as L1. Of course, when taking into consideration learners whose L1 already involves a perfectivity contrast similar to Spanish (e.g. French/Italian/Portuguese learners; see Diaubalick & Guijarro-Fuentes, 2019;Diaubalick, 2019), it seems little obvious to consider the DPTH as a universal. By limiting us to learners whose L1 does not involve a grammaticalized perfectivity contrast as the Spanish system does, the open question of which past tense is selected as default still remains. Generally, the Preterit is considered the more evident choice as it is phonologically (accentuation on inflectional suffix) and semantically (reference to events in the foreground) more salient (Salaberry & Ayoun, 2005, p. 21). Nonetheless, it has been observed that in narrations the Imperfect can be generalised, too (Salaberry, 2003). Consequently, the selection of the Preterit is neither univocal nor stable.
Considering that English Simple Past has a perfective meaning while Dutch Simple Past has an imperfective meaning (recall Section 2.2) the results of our study can contribute to this theoretical framework. Do English speakers use the Preterit in Spanish mostly because it is a default past tense or because in English the Simple Past is perfective? By looking at the production of Dutch learners, we will be able to shed more light onto that question since a deviation would deliver important counterarguments against universality. If Dutch speakers use the Preterit more, it means that this form is the default past tense, if they use more Imperfect forms, it means that they rely on their L1. Correlating this to Salaberry's findings, we could say that there is a default past tense, but the chosen form is not standard or universal, but depends on the aspectual meaning of the Simple Past in their L1.
For more arguments for the L1 effect, we refer to two recent studies (González & Quintana Hernández, 2018 and González & Diaubalick, in press). First, González & Quintana Hernández (2018) conducted an experimental study with L1 English and Dutch learners of L2 Spanish performing the same written production task (retelling of a story after watching a muted video). The results show that the use of both the Preterit and the Imperfect by English-speaking learners is biased by dynamicity contrasts (dynamicity triggers Preterit). Furthermore, their overuse of the Preterit to complete the written task shows L1 transfer from the English past tense. Finally, the results of L1 Dutch speakers show that the use of Spanish grammatical aspect is biased by the terminative/durative classification, and that they prefer the Present Perfect in perfective contexts, also showing L1 transfer. Second, by comparing studies on Dutch and German learners of L2 Spanish, González & Diaubalick (in press) show that these differences lead to very different L1 effects. Whereas Dutch learners are sensible to aspectual differences and often consider the inherent aspect of a verb phrase, German learners rather focus their attention on entire lexical elements such as adverbials without considering specific grammatical (temporal/aspectual) features. In both cases, the applied strategies do not lead to a complete acquisition for the contrast, but crucially the way in which acquisitional difficulties are handled differ quite strongly. This means that even minimal differences between L1s can lead to a significantly different outcome. Both Dutch and German belong to the Germanic languages and, thus, share many properties including the absence of grammatical aspect. Importantly, however, Dutch contains some basic aspectual (see Section 2.2) contrasts that the German system does not. In German, an exchange of past tenses may lead to stylistic effects but does not affect the semantics. Applying this reasoning to the DPTH, it is easy to see that also here we may assume a significant L1 effect.
Regarding the type of task affecting SLA studies of aspectual features, we refer to Domínguez et al. (2013). They highlight two main methodological issues that need to be considered when checking the validity of SLA theories in aspectual acquisition (Camps, 2005;Montrul & Salaberry, 2003;Salaberry, 2008 andDomínguez et al., 2013).
1. The inherent aspectual information of a predication in combination with a particular narrative structure can lead to contradicting predictions regarding which grammatical aspect form is more likely to be used (Domínguez et al., 2013, p. 1). 2. The elicitation of data using only production or comprehension tasks prevents the researcher from making adequate comparisons between data sets.
Among others, Slabakova (2001) claims that data elicited through carefully designed experiments are necessary to achieve a full understanding of L2 speakers' competence. As Domínguez et al. (2013) already pointed out, L2 aspect acquisition has not been satisfactorily studied partly because of methodological issues affecting the design of the tasks employed (Camps, 2005;Montrul & Salaberry, 2003;Salaberry, 2008). Studies using combined evidence from different types of tasks are not yet available. These methodological issues are addressed in the present study by making use of different tasks.

Research question and specific hypotheses
The specific question is the following: Does the type of task affect the choice between perfective and imperfective morphology by Dutch L2 learners of Spanish?
The hypotheses are the following: Hypothesis 1: As their L1 does not contain a perfectivity contrast, the students will select a default past tense. Along the lines of the DPTH, this should be the Preterit which manifests itself in the fill-in-theblanks task and the free production task where the two possible forms are not presented explicitly (see Salaberry, 2003). Hypothesis 2: In a free production task, learners do not only have to apply rules of how to use the tenses but they must also extract the specific form from their memory. Based on this reasoning, we assume a forced-choice task to affect the learners' performance, and, consequently, it is in the multiple-choice task where both the imperfective and the perfective forms are going to be used (see Domínguez et al., 2013).

Methodology
This paper reports a small-scale study, for which data have been collected 9 aiming to elicit perfective or imperfective verb forms in Spanish by Dutch L2 learners. Our aim is to (a) understand the choice that students make in different tasks, and (b) to check if the type of task is indeed a relevant factor in the use of aspectual morphology. The study presents production and election tasks that evaluate the students' interlanguage.

9.
We thank Laura ten Dijke, at the time BA student at Leiden University, for the data collection.

Participants
All participants (n = 7) were students in their first year of "Latin American Studies (LAS)" undergraduate programme at a Dutch university. They followed the second compulsory language course, at the end of which students are expected to obtain the A2 level, in accordance to the Common European Framework of Reference for Languages (CEFR). 10 Participants had received language instruction on grammatical aspect two weeks prior to the study and knew the basic rules that are applied according to the Spanish norm (formal grammar instruction). The Spanish taught at LAS is the Peninsular standard variety, containing also a Present Perfect form. It is only later in the curriculum that differences in variation are presented. Moreover, most of the learners attended French class in High School but reported to have lost any acquired competences. A control group, consisting of 10 L1 speakers of different varieties of Latin American Spanish, participated as well. It is noteworthy to mention that generally there is no significant dialectal variation in the perfective/imperfective contrast throughout the Spanish speaking world (Rothman, 2008). Although variation in the past tenses is attested to a high degree, this concerns mainly the contrast between the Preterit and the Present Perfect, i.e., the competing forms in the perfective/perfect contrast described above (Schwenter & Torres Cacoullos, 2008;González & Verkuyl, 2017;González et al., 2019). We do not focus on this contrast in our study.

Experimental materials
We used the following written tasks, based on the experiments of Diaubalick & Guijarro-Fuentes (2017, 2019 and Diaubalick (2019): 11 -Fill in the blanks Participants were asked to fill 27 spaces with the verb form they considered most appropriate. The sentences were part of a coherent story in which 14 terminative and 13 durative predications 12 were combined with different adverbials and in different contexts in order to obtain a balanced number of contexts in which neither the inherent aspect nor appearances of any adverbial dragged the answers into one specific direction. 11. Due to space reasons, a full appendix cannot be given here. The interested reader is referred to Diaubalick (2019) where a commented version of the test instruments is embedded in a larger study on another set of L2 learners.
Mi autor favorito siempre ____(adorar) Madrid, y frecuentemente lo ___(elegir) como lugar principal en sus relatos. 'My favourite author always adored Madrid and chose it frequently as main place for his stories. ' 13 -Multiple-choice The participants had to choose the most appropriate verb according to the context given. The task consisted of isolated sentences, 5 of which featured two verbs, one only one verb and one three verbs (i.e. 14 selections in total). Again, we balanced terminative and durative verbs in order to avoid biases in the data. Pablo {vio/veía} una película muy larga en la tele durante la que se {durmió/ dormía} 'Pablo watched a long TV movie during which he fell asleep' . -Free production In a third part, participants were asked to write a school-related past anecdote. They were encouraged to use certain adverbials like every day, normally, then, etc. to make sure their story remained in the past tense. The text ought to contain 60-90 words.
The choice of these tasks was motivated by varying the grade of liberty participants had in choosing their answer: Whereas in the multiple-choice task learners were forced to choose between the Preterit and the Imperfect (contributing to a total control of the form of results), the fill-in-the-blanks task allowed for more possibilities. Although the text was constructed to mainly elicit the two past tenses, the context was open enough to include other options, such as the Past Perfect, the Present Perfect or even the present tense. To avoid biases, no specific instructions regarding which forms to use was given. The free production task, finally, features the lowest grade of control, and consequently the highest grade of liberty. It is noteworthy to mention that this methodology makes it difficult to code the data for accuracy. Given that the items, due to the nature of the experiment, do not contain a conjugated verb form, aspectuality is not eventually determined. As pointed out, for example, by Salaberry (2008) or Haßler (2016), grammatical aspect is often subjective, but neither arbitrary nor capricious. Consequently, the verb form is the element that eventually encodes the speakers' point of view. Especially for isolated sentences, both forms theoretically make sense. Therefore, 13. Note that the same example was cited in Diaubalick & Guijarro-Fuentes (2017) where a different learner population was tested with the same experiment.
although speakers may have a preference for one form or another, 14 it is impossible to deem the respective other form as wrong or misplaced. In most such contexts, it is unjustified to determine only one "accurate" form. Consider 8 and 9 where both forms are accurate (taken from González, 2003): (8) Alex escribió/escribía una novela ' Alex wrote a novel' (9) Ramon bebió/bebía una cerveza 'Ramon drank a bier' Moreover, we are only aiming at describing interlanguages, without making any assumptions on the accuracy of the uses. It is thus more fruitful to investigate if the different groups have different preferences manifested in different tendencies towards one or the other word form, asking which factors (intra-or extralinguistic) determine these preferences.

Data collection
Before starting the experiment, it was expressed that participants should not overly reflect on their answers, which should avoid the excessive application of explicit grammatical knowledge. The students received instruction in the classroom and through the digital learning environment linked to the language course, on how to fill in the online questionnaire.

Analysis and results
Since, as explained above, accuracy does not serve as adequate target variable, in the following we focus on the effect of exterior variables. Concretely, we will compare in how far the type of task interferes in the choice of verb forms, questioning whether learners and native speakers are biased in the same or in another way. Starting with those tasks with a fixed number of items, i.e., the fill-in-theblanks and multiple-choice tasks, Table 1 shows the obtained numbers of forms. Note that the numbers do not always sum up to a 100% since the questionnaire allowed to leave blanks empty (fill-in-the-blank task) or even select several verb forms (multiple choice task). Additionally, some participants (in both groups) provided forms ignored in our analyses (Present, Past Perfect, etc.).
14. Thanks to an anonymous reviewer for pointing out this thought. Although, arguably, the data set is small, it allows us to perform basic statistical analyses for nominal data. The data sets of the two groups have different sizes and are not homogeneous with respect to the grade of liberty (see above). We therefore opted for conducting separate chi-square tests per task, which indeed revealed some significant results. In the fill-in-the-blanks task, there is a significant difference between the learners and the native control group (as per the Chisquare test performed over the raw numbers; χ 2 (1, N = 392) = 16.1, p < .001), given that the relative part of Imperfect forms was evidently higher in the L2 group. Furthermore, we find a significant difference between the learners' performance when comparing the two tasks (χ 2 (1, N = 239) = 3.85, p = .049). It is evident to see that this is because, unlike what happens in the fill-in-the-blanks task, in the multiple-choice task the learners did not show an unequivocal tendency towards one tense form or another. Apart from these findings, no other results were statistically significant, i.e., neither did we find a significant contrast between learners and natives in the multiple-choice task, nor could we confirm the task type to affect the native speakers' performance.
In sum, the results indicate that, whereas our native speakers do not generally prefer one or the other tense form within both tasks (thus reflecting the balanced test design), the tendency among the learners was towards the imperfective which manifested itself in the fill-in-the-blanks task only.
For the analysis of the remaining free production task, it is important to note that only 5 learners participated. Since they were asked to write a maximum of 90 words (although one participant surpassed the limit considerably), we did not collect enough data to be able to carry out a sound statistical analysis and, thus, are limited to a tentative data exploration. The task revealed some individual differences that complicate the possibility of meaningful calculations: the number of past forms per person ranged between 6 and 16, and the texts were between 53 and 149 words long. Nevertheless, there is a remarkable finding in the data showing the potential of comparing different tasks: In total, the participants produced 33 imperfective forms, in contrast to only 19 perfective forms. Consequently, the tendency to prefer the imperfective grammatical aspect instead of the perfective form confirms the patterns observed in the other tasks. 15

Conclusions
In this pilot study, we sought to find a (tentative) answer to the following research question: Does the type of task affect the choice between perfective and imperfective morphology by Dutch learners of Spanish as L2?
We hypothesized that, due to an L1 effect, learners would use a Default past tense which would manifest itself differently in the several tasks.
As the results have shown, despite the reduced size of the data set, we have uncovered two interesting findings: the task type seems to affect the learners' performance and there is a general tendency towards the Imperfect among Dutch learners. In this sense, the research question can be answered affirmatively. Our hypotheses, however, are not entirely confirmed since, in line with previous studies, we had expected the Preterit to appear more often. This specific prediction (a core element of the DPTH) has not been corroborated by our data. Although we have found certain evidence of a default past tense, it is the Imperfect which is favoured by the learners. Of course, as this paper reported only some basic statistical analyses, we are not able to discard contrary ideas. Nonetheless, our findings converge with the explanations given above, that is, the patterns can be attributed to the L1: different from English, the Dutch Simple Past has an aspectual value closer to the imperfective aspect, which suggests that Dutch students of Spanish resort to their L1 Simple Past, especially when options are not explicit. This observation relates to McManus' (2015) interpretation of the DPTH as operationalization of the L1/L2 differences and is thus, despite our limited data set, congruent with theoretical deliberations (see also Salaberry, 2008). We therefore believe that our analyses point into the right directions and are not mere consequences of random observations. Simultaneously, our experiment has shown that the learners' use of a default tense does not appear globally in all tasks, as it was not visible in the multiplechoice-task. In other words, the task effect (Domínguez et al., 2013) is clearly visible. This finding involves an important consequence for methodological questions: although in a forced-choice task we may not find any evidence of a default, it can still appear in other types of experiments. Such behavior may be attributed to the fact that the L1 factor is still present in this stage of acquisition and 15. Regarding the influence of temporal markers, the interested reader is referred to González & Diaubalick (in press) where more details on these elements are given.
interlanguage. Although the students have received instruction on both forms of grammatical aspect, they resort to the Imperfect as a similar form to the Dutch Simple Past, especially in tasks where they can choose a form more freely. As for the task type, we have shown that, whereas native speakers are not affected by the experimental design, non-native speakers certainly are. This confirms that the choice between aspectual forms is not arbitrary for native speakers but reflects clear decisions based on the sentences' context. In the interlanguage of the nonnative speakers, in contrast, such decisions seem to be considerably less stable, as they do not only depend on inner-linguistic criteria but are heavily affected by the task type.
Although, evidently, our results need to be considered preliminary, they give very interesting indications for future studies. To get clearer insights on the L2 acquisition of grammatical aspect in Spanish, a follow up study needs to include more participants, and additional controls for extralinguistic factors, also in order to allow for more sophisticated statistical analyses (e.g. mixed effect logistic regression). The results of the free production task -despite their numeric limitations -are particularly interesting and call not only for more participants but also for a task where they are asked to produce longer essays. An extra oral task would also shed more light into this discussion as it gives learners less time to reflect augmenting the chance for a default tense to manifest itself (see e.g. Ellis, 2005 for a broad argumentation).
Further research could also benefit from an expansion towards the use of online methods in order to see if the degree of liberty of a tasks affects the processing time, as well as from the exploration of other L1-L2-combinations, possibly beyond the Germanic and Romance languages.