1. Introduction
The production of LREs, defined as “any part of a dialogue where the students talk about the language they are producing, question their language use, or correct themselves or others” (
Swain and Lapkin 1998, p. 326) and thus considered learning in progress (
Swain and Watanabe 2013), has been the target of several investigations with adults (i.e.,
García Mayo and Azkarai 2016;
Leeser 2004;
Niu 2009;
Payant and Kim 2019;
Ross-Feldman 2005,
2007;
Storch and Aldosari 2013) and more recently, with young learners (i.e.,
García Mayo and Imaz Agirre 2019;
Martínez-Adrián et al. 2021), defined as children up to 12 years old (
Pinter 2011). In this last respect, the latter are receiving increasing attention in the literature on the grounds that children are unique and approach the language learning process in different ways from adults (
Oliver and Azkarai 2017). In addition, ‘childhood cannot be considered as a homogeneous period; on the contrary, children develop quite fast socially and cognitively and show very different characteristics as they grow older’ (
Lázaro-Ibarrola and Azpilicueta-Martínez 2015, p. 3).
This paper in particular will narrow the scope of meaning- and form-focused episodes by providing a detailed subclassification in terms of the features targeted in them, as well as by delving into the elaboration of LREs. As argued by
Zhang and Plonsky (
2020), studies examining LRE production have mainly targeted the focus of the LREs (e.g., grammar, lexis, mechanics), to a lesser extent the resolution of the LREs, and sporadically the nature or quality of the LREs. “LREs of varying qualities represent different degrees of attention to the gap between learner language and the target gap, and thus are reflective of L2 learning opportunities of different qualities” (
Zhang and Plonsky 2020, p. 13). In this respect, more turns per LRE have been suggested to be an indication of more engagement (
Storch and Wigglesworth 2010).
This paper is organized as follows.
Section 2 and
Section 3 review studies that have specifically examined meaning- and form-focused LREs, as well as the elaboration of the LREs.
Section 4 addresses the research questions of the study, while
Section 5 presents the methodology of the study. Results are offered and discussed in
Section 6.
Section 7 concludes the paper.
2. Target Features in LRE Production
In general, learners focus on lexical items over other aspects of language (
Loewen and Sato 2018). In the observation of adult intensive English as a Second Language (ESL) classes,
Williams (
2001) attested that more than 80% of LREs were lexical rather than about morphosyntax. This could be accounted for by the fact that nouns carry more important value when negotiating for meaning (
Spada and Lightbown 2008).
Other studies exploring the relationship between participants’ stimulated recall comments and LRE features have shown how adult learners notice morphosyntactic and lexical LREs equally (
Egi 2007;
Loewen 2019), while in other studies learners were more likely to report noticing lexical, semantic, and phonological items rather than morphosyntactic ones (
Mackey et al. 2000).
Those investigations that have specifically examined the effect of task- modality on LRE production by comparing oral with oral + written tasks indicate that adult learners produce and resolve a greater number of LREs in those tasks that combine oral and written modalities (
Adams and Ross-Feldman 2008;
García Mayo and Azkarai 2016;
Niu 2009;
Payant and Kim 2019). In addition, they also produce more form-focused LREs in oral+ written tasks (
García Mayo and Azkarai 2016;
Payant and Kim 2019). A written component also seems to lead to more accurate resolutions (
Payant and Kim 2019). Even if compelling evidence exists in studies carried out with adults on LRE production, very little research has been conducted on the specific targets of LREs. In the study conducted by
Niu (
2009) with Chinese university learners of L2 English performing a reconstruction task collaboratively, a greater number of LREs were produced in the case of written-output pairs than in oral-output pairs, as well as when lexis-focused, grammar-focused and discourse-focused LREs were considered separately. As regards the specific targets of LREs, oral-output pairs and written-output pairs focused their attention on similar aspects such as word pronunciation, spelling, word stress, minimal pairs, parts of speech, word choice, word use and word search. In contrast, written-output pairs focused on a wider range of features in both grammar-focused and discourse-focused LREs. While oral-output pairs and written-output pairs centred their attention on pronouns, word order, prepositions, tense, verb forms, possessive forms and articles, written-output pairs additionally focused on usage of verbs, plural forms, the choice of prepositions/present participles, gerunds, and choice of affirmative and negative forms.
As in the case of adult learners, oral tasks performed by children lead to more meaning-focused LREs (
Pladevall-Ballester 2021) whereas oral+written tasks result in a greater number of LREs and form-focused LREs (
Gallardo-del-Puerto and Martínez-Adrián 2022;
Gutiérrez-Mangado and Basterrechea, forthcoming;
Martínez-Adrián and Gallardo-del-Puerto 2021). However, unlike adults, the LREs produced in both modalities are not so elaborated, as young learners tend to resolve those episodes without further justifications or with scant metalinguistic explanations (
Gallardo-del-Puerto and Martínez-Adrián 2022;
Martínez-Adrián and Gallardo-del-Puerto 2021). They also tend to incorporate the accurate resolutions of their LREs in the final product of oral + written tasks (
Gallardo-del-Puerto and Martínez-Adrián 2022;
Martínez-Adrián and Gallardo-del-Puerto 2021). Those investigations have also revealed that the possibility of revision and editing of the final product boosts more accurate resolutions irrespective of an oral or written output (
Martínez-Adrián and Gallardo-del-Puerto 2021). Hence, task-modality effects are evident in the focus of LREs and in the incorporation of target-like resolutions in the final product. This result mirrors the findings obtained in other studies conducted with primary-school learners to whom a dictogloss (written) task was administered (
Calzada 2021;
Calzada and García Mayo 2021). In this case, the dictogloss task was equally effective at drawing children’s attention to grammar and mechanics, but not to the targeted feature selected. Despite the fact that studies on LRE production by young learners have grown steadily in the last years, to our knowledge, there is a lack of investigations offering a more comprehensive analysis of the targets of LREs (see
Niu 2009, for adult learners)
3. Elaboration of LREs
As aforementioned, the examination of the nature of LREs (i.e., level of engagement of the learners during the LREs; elaboration of LREs) has received little attention in the literature. In addition, elaboration of LREs has been codified in different ways.
Storch (
2008) used the term level of engagement in LREs. She distinguished between elaborate LREs (E LREs) and limited LREs (L LREs). E LREs are those episodes in which the pair joins in the discussion, deliberates or provides alternatives. In contrast, L LREs refer to those episodes in which there is no deliberation, discussion or suggestion by one of the peers; or by any of members of the dyad. In
Niu (
2009), LREs were also classified in terms of range of turns so as to examine elaboration.
In the studies conducted with adult learners in ESL settings while performing collaborative writing tasks such as text reconstruction, text composition and report, Storch and colleagues (
Storch 2008;
Storch and Wigglesworth 2010;
Wigglesworth and Storch 2012) attested an elaborate level of engagement. This level of engagement positively correlated with language development and higher uptake. Along the same lines as
Storch (
2008),
McDonough and González (
2019) examined lexis-focused LREs, their level of engagement, resolution and the rate of incorporation of those LREs in collaborative writing tasks. Unlike previous studies with adults, the vast majority of LREs (65%) were simple LREs, which could be explained in the light of a goal-oriented approach that students may take, which contrasts with how teachers exploit language learning opportunities by encouraging students to elaborate their discussions. Among the 12 words that emerged during LRE production and that were subsequently incorporated in the texts, only 3 words occurred in negotiated (elaborate) LREs.
In Foreign Language (FL) settings, L2 learners who show a collaborative pattern during dyadic work tend to produce more LREs (
Edstrom 2015) as well as more E LREs (
Edstrom 2015;
Azkarai 2015;
García Mayo and Azkarai 2016). Other studies in FL settings have confirmed an interaction between task-modality and level of engagement. In the study conducted by
Niu (
2009) with Chinese adult learners of L2 English, the written output pairs, in line with the greater number of LRE turns, produced more extensive discussions about the language features they focused on, while the oral output pairs were found to simply make the choice.
Very recently, level of engagement has started to be the focus of investigations with young learners in FL settings (
Azkarai and Kopinska 2020;
Calzada 2021;
Calzada and García Mayo 2021). The study by
Azkarai and Kopinska (
2020) with 6th grade primary school learners performing a collaborative dictogloss task targeting the third person singular
-s morpheme showed that the E LREs were prevalent in all the pairs. In the study by
Calzada and García Mayo (
2021) with 6th grade learners to whom a dictogloss task focusing on the same morpheme was administered, the elaboration of LREs was measured in terms of the number of turns. Despite the fact that this task generated more LREs about grammar than vocabulary, lexis-focused LREs were lengthier. This finding also aligns with
Calzada (
2021) where more E LREs were obtained in the case of lexis-focused LREs than in grammar-focused LREs and mechanics-focused LREs.
Even if level of engagement/elaboration of LREs is receiving increasing attention in the literature on young learners, research is still in its infancy. In addition, the interface between task-modality and level of engagement/elaboration of LREs is clearly an under-researched topic in this population.
6. Results and Discussion
Results are presented and discussed according to the four research questions of the study. The analysis regarding task-modality differences in the production of LREs (see
Table 1) indicated that differences were greater for form-based LREs than for meaning-based LREs. As for the latter, no significant differences were found in the ‘overall’ and in the ‘word-meaning’ categories, even though in the case of the category ‘word choice’ a marginal difference with a small-to-medium effect size was reached in favor of the OW task. With regard to form-based LREs, the statistical analyses (Wilcoxon signed rank test) revealed that this type of LREs were significantly more frequent in the OW than in the ED task. This was particularly true for the LREs which focused on ‘prepositions’ and ‘spelling’, which rendered no occurrences in the OE task. Effect sizes were of a medium-to-large magnitude for these contrasts. No statistical significance was reached, however, for the categories ‘phonology’ and ‘morphosyntax’. Examples of the various LRE categories from our database are provided in
Table 2.
The results regarding the production of LREs indicated that it is the written modality that boosts the production of LREs, as compared to the oral modality. As in previous research with adults, the task combining an oral and a written component led to more language attention. It is particularly revealing to note that the superiority of the oral+written modality is found in our study, as the oral-only task is accompanied by an editing phase which was not present in the oral tasks in prior research (
Adams 2003,
2006;
Adams and Ross-Feldman 2008;
Niu 2009;
Payant and Kim 2019). It is also observed that both task modalities fostered meaning-based LREs mainly, though to different degrees. We discovered a similar pattern as in
Niu’s (
2009) study with adults, in which no differences were attested between oral-output pairs and written-output pairs in terms of meaning-based LREs. However, as will be reported in the intra-task analyses below, this pattern is not entirely equal. With regard to form-focused LREs, in spite of being less frequent than meaning-based LREs, the impact of task modality was stronger, particularly for spelling and prepositions. This particular focus on prepositions goes in line with how adult learners have behaved in previous research (
Niu 2009). The enhanced attention to spelling has also been recently observed in children performing a written task, namely a dictogloss task (
Calzada 2021;
Calzada and García Mayo 2021).
As for the second research question,
Table 3 presents the results of the inferential statistical analyses performed to discover which targets are more common in meaning-based and form-focused LREs in the OW task. Regarding the former, a t-test was computed to compare ‘word meaning’ and ‘word choice’ LREs. These two categories yielded a similar mean number of LREs, no statistical significance being attested. As for form-based LREs, Wilcoxon signed rank tests indicated that differences reached statistical significance when ‘phonology’ was involved in the binary comparison, as the production of this target was significantly lower than the production of any of the other form-based LREs, namely ‘morphosyntax’, ‘preposition’ and particularly ‘spelling’. Effect sizes were of a medium and medium-to-large magnitude in these cases. In addition to this, a marginal difference with a small-to-medium effect size was discovered when the two last targets were compared, ‘spelling’ being more abundant than ‘preposition’.
Regarding the OE task (see
Table 4), the Wilcoxon signed rank tests pointed out that ‘word-meaning’ targets were significantly more common than the meaning-based LREs focusing on ‘word choice’, a medium effect size being discovered for this contrast. The binary comparisons in the case of form-based LRE targets, however, did not reveal significant differences, even though some marginal differences with a small-to-medium effect size were found. The lack of LRES targeting ‘spelling’ and ‘prepositions’ seems to make these targets distance themselves from the categories ‘phonology’ and morphosyntax’. The higher mean number of ‘morphosyntax’ occurrences, however, did not statistically differ from that the lower mean number of ‘phonology’ LREs.
All in all, intragroup analyses for meaning-based LREs have thus revealed two clear tendencies which might be related to learners’ age if we take into account
Niu’s (
2009) results for adults where we can observe a slightly different trend—the OW task seems to be more oriented towards word choice whereas the OE task seems to focus on word meaning further. This pattern could be ascribed to modality differences regarding the level of control of processing, which must be greater in oral tasks given their inherent immediacy (
Bialystok 2001), or of pace, which is typically more self-determined in written tasks as learners can better decide on when to concentrate on the planning phase or on word selection (
Kuiken and Vedder 2012). As for form-based LREs, the data clearly pointed to a focus on morphosyntax in both modalities. In addition to this, the OW task promoted more attention to spelling, as in the children in
Calzada (
2021) and
Calzada and García Mayo (
2021), whereas the OE task resulted in a focus on pronunciation. This result is not in agreement with
Niu’s (
2009), since both oral-output and written-output adult dyads tackled pronunciation issues similarly in a text-reconstruction task.
Table 5 displays the results that enable us to find out whether there are any differences in terms of the elaboration of the LREs. With regard to inter-task differences, we observe that it is in the OW task that dyads produced the LREs that contained a higher number of turns, with the exception of form-based LREs focusing on phonology, which displayed a wider range of turns in the OE task. The greater elaboration observed in the OW task aligns with the results obtained in previous research with adult learners (
Niu 2009). The submission of a written product, which makes the OW task more visual, might prompt the learners to engage with the language to a higher extent than in the OE task. In turn, in the light of the positive correlation attested in the literature between a higher level of engagement and language development (
Storch 2008;
Storch and Wigglesworth 2010), we could claim that the OW task might generate more and better learning opportunities.
Table 5 also allows us to look into intra-task comparisons. As for the OW task, we observe that although the maximum number of LREs per dyad is the same in both meaning-based and form-based LREs, the range of turns is clearly wider for meaning-based than for form-based LREs. Besides, the range of turns is wider in the meaning-based LREs which target ‘word choice’ than in those focusing on ‘word meaning’. This tendency is also mirrored by the maximum number of LREs per dyad, which was higher for ‘word choice’ than for ‘word meaning’. Regarding form-based LREs, the range of turns for ‘spelling’ was the highest one, followed by prepositions, morphosyntax and, finally, phonology. A similar pattern is yielded by the maximum number of LREs per dyad, the ‘spelling’ target being the one that achieved the highest number and phonology the lowest.
As to the elaboration of the LREs in the OE task,
Table 5’s data indicated that the range of turns is slightly wider for meaning-based than for form-based LREs, an outcome which is accompanied by a far higher number of LREs per dyad in meaning-based than in form-based LREs. It was also discovered that meaning-based LREs targeting ‘word meaning’ yielded a wider range of turns than those aiming at ‘word choice’. The maximum number of LREs per dyad also resembles this result, as a strikingly higher number of LREs was produced for ‘word meaning’ targets, as compared to ‘word choice’. As for the various form-based LRE categories, it should be first noted that no cases of LREs discussing ‘spelling’ or ‘prepositions’ were found. Second, the range of turns discussing ‘phonology’ was wider than that of those discussing ‘morphosyntax’, even though the highest maximum number of LREs per dyad was found for the latter.
The analysis of the most elaborate targets in each task seems to indicate that meaning-focused LREs are lengthier than form-focused LREs, this tendency being more marked in the OW task. This result is consistent with the findings from other studies with young learners performing written tasks such as dictogloss (
Calzada 2021;
Calzada and García Mayo 2021). It is striking though to see how the maximum number of LREs per pair is higher in the case of meaning-focused LREs in the OE task than in the OW task, a finding that could be ascribed to the nature of the task (storytelling), but this OE mode may not invite the learner to engage in elaborate meaning-focused discussions. This finding is also related to the trend observed for the elaboration of the different types of meaning-focused LREs. The OW task generates more elaborate word choice LREs and the OE more elaborate word meaning LREs. It seems as in the OE learners stopped at moving the task forward and in the OW learners were geared towards finding the appropriate word.
If we tackle form-focused LREs, we can observe how learners engage in lengthier discussions in the case of spelling in the OW task and in phonology in the OE task, despite the attention to morphosyntax shown by the learners in either task. This result does not seem to be surprising if we take into account the relationship with sound production that both phonology and spelling may entail. It seems as if in both task modalities learners would decide to focus on the most superficial layer of form. As
Izumi (
2003, p. 190) has claimed, ‘the limited capacity of the human attentional system is likely to affect the efficiency of the encoding and monitoring processes in production’. This effect might be stronger in L2 learners due to their greater need to exercise controlled processing that requires attentional control (
Kormos 1999; as cited in
Izumi 2003). Elaboration of both aspects would come up thanks to the inherent editing of the written modality and the explicitly required editing of the oral modality, as well as thanks to their available resources and knowledge to discuss these language aspects, which may be developed further than those resources for discussing grammar, given the young age of these learners. Most pronunciation and spelling discussions revolved around a common issue, namely that of the orthoepic competence (
Council of Europe 2002), that is, the ability to retrieve sounds from graphemes and vice versa (
Fernández Martín 2009). Schoolchildren are particularly accustomed to dealing with these issues, as literacy in the mother (and foreign) language is one of main concerns in primary education classrooms. In fact, as reported in (
Muñoz 2014, as cited in
Villarreal and Munarriz-Ibarrola 2021, p. 91), “spelling and word combinations appear as early concerns for young English language learners, while more grammatical concerns appear as they grow older”. In addition, the lower level of elaboration in the case of grammar aspects could also be ascribed to the fact that spelling and pronunciation aspects may interfere with communication more severely than morphosyntax, as research on corrective feedback has revealed (
Mackey et al. 2000).
7. Conclusions
This paper has contributed to further exploring task-modality effects on learning opportunities operationalized as LREs in young learners. Our previous study (
Martínez-Adrián and Gallardo-del-Puerto 2021) evinced the influence of task-modality at the level of the incidence, nature and incorporation of LREs in the sense that tasks with a writing component promote a greater number of LREs, more form-focused LREs and the incorporation of target-like resolutions in the final product. The present investigation has also revealed that OW tasks foster a higher level of engagement with the language and that in general terms, learners tend to elaborate more meaning-focused LREs than form-focused LREs in either task. However, this picture is not entirely crystal-clear. Analyzing the data from a top-down perspective has offered a fairer comparison across tasks in terms of the type of targets and engagement with the language during collaborative dialogue.
While in terms of meaning-focused LREs, the OW task led the learners to attend to and elaborate discussions on word choice, the OE task enhanced learners’ focus and engagement in word meaning. In the case of form-focused LREs, the OW task fostered a greater focus on morphosyntactic aspects and spelling, with a higher engagement observed in the latter. Morphosyntactic aspects were also the target of learners’ discussions in the OE task together with phonological aspects, but slightly more elaborate discussions were attested in the latter.
In the light of the results obtained and the idea that a more elaborate engagement is more likely to lead to L2 learning (
Fernández Dobao 2014;
Storch 2008;
Storch and Wigglesworth 2010), learning gains would potentially be more evident in certain features than others. In this regard, given that discussions about morphosyntactic aspects were not so elaborate in the present study, focused tasks (see
García Mayo 2018;
Nassaji and Fotos 2011 for a review) with a noticing and an awareness component (
Lyster 2007,
2015) could enhance learners’ discussions in this respect.
For future research, it would be convenient to increase the sample size so as to improve the generalizability of the results. Likewise, taking into account the interplay between the pairing method, patterns of interaction and the use of previously known languages during task-based interactions in the case of young learners (see
Martínez-Adrián et al. 2021), a follow-up study including these variables would be desirable.