The impact of imposing processing strategies on L2 learners' deliberate study of lexical phrases

System 56 (2016) 127e139

Contents lists available at ScienceDirect

System journal homepage: www.elsevier.com/locate/system

The impact of imposing processing strategies on L2 learners' deliberate study of lexical phrases June Eyckmans a, *, Frank Boers b, Seth Lindstromberg c Ghent University, Department of Translation, Interpreting and Communication, Groot-Brittanni€ elaan 45, 9000 Ghent, Belgium Victoria University of Wellington, School of Linguistics and Applied Language Studies, Kelburn Parade, Wellington 6012, New Zealand c Hilderstone College Kent, St. Peter's Road, Broadstairs, Kent CT10 2JW, United Kingdom a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 March 2015 Received in revised form 14 November 2015 Accepted 2 December 2015 Available online xxx

Lexical phrases (e.g., collocations and idioms) have in recent years attracted a fair amount of interest in the fields of SLA and language pedagogy. However, there is still a shortage of empirical studies of instructional methods and techniques. The present study focuses on the deliberate memorization of L2 phrases and tests the effectiveness of two proposals for how to help learners accurately recall the lexical composition of L2 phrases. EFL learners (N ¼ 65) studied a list of 32 figurative VERB þ NOUN-PHRASE expressions (e.g., turn the tide) glossed in L1. In one condition, they were only instructed to study the list. In a second, they were additionally instructed to identify phrases whose verb is non-congruent with the L1 translation. In a third, the instruction was to identify phrases that alliterate. The 32 phrases included roughly equal numbers of congruent and alliterative ones. It was hypothesized that the extra instructions would cause learners to pay closer attention to the verbs in the phrases and that this would result in better recall. Comparisons of pre- to post-test gains suggest that the instruction to look out for alliteration enhances retention, whereas the instruction to compare with L1 does not. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Collocations Idioms Memorization Lexical selection Congruency Alliteration

1. Introduction The past couple of decades have witnessed a proliferation of research into the formulaic dimension of language, including a steep increase of studies on the role and nature of lexical phrases (also known as collocations, idiomatic expressions, formulaic sequences, phrasal expressions and multiword units, among other terms). It is now widely acknowledged that knowledge of phrasal lexis benefits second (or foreign) language learners in various ways. For example, it facilitates fluent real-time language processing (Skehan, 1998; Wray, 2002), both receptive (Conklin & Schmitt, 2008; Ellis, Simpson-Vlach, & Maynard, 2008) and productive (Boers, Eyckmans, Kappel, Stengers, & Demecheleer, 2006; Wood, 2010); and it helps learners to produce ‘idiomatic’ language rather than unconventional word strings which may hinder smooth communication (Millar, 2010; Pawley & Syder, 1983) and which may negatively affect learners' scores on proficiency tests (e.g., Boers et al., 2006; Dai & Ding, 2010; Stengers, Boers, Housen, & Eyckmans, 2011). In short, mastery of a sizeable repertoire of lexical phrases is one of the characteristics of advanced proficiency in a second language (Durrant & Schmitt; 2009; Granger & Bestgen, 2014).

* Corresponding author. Tel.: þ32 474 32 07 79. E-mail addresses: [email protected] (J. Eyckmans), [email protected] (F. Boers), [email protected] (S. Lindstromberg). http://dx.doi.org/10.1016/j.system.2015.12.001 0346-251X/© 2015 Elsevier Ltd. All rights reserved.

128

J. Eyckmans et al. / System 56 (2016) 127e139

However, in the absence of massive amounts of exposure to the target language, learners tend to be very slow at developing phrasal competence (Forsberg, 2010; Kuiper, Columbus, & Schmitt, 2009; Laufer & Waldman, 2011; Li & Schmitt, 2010) and so initiatives to accelerate this learning process would be very welcome (Boers & Lindstromberg, 2009; Lewis, 1993; Nattinger & DeCarrico, 1992). Several types of pedagogic intervention with this aim have been experimentally tested in recent years. Some of these interventions direct learners' attention to lexical phrases encountered in texts. This can be done by manipulating the frequency of the occurrence of particular phrases in a text (Webb, Newton, & Chang, 2013), by typographically enhancing phrases (Peters, 2009; 2012; Sonbul & Schmitt, 2013), and/or by providing glosses in the case of semantically non-transparent phrases (Bishop, 2004). Such interventions are in keeping with Schmidt's (1990; 2001) Noticing Hypothesis, which holds that attention is a crucial first step towards retention. Other interventions assign a more active, discovery role to the learners. For example, students can be asked to identify lexical phrases in texts themselves (Jones & Haywood, 2004) and they may be asked to consult dictionaries (Komuro, 2009) or web-based resources (Chen, 2011) to check the idiomaticity of their L2 output. Benefits from such approaches would be in accordance with Laufer and Hulstijn's (2001) Involvement Load Hypothesis, which holds that the chances of incidental uptake of lexis are enhanced when learners feel the need to search for (the meaning of) lexical items and then evaluate whether the outcome of that search has fulfilled the need. Other interventions present sets of preselected lexical phrases to students for deliberate study (e.g., Webb & Kagimoto, 2011). While a language's vocabulary (including multiword lexis) is clearly too vast to be acquired through deliberate study alone (Nation, 2013: 92), deliberate study can at least speed up the learning process (Laufer, 2005). Given the growing recognition of the importance of formulaic language, it would therefore seem a welcome trend for contemporary textbooks to include exercises focusing on lexical phrases, such as exercises where learners are asked to match separated constituents of phrases or to supply the missing parts of collocations in gapped sentences. This broad type of exercise is a staple of course books for learners of English. The effectiveness of matching and blank-filling exercises which require learners to reassemble split phrases remains under-researched (but see Boers, Demecheleer, Coxhead, & Webb, 2014; Boers, Eyckmans, & Lindstromberg, 2014; Boers, Lindstromberg, & Eyckmans, 2014, for preliminary evidence suggesting that the learning gains they bring about may be small). While it is safe to say that the aforementioned interventions help raise learners' awareness of the phraseological dimension of language generally and to direct their attention to particular lexical phrases, it is possible that more “engagement” (Schmitt, 2008) with the meaning and the form of these phrases is required for learners to form accurate and durable representations of them in long-term memory. In the realm of single-word learning, explorations of the benefits of mnemonic techniques, such as the Keyword Method (Atkinson, 1975), have a long history (e.g., Avila & Sadoski, 1996; Barcroft, Sommers, & Sunderman, 2011; Sagarra & Alba, 2006; Shapiro & Waters, 2005). However, there have been far fewer initiatives regarding the design and evaluation of interventions intended to help learners commit multi-word items to memory. Most of those initiatives have to do with figurative idioms (e.g., on the ropes) and with how the imageability of a targeted idiom can be enhanced by informing learners about the context (e.g., boxing) in which the targeted expression is, or formerly was, used with a concrete meaning (Boers, 2013, for a review). A limitation of this technique is that it cannot be applied to all lexical phrases. Besides, it seems to foster retention of the meaning of idioms, first and foremost, but not necessarily their precise lexical composition. This matters because many lexical phrases are problematic for learners not because of their meanings or communicative functions but because of some feature of their form. For example, if one understands the words tell and lies, then Don't tell lies probably presents no challenge to interpretation e hence the distinction that is often made between idioms, whose meaning transcends the meaning of the constituent words, and collocations, whose overall meaning may be inferred relatively easily from the constituent words themselves. Especially if the objective is the accurate, productive use of lexical phrases, the challenge for learners includes mastering their lexical composition. For example, one needs to know that lies is typically preceded by tell rather than by a semantically related verb such as say, and that a photo is typically preceded by take rather than by make, and that mistakes follows make rather than do. The semantics of the verbs in these examples offer no obvious clues as to why one rather than the other accompanies the given noun. This arbitrariness can further be illustrated by the many examples of non-congruent verb selection across languages which share near-equivalent phrases (e.g., the Dutch counterpart of make an effort is een inspanning doen, ‘do an effort’). Indeed, making the right choice of verb seems to be particularly difficult for L2 learners, according to several studies (Laufer & Waldman, 2011; Nesselhauf, 2005; Peters, 2016), which is why expressions consisting of verb þ noun-phrase (NP) were chosen as targets for learning in the quasi-experimental study we report in this article. The study is part of an on-going quest for ways to engage learners in kinds of processing likely to help them durably entrench lexical phrases in memory. In this particular investigation, the focus is on learners' retention of non-substitutable verbs in phrases of the targeted type. We asked three groups of EFL learners to memorize a set of English intact lexical phrases, accompanied by their L1 translations. One group (n1 ¼ 22) was given no further instructions of how they should go about this. The other two groups (n2 ¼ 22; n3 ¼ 21) were given additional instructions intended to cause them to consider the composition of the phrases more closely. The nature of those instructions and the rationale behind them are explained in the next section. 2. Stimulating elaboration We borrow the term elaboration from Levels-of-Processing Theory (e.g., Craik & Tulving, 1975), where it is used to refer to learners' mental operations regarding to-be-learned lexis. It was originally proposed that word retention is aided most by semantic elaboration, that is, by forming various associations with the meaning of the to-be-learned word. However, it has been recognized that semantic elaboration may not be sufficient if the aim is for learners to recall the precise form of words (cf.

J. Eyckmans et al. / System 56 (2016) 127e139

129

the notion of Transfer-Appropriate Processing; Morris, Bransford, & Franks, 1977). For the latter purpose, structural elaboration (Barcroft, 2002) may be more useful, that is, the performance of mental operations requiring attention to the form of the target item. In the present study, we are interested in learners' retention of the composition of lexical phrases. Insofar as this is a formal rather than a semantic feature, the elaboration which our treatment procedures aim to stimulate can be considered to be of the structural rather than the semantic kind. The following two proposals for stimulating elaborations concerning the lexical composition of lexical phrases e with a special interest in the verb e informed the interventions that were tested in the present study. One proposal, by Laufer and Girsai (2008), is for teachers (or materials writers) to encourage learners to compare and contrast the lexical composition of L2 phrases with the lexical composition of L1 counterparts of those phrases in order to forestall errors arising from the well documented tendency of learners to translate L1 collocations into L2 word by word (e.g., Nesselhauf, 2005; Wolter & Gyllstad, 2011; Yamashita & Jiang, 2010). Evidence of the benefits of helping learners to engage in contrastive analysis of this kind has come from a quasi-experimental study (Laufer & Girsai, 2008) in which greater learning gains were observed in the contrastive-analysis treatment condition than in either a treatment that included no explicit focus on the composition of the target phrases or in another treatment which involved gap-fill and multiple-choice exercises without contrastive considerations. However, the better results of the contrastive-analysis group should be interpreted with some caution because the test used a translation format, and the contrastive-analysis group had used translation as part of their learning procedure whereas the others had not. Further investigations of the merits of asking learners to engage in L1eL2 comparisons of lexical phrases would therefore be welcome. The second proposal for stimulating engagement with the lexical makeup of phrases is for teachers to draw learners' attention to phonological patterns such as alliteration (e.g., slippery slope; time will tell; make a mess) and assonance (e.g., small talk; high time; take a break), whose incidence in formulaic language is far above-chance (Boers, Demecheleer, et al., 2014; Boers & Lindstromberg, 2009: 114; Gries, 2011). While the preference for a particular word over a near-synonym in a lexical phrases often defies semantic explanations, there quite often appear to be plausible phonological explanations (e.g., time will tell alliterates, whereas time will say/show/reveal do not). This ‘words-of-a-feather-flock-together’ phenomenon suggests that patterns such as alliteration and assonance can make word strings relatively memorable. This possibility is in line with strong findings from four major streams of psycholinguistics research concerned with the effects of patterns of sound repetition, or ‘phonological similarity’, on the accessibility in memory of lexical forms. Here, we can touch only briefly on just a few of these streams. One stream, comprising many hundreds of studies, concerns the learning of lists of paired-associates, these being mostly short, discrete, more or less arbitrarily paired L1 items such as catehat (e.g., Bower & Bolton, 1969; Nelson & Garland, 1969; see Rubin, 1995, for an overview). Another stream has to do with the short-term serial and nonserial recall of items from various kinds of lists (e.g., all-alliterative lists or lists where only some items alliterate, or where none do) (e.g., Watkins, Watkins, & Crowder, 1974; for a review see Gupta, Lipinski, & Actunc, 2005). A third stream is concerned with effects of phonological similarity on implicit priming of targeted response words (see Levelt, 1999, for a brief summary). A fourth stream concerns the facilitative role of patterns of phonological similarity (especially rhyme and alliteration) in the memorization of L1 oral texts such as song lyrics (Rubin, 1995). A common observation, especially in the paired-associates studies, is that the patterns of phonological similarity have robust effects in the wake of an intervention to promote noticing of phonological form but weak effects, or even none, in the absence of such intervention (Rubin, 1995). In our field (of SLA and language pedagogy), there have been relatively few experimental studies of phonological similarity effects on shortterm or long-term learning, but findings from a number of those studies that have been carried out suggest that the mnemonic potential of alliteration tends to remain dormant except when there has been a pedagogical intervention likely to induce noticing of phonological form (Boers, Lindstromberg, et al., 2014; Lindstromberg & Boers, 2008a, 2008b). However, it is not yet clear how interventions that direct learners' attention to phrases manifesting alliteration or a similar phonological trait affect the recall of those phrases in a to-be-learned set of phrases which do not manifest the phonological trait. It is conceivable that stimulating engagement with target phrases which manifest a stipulated trait privileges recall of those phrases at the cost of other to-be-learned phrases (Boers, Eyckmans, et al., 2014), thus dampening the overall benefits of the intervention. So, clearly this present proposal for engaging learners with the composition of lexical phrases requires further evaluation as well. Given the above considerations, we asked three groups of EFL learners to study a set of verb-noun expressions containing (a) items that are congruent and that are non-congruent with their L1 and (b) items that alliterate and that do not alliterate. One group was asked to study the items without any further instruction, while the other two groups received additional instruction, intended to promote their engagement with the composition of the expressions: One group was asked to decide if the verb was different from the translation equivalent in L1; the other group was asked to decide if the verb alliterates with the noun.

3. Research questions The research questions are as follows: 1) Do instructions that promote learners' attention to the lexical makeup of a set of phrases during a deliberate memorization task lead to better recall of that set of phrases than a condition where learners are given no such instructions?

130

J. Eyckmans et al. / System 56 (2016) 127e139

2) Does the type of instruction, that is, checking for non-congruency with L1 or checking for alliteration, make a difference to the overall recall rates? 3) Does the type of instruction, that is, checking for non-congruency with L1 or checking for alliteration, privilege recall of the phrases which display the trait to be sought, possibly at the expense of others?

4. Method 4.1. Participants Three intact classes of EFL pupils (total n ¼ 65, 22 male and 43 female) at a secondary school in Flanders, Belgium, participated in this quasi-experimental study. Sixty-one were native speakers of Dutch, four were bilinguals with Dutch as one of their native languages. Since the pupils were enrolled in different study programs (i.e., Latin-Modern Languages, LatinMathematics, and Sciences) that involve differing numbers of English classroom hours, pupils from these different classes were proportionally distributed across three groups. Each group was randomly assigned to one of the three treatment conditions. The pupils' age ranged from 13 to 14 years old, and their proficiency in English was estimated by their teacher to be at least at level A2 of the Common European Framework of Reference. 4.2. Materials A set of 32 verb-noun phrases were selected from sources such as McCarthy and O'Dell's books for independent study of collocations and idioms (2002; 2003; 2005; 2008). It is because some of the target phrases used in the experiment would be considered as collocations and others as idioms, according to criteria such as semantic (non)decomposability (Cowie, 1981; Gibbs, Nayak, & Cutting, 1989), that we use (lexical) phrase as an umbrella term throughout this article. We chose the 32 target phrases also with a view to creating four subsets of roughly equal size of phrases that are (a) alliterative and congruent with L1 (b) alliterative and not congruent with L1 (c) not alliterative and congruent with L1, and (d) not alliterative and not congruent with L1. The 32 target phrases were listed in random order on a handout together with their translations in the pupils' L1. The list was laid out in two columns so as to make it fit on one side of the handout sheet (See Appendix A). To counter the possibility that target items presented at the top or at the bottom of a column would enjoy a comparative advantage in subsequent retrieval, four fillers items were added to occupy those places on the handout sheet. These four fillers were not included in the pre/post-tests. Two tests with differing degrees of difficulty were created to gauge the pupils' knowledge of the target phrases. The tests consisted of gapped sentences each incorporating one of the 32 target phrases in a meaningful context. In the pre-test (see Appendix B), only the verb of each target phrase was missing. Requiring pupils to supply the complete phrase would have carried the risk of them filling the blanks with other arguably appropriate phrases than the ones targeted in the treatment. This would then not have allowed us to conclude the pupils had inaccurate or no prior knowledge of the actual target phrases. To gauge their gain in knowledge as a result of the treatment (i.e., the deliberate study phase), the same test (but with the gapped sentences presented in another order) was used as a delayed post-test. An immediate post-test also consisted of a series of gapped sentences, but this time the pupils were asked to supply the whole lexical phrase. At this point the pupils were informed that they were expected to fill the blanks with phrases they had just studied. This more challenging test was deemed appropriate as it followed closely on the deliberate study phase. The tests did not give L1 translations of the target phrases as prompts in addition to the sentential co-text. Such prompts could have been to the advantage of the group of pupils who were led in the study phase to compare and contrast the L2 phrases with their translations. 4.3. Procedure The pre-test was administered one week before the study phase. In this test, the secondary school teacher who agreed to gather the data for this study, asked the pupils to supply the verb of the 32 target phrases, all of which were embedded in a sentence (e.g., When he was first elected mayor he believed he was going to be able to ______________ mountains and make life better for everyone in the city. He soon realised there were limits to what he could achieve.). Pupils were given 15 min to complete the test and this was sufficient. The test sheets were collected and no feedback was given. In the subsequent treatment (or study) phase, their teacher first illustrated the relevance of studying lexical phrases by giving examples of ‘odd’ word combinations (e.g., a pretty man) and examples of how substituting one word for another can change the meaning of the phrase as a whole (e.g., a big woman vs. a tall woman). She then handed out the sheet with the list of English phrases (and their translation) preceded by the instruction to study these as well as possible in preparation for a memory test in which they would be required to use the expressions in gapped sentences. This instruction to study the phrases as well as possible was supplemented in two of the groups with an additional instruction (see Appendix C). In one group, an example of alliteration was given and the pupils were given the task of putting a tick on their handout before phrases that display alliteration. They were to do this as they studied the phrases. In the other group, an example was given of an English verbenoun collocation

J. Eyckmans et al. / System 56 (2016) 127e139

131

whose verb differs from its Dutch equivalent (e.g., drive someone crazy, iemand gek maken ‘make someone crazy’) and the pupils were given the task of putting a tick before phrases on their handout that were different in that way from their Dutch translations. We decided to have pupils tick the non-congruent phrases in the latter condition in keeping with the finding that transfer from L1 collocations comes rather naturally and that it is therefore cases where this transfer leads to infelicitous word combinations that invite intervention to curb it. In all three groups, the pupils were given 15 min to study the list of phrases in preparation for the announced memory test. When time was up, the teacher collected the handouts. After a short interlude activity that consisted of watching an unrelated video clip, the pupils were given the immediate post-test, which they completed in 15 min. The test sheets were collected and no feedback was offered at this stage e pupils were promised their work would be returned to them with feedback after a week or so. Ten days later the delayed post-test, which had not been announced, was administered. In this test e as in the pretest e the pupils were asked to supply only the verb of the lexical phrases (e.g., The authorities have decided to ______________ new standards for health and safety at work because the old ones are out of date.). They were again given 15 min to complete this test. In all the tests, scoring was strict with responses considered either right or wrong. Spelling mistakes (e.g., *seak solace) were practically non-existent in the tests and were considered wrong responses. This corresponded to the correction practice that the pupils were used to.

5. Results Before comparing the three groups' overall test scores and their scores for each subset of target phrases in particular, we needed to ascertain that the students who received instruction to tick alliterative phrases or to tick non-congruent phrases had adequately followed this instruction. Inspection of the students' study sheets revealed that the students who were asked to tick alliterative phrases had done so very successfully: Only seven instances (2.3%) of alliteration were overlooked and just one instance of a non-alliterative phrase was ticked by mistake. It is thus safe to say that these participants generally considered the same phrases to be alliterative that we did. However, inspection of the study sheets collected from the contrastive-analysis group, indicated that these students considered as non-congruent a couple of the phrases which we had considered to be congruent and vice versa. For example, we had included seek solace as a phrase congruent with its Dutch translation ‘troost zoeken’, but the majority of the students ticked this item to indicate they considered it non-congruent. It is possible that these students were not yet familiar with the verb seek, which may have prevented them from recognising the congruency. For these students, the more familiar verb search is likely to spring most readily to mind as the equivalent of Dutch ‘zoeken’. Conversely, the phrase hit the target, which we expected the students to tick as non-congruent with the Dutch translation ‘doel treffen’ (because we thought the students would associate hit with the Dutch verb ‘slaan’ rather than ‘treffen’) was in fact ticked by very few. We shall return to such issues of polysemy in determining L1-L2 congruency in the discussion section further below. As we felt the data analysis should reflect the way the participants had performed the task and how they had engaged with the phrases (rather than how we had expected them to), we decided to re-assign phrases for which the majority of the participants' responses diverged from our initial expectations. This resulted in adjusted subsets: (a) 7 alliterative congruent phrases, (b) 9 alliterative non-congruent ones, (c) 9 non-alliterative congruent ones, and (d) 7 nonalliterative non-congruent ones. Appendix A shows which 16 phrases were deemed congruent by the majority of the participants. The descriptive statistics concerning the three groups' test performance are summarized in Tables 1 and 2. Table 1 displays the overall mean test scores per group, while Table 2 lists the mean per-item score for each phrase category (congruent alliterative items, congruent non-alliterative items, incongruent alliterative items, and incongruent non-alliterative items). The information in Table 1 suggests that the three groups had roughly equivalent prior knowledge of the set of 32 target phrases. Table 1 further suggests that the group that was encouraged to engage in contrastive analysis performed relatively poorly in the post-tests, whereas the group that was asked to check items for the presence of alliteration performed relatively well. It may seem surprising that the immediate post-test scores are lower than the delayed post-test scores, but recall that the former was more challengingepupils were required to supply not just the verb but the whole lexical phrase. This difference needs to be kept in mind when interpreting the pre-test to immediate post-test gains. It is the comparison between the pre-test and the delayed post-test that is the most legitimate because these tests were identical (apart from the item order). Our analysis therefore focuses on the delayed post-test e although it is worth mentioning that the immediate postTable 1 Mean overall test scores (and standard deviations) per group (max ¼ 32). Pre-test

Immediate post-test

Delayed post-test

No additional instructions (n 22)

4.23 (3.89)

10.32 (8.00)

Tick non-congruent phrases (n 21)

4.05 (3.19)

8.62 (6.43)

Tick alliterative phrases (n 22)

4.55 (3.04)

11.50 (7.88)

13.77 (7.83) Gain: 9.54 (29.8%) 12.00 (6.83) Gain: 7.95 (24.8%) 15.95 (8.24) Gain: 11.4 (35.6%)

132

J. Eyckmans et al. / System 56 (2016) 127e139

Table 2 Mean success rates (and standard deviations) in percentages for each phrase category. Learning condition

Type of phrase

Pre-test

Immediate post-test

Delayed post-test

No additional instructions (n 22)

þ þ þ   þ   þ þ þ   þ   þ þ þ   þ  

11 (15)

18 (18)

3 (5)

29 (27)

30 (13)

39 (16)

8 (8)

41 (26)

10 (15)

16 (22)

3 (3)

23 (22)

28 (22)

37 (18)

10 (10)

31 (24)

11 (25)

23 (24)

3 (3)

34 (23)

33 (20)

50 (17)

8 (8)

37 (24)

38 (22) Gain: 27 34 (20) Gain: 31 58 (19) Gain: 28 40 (19) Gain: 33 29 (20) Gain: 19 28 (14) Gain: 25 54 (23) Gain: 26 37 (23) Gain: 27 43 (26) Gain: 32 50 (16) Gain: 47 66 (12) Gain: 33 40 (22) Gain: 32

Tick non-congruent phrases (n 21)

Tick alliterative phrases (n 22)

alliterative congruent alliterative congruent alliterative congruent alliterative congruent alliterative congruent alliterative congruent alliterative congruent alliterative congruent alliterative congruent alliterative congruent alliterative congruent alliterative congruent

test data show perfectly parallel trends: screening the phrases for alliteration was more associated with learning gains than deciding whether the phrases contained a non-congruent verb. Scores on both the immediate as the delayed post-test also showed that the learning gains of the no-intervention group were in between those of the alliteration-identification group and the contrastive-analysis group. The pupils whose task it had been to tick non-congruent phrases gained on average 7.95 items between the pre-test and the delayed post-test, whereas those whose task it had been to tick alliterative phrases gained on average 11.40 items. The mean gain for the pupils who had not received any additional instructions (henceforth the no-intervention group) was 9.54. The following chi-square analysis confirms the impression that the instruction to tick alliterative phrases was more helpful than the instruction to tick non-congruent ones. The number of failed pre-test responses was 611 (86.88%) in the no-intervention group, 587 (87.35%) in the group instructed to check for non-congruency, and 604 (85.8%) in the group instructed to look for alliteration. The proportions of these unsuccessful pre-test responses that were followed by successful delayed post-test responses were 35.27%, 29.64% and 42.74%, respectively. The difference is significant (c2 ¼ 14.84; p < .001, d ¼ 0.47, CI95% [0.35, 0.60]).1 The task of identifying alliterative phrases was more strongly associated with gains than the contrastive-analysis task (c2 ¼ 21.82, p < .0001, d ¼ 0.32, CI95% [0.18, 0.45]).2 Moreover, the latter was associated less with gains than the condition where pupils were left to their own mnemonic devices (c2 ¼ 4.30, p ¼ .04, d ¼ 0.20, CI95% [0.39, 0.01]). On the other hand, the intervention which raised the pupils' awareness of the presence of alliteration does appear to have been helpful, since it was more strongly associated with gains than the no-intervention condition (c2 ¼ 7.02, p ¼ .008, d ¼ .17, CI95% [0.04, 0.30]). The next question is whether asking learners to identify phrases which meet the criterion stipulated by the instruction (i.e., in our case, identifying phrases that manifest either non-congruency or alliteration) might have a positive effect on their retention of these phrases, but possibly at the expense of the other phrases in the to-be-learned set. Let's first consider the (non-)congruency factor. Per-item analysis shows that the set of congruent phrases attracted roughly 16% more correct responses in the pre-test than the set of non-congruent ones, which is consistent with the facilitative role of congruency observed in previous studies (e.g., Wolter & Gyllstad, 2011). Congruency indeed correlates significantly with the mean pretest scores calculated per target phrase (r2 ¼ .25; p ¼ .003). This correlation is even stronger than that computed for corpus frequency (extracted from the Corpus of Contemporary American English or COCA, http://corpus.byu.edu/coca/ in December 2014, see Appendix D)efrequency being a proxy for the likelihood that learners had incidentally encountered the phrases in English discourse prior to the experiment (r2 ¼ 20; p ¼ .01). Non-congruent phrases did not appear to resist deliberate

1 Throughout this article a ¼ .05, two-sided. Values of c2 derive from the proportion test. Values of Cohen's d and 95% confidence intervals (CI95%) were derived online from: http://www.campbellcollaboration.org/escalc/html/EffectSizeCalculator-SMD10.php. In two cases this calculator failed to return a CI. We calculated these CIs in R using function ‘chies’ in the ‘compute.es’ package. The standard default interpretations of values of d are that 0.80 indicates a large effect (80% of a standard deviation) while 0.50 and 0.30 indicate a medium and a small effect, respectively. In educational research an observed effect of 0.30 is slightly above average (Grissom & Kim, 2012). 2 Because a ¼ .05, all CIs are 95% CIs.

J. Eyckmans et al. / System 56 (2016) 127e139

133

learning more than congruent ones, however. The mean pre-test to delayed post-test gain was actually slightly better (32.5% vs. 28%) for the set of non-congruent phrases. The contrastive-analysis group does not stand out in this regard (See Table 2). Interestingly, their mean gain (26%) was also low on the set of non-congruent items, as compared to the other two groups. Chi-square analysis even indicates that the non-congruent items were generally better learned under the alliterationidentification than the contrastive-analysis procedure (c2 ¼ 10.88, p ¼ .001, d ¼ 0.30, CI95% [0.12, 0.48]). There is thus no evidence in our data to suggest that asking learners to check items for non-congruency with L1 counterparts during memorization of a set of lexical phrases fosters either recall of the set in general or recall of the non-congruent items in particular. We now turn to the factor of alliteration. Per-item analysis shows that the set of alliterative phrases generated lowest mean pre-test successes than the set of non-alliterative ones (7% vs. 19.5%). The phrase tell the truth excepted, the alliterative phrases were in general less frequent than the non-alliterative ones according to counts in COCA, and so the pupils were perhaps less likely to have encountered them before. Overall, the mean learning gains were virtually identical for the alliterative and nonalliterative phrases (approximately 30%), except in the group whose task it was to identify alliterative phrases, where the gains were better for the subset of alliterative items than the non-alliterative ones (see Table 1). More learning of alliterative phrases occurred in this group than in the other two groups: c2 ¼ 19.24; p < .001, d ¼ 0.31, CI95% [0.14, 0.47], with the alliteration-identification group outperforming both the non-intervention group (c2 ¼ 5.51, p ¼ .02, d ¼ 0.21, CI95% [0.39, 0.03]) and the contrastive-analysis group (c2 ¼ 19.08, p < .001, d ¼ 0.41, CI95% [0.22, 0.59]). As no advantage for alliterative phrases was observed in the no-intervention and contrastive-analyses groups, the data is consistent with the thesis that it takes conscious noticing of alliteration to harness its mnemonic potential. At the same time, there is no evidence that the conscious noticing of alliteration brought about by the instruction to inspect forms for the presence of alliteration came at a cost. The learning gain on the non-alliterative phrases was at least as good as in the no-intervention group (c2 ¼ 1.90, p ¼ .17, d ¼ 0.12, CI95% [0.05, 0.29]) and significantly better than attested in the contrastive-analysis group (c2 ¼ 4.46, p ¼ .03, d ¼ 0.18, CI95% [0.01, 0.35]). A problem with a gain score (i.e., post-test score e pre-test score) is that it does not account for the fact that a low pretest score tends to be associated with a high gain score just because a participant with a low pretest score (compared to a participant with a high pretest score) has more test items on which to improve, including relatively easy items that higher scoring participants already knew. A way of addressing this bias is to calculate a ‘relative gain’ (RG) score for each participant (e.g., Horst, Cobb, & Meara, 1998) as follows: RG ¼ [(gain score)/(maximum possible  pretest score)] * 100. The italicized portion of the equation is a proportion, namely, amount of gain per amount of available gain. Multiplication by 100 transforms the proportion into a percentage. Table 3 shows the mean of per-participant relative gain scores for each type of idiom in each treatment group. With the non-intervention group being taken as baseline, it can again be seen that direction of attention to alliteration had an overall positive effect whereas direction of attention to (in)congruence with L1 had an overall negative effect. So far, we have assumed that the differences in learning gains between the three groups are due to the way the instructions channelled the pupils' processing of the to-be-learned phrases. It may be worth verifying this assumption by investigating whether item characteristics that predict learning in the no-intervention group are less influential in the groups whose engagement with the items was manipulated through the instruction they received. An item characteristic that is known to positively influence the likelihood of lexical retention is imageability of meaning (e.g., Hamilton & Rajaram, 2001); it also plays a part in deliberate idiom learning (Steinel, Hulstijn, & Steinel, 2007). Imageability of meaning is a semantic factor, and its relative impact on learning gains may serve to answer the question of whether its influence is affected by the instruction-induced elaboration of the lexical composition of the phrases, an elaboration we characterized as structural rather than semantic. Two weeks after the study phase we re-presented the list of 32 target phrases to 16 of the pupils who had participated in the study (a roughly equal number of pupils across the three learning conditions) and asked them to rate each phrase (on a 9-point scale) according to how easy they found it to form an image of the phrase's meaning. We calculated the respondents' mean ratings (see Appendix E) and used these to explore the degree to which imageability advantaged some

Table 3 Deviation of mean per-participant relative gain scores from the mean expected under the null hypothesis (DevE) and differences from the non-intervention group percentages (DiffB). Idiom types

þA þC þA C A þC A C Totalb

No additional instructions (n 22)

Tick non-congruent phrases (n 21)

Tick alliterating phrases (n 22)

DevE

NA

DevE

DiffB

DevE

DiffB

7.4% 7.0% 2.0% 6.0% e

e e e e e

9.4% 8.1% 6.6% 11.8% e

16.8* 1.1 8.6* 5.5 8.2

1.4% 13.1% 1.8% 11.0% e

6.0*a 20.1* 0.2 17.0 26.1

Notes: a Asterisks mark values relating to presence of the feature that participants' attention was directed to by the treatment. b Total of the values marked by asterisks.

134

J. Eyckmans et al. / System 56 (2016) 127e139

phrases over others per treatment group. Imageability was found to correlate with learning gains most strongly in the nointervention group (r2 ¼ .25; p ¼ .004), followed by the contrastive-analysis group (r2 ¼ .19; p ¼ .01). However, the correlation was negligible in the group instructed to look for alliteration (r2 ¼ .01; p ¼ .53), which lends support to the assumption that this instruction indeed prompted the pupils to process the target items in a way that partially overrode the influence of semantic factors on their processing. 6. Discussion and conclusion The preliminary answer to our research question concerning the merits of channelling learners' engagement during deliberate study must be that the merits depend on the type of channelling. In these data, prompting learners to engage in contrastive-analysis did not appear to be helpful in comparison with a condition where learners were left to their own mnemonic devices. On the other hand, prompting learners to look for alliteration did appear to be conducive to form retention. It seems plausible that the latter finding is generalizable to other easily-recognizable sound patterns, such as rhyme and assonance. The relatively poor effectiveness of the intervention intended to help learners notice differences between the lexical makeup of L2 and L1 phrases illustrates that not all efforts to manipulate learners' processing during a memorization task necessarily bring about better outcomes than if the learners use their own, habitual mnemonic strategies (see, e.g., Barcroft, et al., 2011). This parallels a recent study by Wei (2015) on the memorization of single words. In that study, the mnemonic intervention of instructing students to use the Keyword Method was less fruitful than unguided rote-learning while another intervention (instructing students to analyse the words for meaningful morphemes) did prove helpful. Leaving learners to their own mnemonic devices does, of course, not preclude the possibility that they engage cognitively with the to-be-learned items in ways that resemble the engagement prompted by a particular instruction. In our experiment, for example, we cannot rule out the possibility that students in the no-intervention condition also, spontaneously, resorted to L2-L1 comparisons as a way (or one of the ways) of memorizing the phrases. The implication, therefore, is not that comparing and contrasting L2 phrases with L1 counterparts is ineffective. Rather, the implication is that the way this kind of engagement was pushed in the contrastive-analysis group was not helpful. Both interventions tested in our study have in common that they asked pupils to identify items in the to-be-learned set that met a stipulated criterion (either non-congruency or alliteration). In a study by Boers, Eyckmans, et al. (2014), it was suggested that such a ‘positive discrimination’ task benefits students' retention of items deemed to meet the criterion but not their retention of items that students may give no further notice the instant they see that the stipulated trait is absent. The data in the present study do not support that idea. The learning gains for congruent and non-congruent phrases were similar in the contrastive analysis group, and so were the gains for alliterative and non-alliterative phrases in the alliteration-identification group. However, the present study is different from Boers, Eyckmans, et al. (2014) in that here a memory test on the complete set of to-be-learned phrases was announced during the treatment phase, and this made all the stimulus items equally important for that purpose. An intriguing finding from our study is the poor outcome under the contrastive-analysis treatment. It is possible that the effort invested in determining whether the verb in the L1 translation was equivalent to the verb in the L2 phrase involved excess attention to that translation equivalent, which detracted from processing the L2 target phrase itself. The task may also have been too challenging. Due to polysemy, determining the equivalence of an L2 and an L1 verb in lexical phrases is not as straightforward as one might think. For example, a pupil may associate draw in draw conclusions with its sense of drawing lines on paper and the verb ‘trekken’ in the Dutch counterpart ‘conclusies trekken’ with its core, literal sense of ‘pulling’ (as in pulling on a rope). If so, the verdict will be one of non-congruency. However, Dutch ‘trekken’ happens to have a secondary sense which is in fact equivalent to the use of draw in drawing lines. Recognising this secondary usage of Dutch ‘trekken’ will shift the balance in favour of a congruency verdict. Given such complications, it is not surprising that pupils in the contrastive-analysis group did not consistently tick the same phrases for noncongruency. Analysis of their worksheets shows 17.1% disagreement. In comparison, the alliteration-identification task was very simple, and this is reflected in the near absence of instances (only 2.3%) on the pupils' worksheets where alliterative targets were overlooked. The effort required for determining whether a given L2 target phrase manifests alliteration is probably fairly minimal, and it does not require processing the lexical composition of the translation equivalent at the same time. Also, looking for alliteration in the lexical phrases may have induced more attention to the precise orthographic form of the phrases, the benefits of which are shown in the gain scores of the alliterating as well as the non-alliterating phrases. Consideration of the role of the L1 translations raises the question of whether the treatment conditions may have led to differing learning gains in terms of pupils' retention of the meaning of the target phrases. A limitation of our study is that we did not test the pupils specifically on their comprehension of the phrases e the focus of the interventions was on knowledge of their formal lexical makeup. Although the pupils were arguably not under extreme time pressure in the study phase of our experiment, the attention they gave to formal aspects of to-be-learned lexis may nevertheless have come at the cost of processing resources that would otherwise have been allocated to semantics (e.g., Barcroft, 2002). The closest to a measure of meaning retention in our materials is the immediate post-test, where the pupils were asked to supply the whole lexical phrase in a meaningful sentential co-text. This required them to determine the extent to which the lexical phrase (if they managed to retrieve it) fits the semantics of the sentence, without the availability of the noun of the phrase as an additional

J. Eyckmans et al. / System 56 (2016) 127e139

135

cue (as in the pre-test and delayed post-test format). As shown in Table 1, the three groups' performance on the immediate post-test parallels their performance on the delayed post-test, with the alliteration-identification group obtaining a better score and the contrastive-analysis group a poorer score than the no-intervention group. This suggests that at least the (relatively straightforward) task of checking phrases for alliteration did not compromise uptake of their meaning in the deliberate study procedure that was used here. Our experimental design required sufficient numbers of phrases to create contrasting sets of alliterative vs. non-alliterative and congruent vs. non-congruent items. As a consequence, the pupils were asked to memorize a large number of phrases (36, including the four filler items), many of which they had little or no prior knowledge of (to judge by the pre-test scores). To be clear, we do not advocate using such ‘massed’ learning tasks in actual pedagogic practice. Being confronted with so many phrases is likely to have been overwhelming and it is likely to have made the memorization task extra laborious due to the need for pupils to block out cross-item interference, especially when semantically related verbs were involved (e.g., pull strings vs. draw conclusion; cut corners vs. break the silence; speak volumes vs. tell the truth) (see, e.g., Erten & Tekin, 2008; Finkbeiner & Nicol, 2003; Ishii, 2015 on the problems of memorizing semantically related words). There were in fact six instances in our data where correct pre-test responses were replaced with wrong ones in the delayed post-test, which suggests that some confusion may indeed have been engendered by the deliberate study of so many items in one go. Better outcomes can be expected if sets of study items are more manageable and if the risk of cross-item interference is minimized (Webb & Kagimoto, 2011). That said, the learners in this experiment did on average manage to supply correct responses in a delayed post-test for over 30% of the items that they were unsuccessful at in the pre-test, despite the heavy learning burden they had been given. Considering the brief study time, this general outcome does not compare at all unfavourably with the learning gains that have been attested in quasi-experimental studies on L2 learners' incidental acquisition of lexical phrases, which tended to take more time (e.g., Webb et al., 2013). What we cannot tell on the basis of our test data is how durable the phrasal knowledge obtained through deliberate study is in the longer run, and how it compares to phrasal knowledge acquired through incidental procedures in this regard. Keeping that caveat in mind, and recognizing that deliberate study can by no means be a blanket substitute for the incremental acquisition of words and phrases from meaning-focused input and interaction, such deliberate study may nevertheless be a worthwhile activity. Whether it is most beneficial to guide learners' mnemonic processing of phrases during deliberate study or whether it is best to leave learners to their own mnemonic devices is likely to depend on the kind of guidance that is chosen. In the present study, an advantage was found for an intervention that directs learners' attention to the presence (or absence) of a catchy sound pattern, but not for an intervention that engages learners in contrastive analysis. This does not mean, of course, that the latter intervention might not be pedagogically fruitful when it is used in different circumstances, such as when a teacher alerts learners to the makeup of a particular phrase as it occurs in a text or in the course of remedial teaching. Acknowledgements We would like to express our gratitude to the pupils who agreed to participate in the experiment.

Appendix A. The 32 target phrases with their translations in Dutch and the four filler items (in italics). The phrases deemed congruent by the majority of the students are underlined

make a presentation (een spreekbeurt geven) take revenge (on) (wraak nemen) gain ground (veld winnen) take a toll (een tol eisen) spread the word (een gerucht verspreiden) find fault (with) (op een fout wijzen) seek solace (troost zoeken) make a fortune (fortuin maken) crack the case (de oplossing vinden) hold your breath (je adem inhouden) tell a lie (een leugen vertellen) sell your soul (je ziel verkopen) wage war (oorlog voeren) turn the tide (het tij keren) lose heart (de moed verliezen) take a test (een test afleggen) set the tone (for) (de toon aangeven) cause a commotion (commotie veroorzaken)

face the facts (de feiten onder ogen zien) catch fire (vuur vatten) cut corners (het niet zo nauw nemen) tell the time (de tijd lezen) play a sport (een sport doen) draw conclusions (conclusies trekken) set a standard (een norm stellen) run a business (een zaak beheren) pull strings (invloed aanwenden) break the silence (de stilte breken) tell the truth (de waarheid vertellen) speak volumes (boekdelen spreken) reap the reward (vruchten plukken) hit the target (doel treffen) run a risk (een risico lopen) meet a deadline (een deadline halen) move mountains (bergen verzetten) launch a campaign (een campagne lanceren)

136

J. Eyckmans et al. / System 56 (2016) 127e139

Appendix B. The pre-test consisting of gapped sentences for each one of the 32 target phrases VERB þ NOUN EXPRESSIONS. TRY TO FILL IN THE BLANKS WITH ONE VERB EACH. 1. “Instead of wasting your time on computer games, you should choose to _____________ a sport”, my Dad pleaded, adding that someone my age should be in better physical condition. 2. After having been an ardent communist for so many years, when he accepted the managerial post at a multinational company he felt like he was _________________ his soul. 3. After he lost his wife, he got depressed and he started to _______________ solace in alcohol. 4. Alcohol addiction can _________________a heavy toll on your health, and can shorten your life by many years. 5. As part of the job application procedure you will need to ______________ a test. It contains a diversity of questions meant to evaluate your knowledge and reasoning skills. 6. Because we haven't heard all the facts, I don't think it's wise to ________________ conclusions yet. 7. David probably thinks his feelings for Jennifer are a well-hidden secret, but he doesn't realise that the way he looks at her _________________ volumes. 8. Despite the long distance, Robin Hood managed to ______________ the target with his last arrow and win the prize. 9. Few people believed Peggy had the right qualities to _____________ a business, but she proved them wrong. She's now head of one of the most profitable firms in the country. 10. For five years, Nazi Germany _____________ war on its neighbours. 11. For several minutes nobody said anything, but then Bill decided to _______________ the silence and inquired if anyone had been to the movies lately. 12. Her daughter is barely four years old but she can already _______________ the time, even when a clock is up-side-down! 13. I know it's been hard work, but you'll soon ______________ the rewards of all the efforts you've made. 14. I know you must feel disappointed by yet another rejection letter. But don't _______________ heart; I'm sure you'll get a good offer soon. 15. I told Bill to move the barbecue away from the garden hedge. I was afraid it would _______________ fire, since that kind of tree burns very easily. 16. I'll have to work late tonight if I want to _______________ the deadline for the assignment. 17. If you ______________ lies all the time, no-one will believe you when you're truthful for a change. 18. If you go out with Sarah, you'll ______________ the risk of losing Jenny's friendship because Sarah is her worst enemy. 19. It took a long time for Darwin's theory of evolution to ______________ ground because many people objected to the idea that apes and men could share the same ancestry. 20. James was arrested for drunk driving but his dad managed to ________________ strings and have the charges dropped. His dad must be a very influential man. 21. Jerry wanted to sell his CD collection, and so he told his friends to _______________ the word and make sure as many people as possible would know about it. It wasn't long before his room was crowded with people making their choice. 22. Many action movies follow the typical story-line: the hero's loved ones are killed or badly hurt by the bad guys, but he will ______________ revenge on them. 23. Mrs Hamilton opened the meeting by cracking a couple of jokes. That ____________ the tone for the whole of the meeting, which was relaxed and friendly. 24. My parents always _________________fault with the friends I go out with. I'm beginning to believe they don't want me to go out at all. 25. Sherlock Holmes was the only one who could _______________ the case. No other detective could find clues like he could. 26. Sweatshop bosses pay their workers low wages for long hours, and they ____________ corners on health and safety, too. 27. The authorities have decided to ______________ new standards for health and safety at work because the old ones are out of date. 28. The court expects complete honesty, so please swear to _____________ the truth and nothing but the truth. 29. The new Italian government is determined to ______________ the tide on fraud and corruption, which has been a major problem for many years. 30. The nurse asked me to _______________ my breath while the X-ray of my chest was taken. 31. They think they've got a great idea for a new business. They hope to _____________ a fortune in just a few years, retire, and lead a life of luxury forever after. 32. When he was first elected mayor he believed he was going to be able to ______________ mountains and make life better for everyone in the city. He soon realised there were limits to what he could achieve.

J. Eyckmans et al. / System 56 (2016) 127e139

137

Appendix C. The additional instructions for the alliteration-identification group and the contrastive-analysis group respectively Study the following list of expressions as well as you can. You'll be asked later on to use the expressions to fill blanks in sentences. Quite a few of the English expressions alliterate e the first consonant of two words is the same, as in make a mess. Put a tick (Ѵ) before the English expressions that show this pattern. Study the following list of expressions as well as you can. You'll be asked later on to use the expressions to fill blanks in sentences. In some of the English expressions, the verb is the same as in the Dutch counterpart (e.g. give a kiss e een zoen geven). In other expressions, the verb is different from the verb used in the Dutch translation (e.g. drive someone crazy e iemand gek maken) Put a tick (Ѵ) before the English expressions where the verb is different from the Dutch one.

Appendix D. Frequency and Mutual Information scores for the 32 collocations and means per category. Congruent collocations are in italics

Alliterative collocations

Freq

MI

Non-alliterative collocations

Freq

MI

gain ground set a standard sell your soul turn the tide tell the truth move mountains run a risk Mean crack the case wage war cut corners find fault (with) take a test take a toll tell the time reap the reward seek solace Mean

99 144 13 159 2157 49 22 377.6 34 320 246 108 107 139 27 3 30 112.7

9.93 8.93 6.90 9.56 8.68 9.66 9.40 9.01 8.82 8.76 9.56 4.66 3.01 7.87 6.72 13.36 11.83 8.29

run a business catch fire play a sport meet a deadline pull strings set the tone (for) draw conclusions Mean hit the target hold your breath make a fortune spread the word break the silence speak volumes lose heart tell a lie take revenge Mean

176 264 29 24 22 316 226 151.0 97 268 141 446 105 86 56 138 144 164.6

9.01 10.18 9.28 3.08 11.28 8.93 11.73 9.07 10.00 6.70 7.56 9.94 10.05 10.64 9.81 8.71 7.87 9.03

Appendix E. Respondents' imageability ratings (on a 9-point scale) for the 32 target phrases and mean and standard deviation per category

Target phrase

Imageability rating

gain ground (veld winnen) set a standard (een norm stellen) sell your soul (je ziel verkopen) turn the tide (het tij keren) move mountains (bergen verzetten) run a risk (een risico lopen) tell the truth (de waarheid vertellen) Mean (SD) for congruent, alliterative items reap the reward (vruchten plukken) crack the case (de oplossing vinden) wage war (oorlog voeren) find fault (with) (op een fout wijzen) take a test (een test afleggen) take a toll (een tol eisen) seek solace (troost zoeken) cut corners (het niet zo nauw nemen) tell the time (de tijd lezen) Mean (SD) for incongruent, alliterative items break the silence (de stilte breken) speak volumes (boekdelen spreken) tell a lie (een leugen vertellen)

4.38 4.81 5.25 4.13 6.06 6.13 7.75 5.50 (1.25) 3.63 7.13 5.56 6.06 7.56 4.81 4.94 3.31 6.94 5.55 (1.51) 7.81 3.63 7.81 (continued on next page)

138

J. Eyckmans et al. / System 56 (2016) 127e139 (continued ) Target phrase

Imageability rating

take revenge (on) (wraak nemen) hit the target (doel treffen) hold your breath (je adem inhouden) lose heart (de moed verliezen) spread the word (een gerucht verspreiden) make a fortune (fortuin maken) Mean (SD) for congruent, non-alliterative items run a business (een zaak beheren) play a sport (een sport doen) meet a deadline (een deadline halen) pull strings (invloed aanwenden) set the tone (for) (de toon aangeven) draw conclusions (conclusies trekken) catch fire (vuur vatten) Mean (SD) for incongruent, non-alliterative items

7.44 6.63 7.88 5.56 6.50 7.19 6.72 (1.39) 6.50 7.63 6.31 2.81 6.06 6.00 7.19 6.07 (1.56)

References Atkinson, R. C. (1975). Mnemotechnics in second-language learning. American Psychologist, 30, 821e828. Avila, E., & Sadoski, M. (1996). Exploring new applications of the keyword method to acquiring English vocabulary. Language Learning, 46, 379e395. Barcroft, J. (2002). Semantic and structural elaboration in L2 lexical acquisition. Language Learning, 52, 323e363. Barcroft, J., Sommers, M. S., & Sunderman, G. (2011). Some costs of fooling mother Nature: a priming study on the keyword method and the quality of developing L2 lexical representations. In P. Trofimovich, & K. McDonough (Eds.), Applying priming methods to L2 learning, teaching and research (pp. 49e72). Amsterdam: John Benjamins. Bishop, H. (2004). The effect of typographic salience on the look up and comprehension of unknown formulaic sequences. In N. Schmitt (Ed.), Formulaic sequences (pp. 227e247). Amsterdam: John Benjamins. Boers, F. (2013). Cognitive semantic ways of teaching figurative phrases: an assessment. In F. Gonzalvez-Garcia, M. S. Pena Cervel, & L. Perez Hernandez (Eds.), Metaphor and metonymy revisited beyond the contemporary theory of metaphor (pp. 229e263). John Benjamins. Boers, F., Demecheleer, M., Coxhead, A., & Webb, S. (2014). Gauging the effects of exercises on verb-noun collocations. Language Teaching Research, 18, 54e74. Boers, F., Eyckmans, J., Kappel, J., Stengers, H., & Demecheleer, M. (2006). Formulaic sequences and perceived oral proficiency: putting a lexical approach to the test. Language Teaching Research, 10, 245e261. Boers, F., Eyckmans, J., & Lindstromberg, S. (2014). The effect of a discrimination task on recall of L2 collocations and compounds. International Journal of Applied Linguistics, 24, 357e369. Boers, F., & Lindstromberg, S. (2009). Optimizing a lexical approach to instructed second language acquisition. Palgrave Macmillan. Boers, F., Lindstromberg, S., & Eyckmans, J. (2014). Is alliteration mnemonic without awareness-raising? Language Awareness, 23, 291e303. Bower, G., & Bolton, L. (1969). Why are rhymes easy to learn? Journal of Experimental Psychology, 82(3), 453e461. Chen, H. (2011). Developing and evaluating a web-based collocation retrieval tool for EFL students and teachers. Computer Assisted Language Learning, 24, 59e76. Conklin, K., & Schmitt, N. (2008). Formulaic sequences: are they processed more quickly than nonformulaic language by native and nonnative speakers? Applied Linguistics, 29, 72e89. Cowie, A. P. (1981). The treatment of collocations and idioms in learners' dictionaries. Applied Linguistics, 2(3), 223e235. Craik, F. I. M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General, 104, 268e294. Dai, Z., & Ding, Y. (2010). Effectiveness of text memorization in EFL learning of Chinese students. In D. Wood (Ed.), Perspectives on formulaic language: Acquisition and communication (pp. 71e87). New York: Continuum. Durrant, P., & Schmitt, N. (2009). To what extent do native and non-native writers make use of collocations? International Review of Applied Linguistics, 47, 157e177. Ellis, N. C., Simpson-Vlach, R., & Maynard, C. (2008). Formulaic language in native and second language speakers: psycholinguistics, corpus linguistics, and TESOL. TESOL Quarterly, 42, 375e396. Erten, I. H., & Tekin, M. (2008). Effects on vocabulary acquisition of presenting new words in semantic sets versus semantically unrelated sets. System, 36, 407e422. Finkbeiner, M., & Nicol, J. (2003). Semantic category effects in second language word learning. Applied Psycholinguistics, 24, 369e383. Forsberg, F. (2010). Using conventional sequences in L2 French. IRAL, 48, 25e51. Gibbs, R. W., Nayak, N. P., & Cutting, C. (1989). How to kick the bucket and not decompose: analyzability and idiom processing. Journal of Memory and Language, 28, 576e593. Granger, S., & Bestgen, Y. (2014). The use of collocations by intermediate vs. advanced non-native writers: a bigram-based study. International Review of Applied Linguistics in Language Teaching, 52, 229e252. Gries, S. T. (2011). Phonological similarity in multiword units. Cognitive Linguistics, 22, 491e510. Grissom, R., & Kim, J. (2012). Effect sizes for Research: Univariate and multivariate applications. New York: Routledge. Gupta, P., Lipinski, J., & Aktunc, E. (2005). Reexamining the phonological similarity effect in immediate serial recall: the roles of type of similarity, category cuing, and item recall. Memory & Cognition, 33, 1001e1016. Hamilton, M., & Rajaram, S. (2001). The concreteness effect in implicit and explicit memory tests. Journal of Memory and Language, 44, 96e117. Horst, M., Cobb, T., & Meara, P. (1998). Beyond A Clockwork Orange: acquiring second language vocabulary through reading. Reading in a Foreign Language, 11(2), 207e223. Ishii, T. (2015). Semantic connection or visual connection: investigating the true source of confusion. Language Teaching Research, 19(6), 712e722. Jones, M., & Haywood, S. (2004). Facilitating the acquisition of formulaic sequences: an exploratory study. In N. Schmitt (Ed.), Formulaic sequences (pp. 269e300). Amsterdam: John Benjamins. Komuro, Y. (2009). Japanese learners' collocation dictionary retrieval performance. In A. Barfield, & H. Gyllstad (Eds.), Researching collocations in another language: Multiple perspectives (pp. 86e98). Basingstoke, UK: Palgrave Macmillan. Kuiper, K., Columbus, G., & Schmitt, N. (2009). Acquiring phrasal vocabulary. In S. Foster-Cohen (Ed.), Advances in language acquisition (pp. 216e240). Basingstoke, UK: Palgrave Macmillan.

J. Eyckmans et al. / System 56 (2016) 127e139

139

Laufer, B. (2005). Focus on form in second language vocabulary acquisition. In S. Foster-Cohen (Ed.), EUROSLA yearbook 5 (pp. 223e250). Amsterdam: John Benjamins. Laufer, B., & Girsai, N. (2008). Form-focused instruction in second language vocabulary learning: a case for contrastive analysis and translation. Applied Linguistics, 29, 694e716. Laufer, B., & Hulstijn, J. H. (2001). Incidental vocabulary acquisition in a second language: the construct of task-induced involvement. Applied Linguistics, 22, 1e26. Laufer, B., & Waldman, T. (2011). Verb-noun collocations in second language writing: a corpus analysis of learners' English. Language Learning, 61, 647e672. Levelt, W. (1999). Models of word production. Trends in Cognitive Sciences, 3(6), 223e232. Lewis, M. (1993). The lexical approach. Hove, UK: LTP. Li, J., & Schmitt, N. (2010). The development of collocation use in academic texts by advanced L2 learners: a multiple case study approach. In D. Wood (Ed.), Perspectives on formulaic language: Acquisition and communication (pp. 22e46). New York: Continuum. Lindstromberg, S., & Boers, F. (2008a). The mnemonic effect of noticing alliteration in lexical chunks. Applied Linguistics, 29, 200e222. Lindstromberg, S., & Boers, F. (2008b). Phonemic repetition and the learning of lexical chunks: the mnemonic power of assonance. System, 36, 423e436. McCarthy, M., & O'Dell, F. (2002). English idioms in use. Cambridge: Cambridge University Press. McCarthy, M., & O'Dell, F. (2003). English idioms in use. Advanced. Cambridge: Cambridge University Press. McCarthy, M., & O'Dell, F. (2005). English collocations in use: Intermediate. Cambridge, England: Cambridge University Press. McCarthy, M., & O'Dell, F. (2008). English collocations in use: Advanced. Cambridge, England: Cambridge University Press. Millar, N. (2010). The processing of malformed formulaic language. Applied Linguistics, 32, 129e148. Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of processing versus transfer appropriate processing. Journal of Verbal Learning and Verbal Behavior, 16, 519e533. Nation, I. S. P. (2013). Learning vocabulary in another language (2nd ed.). Cambridge: Cambridge University Press. Nattinger, J. R., & DeCarrico, J. S. (1992). Lexical phrases and language teaching. Oxford: Oxford University Press. Nelson, D., & Garland, R. (1969). Amount and locus of stimulus-response overlap in paired associate acquisition. Journal of Experimental Psychology, 82, 297e300. Nesselhauf, N. (2005). Collocations in a learner corpus. Amsterdam: John Benjamins. Pawley, A., & Syder, F. H. (1983). Two puzzles for linguistic theory: native-like selection and native-like fluency. In J. C. Richards, & R. W. Schmidt (Eds.), Language and communication (pp. 191e225). London: Longman. Peters, E. (2009). Learning collocations through attention-drawing techniques: a qualitative and quantitative analysis. In A. Barfield, & H. Gyllstad (Eds.), Researching collocations in another language: Multiple perspectives (pp. 194e207). Basingstoke, UK: Palgrave Macmillan. Peters, E. (2012). Learning German formulaic sequences: the effect of two attention-drawing techniques. Language Learning Journal, 40(1), 65e79. Peters, E. (2016). The learning burden of collocations: the role of interlexical and intralexical factors. Language Teaching Research, 20(1), 113e138. Rubin, D. (1995). Memory in Oral Traditions: The Cognitive Psychology of Epic, Ballads, and Counting-out Rhymes. Oxford University Press. Sagarra, N., & Alba, M. (2006). The key is in the keyword: L2 vocabulary learning methods with beginning learners of Spanish. The Modern Language Journal, 90, 228e243. Schmidt, R. W. (1990). The role of consciousness in second language learning. Applied Linguistics, 11, 129e158. Schmidt, R. W. (2001). Attention. In P. Robinson (Ed.), Cognition and second language instruction (pp. 3e32). Cambridge: Cambridge University Press. Schmitt, N. (2008). Instructed second language vocabulary learning. Language Teaching Research, 12, 329e363. Shapiro, A. M., & Waters, D. L. (2005). An investigation of the cognitive processes underlying the keyword method of foreign vocabulary learning. Language Teaching Research, 9, 129e146. Skehan, P. (1998). A cognitive approach to language teaching. Oxford, UK: Oxford University Press. Sonbul, S., & Schmitt, N. (2013). Explicit and implicit lexical knowledge: acquisition of collocations under different input conditions. Language Learning, 63, 121e159. Steinel, M. P., Hulstijn, J. H., & Steinel, W. (2007). Second language idiom learning in a paired-associate paradigm: effects of direction of learning, direction of testing, idiom imageability, and idiom transparency. Studies in Second Language Acquisition, 29, 449e484. Stengers, H., Boers, F., Housen, A., & Eyckmans, J. (2011). Formulaic sequences and L2 oral proficiency: does the type of target language influence the association? International Review of Applied Linguistics, 49, 321e343. Watkins, M., Watkins, O., & Crowder, R. (1974). The modality effect in free and serial recall as a function of phonological similarity. Journal of Verbal Learning and Verbal Behavior, 13(4), 430e447. Webb, S., & Kagimoto, E. (2011). Learning collocations: do the number of collocates, position of the node word, and synonymy affect learning? Applied Linguistics, 32, 259e276. Webb, S., Newton, J., & Chang, A. C.-S. (2013). Incidental learning of collocation. Language Learning, 63, 91e120. Wei, Z. (2015). Does teaching mnemonics for vocabulary learning make a difference? Putting the keyword method and the word-part technique to the test. Language Teaching Research, 19, 43e69. Wolter, B., & Gyllstad, H. (2011). Collocational links in the L2 mental lexicon and the influence of L1 intralexical knowledge. Applied Linguistics, 32, 430e449. Wood, D. (2010). Formulaic language and second language speech fluency: Background, evidence and classroom applications. New York: Continuum. Wray, A. (2002). Formulaic language and the lexicon. Cambridge, UK: Cambridge University Press. Yamashita, J., & Jiang, N. (2010). L1 influence on the acquisition of L2 collocations: Japanese ESL users and EFL learners acquiring English collocations. TESOL Quarterly, 44, 647e668.