1.1 Learners' Individual Differences

Learners vary in their cognitive, social, affective, and motivational characteristics (^{Dörnyei, 2009}); thus, exploration of the effects of individual differences has recently increased in studies regarding second/foreign language (L2) learning and processing (^{Brothers et al., 2021}; ^{Cheng et al., 2021}; ^{De Wilde et al., 2021}; ^{Elgort and Warren, 2014}; ^{García-Castro, 2015}; ^{Marecka et al., 2020}; ^{Montero Perez, 2020}). However, until now much research has not explored whether, or how, verbal fluency capacity facilitates L2 word learning in adults. For instance, most recent studies on L2 vocabulary learning have researched learners' individual differences in working memory (^{Ansari̇n and Khabbazi̇, 2021}; ^{Lee and Lee, 2021}), vocabulary knowledge (^{Elgort et al., 2018}; ^{García-Castro, 2015}, ²⁰²⁰; ^{Montero Perez, 2020}), phonological awareness (^{Marecka et al., 2020}; ^{Zhang and Roberts, 2021}), verbal fluency in children (^{Friesen et al., 2021}), and verbal fluency in bilingual aphasia (^{Carpenter et al., 2020}). Research on verbal fluency and its possible influence on L2 word learning in adulthood is needed.

Previous studies have demonstrated that verbal fluency capacity is associated to language learning and processing (^{Friesen et al., 2021}; ^{Lam and Marquardt, 2021}; ^{Luo et al., 2010}; ^{Rommers et al., 2015}; ^{Wang et al., 2021}); however, what is less clear is if, or how, verbal fluency capacity contributes to L2 vocabulary learning in adult Spanish native speakers. Thus, it is necessary to explore the possible influence of verbal fluency capacity in L2 vocabulary learners with adults to provide a deeper account of vocabulary learning in that population. Additionally, researching that possible influence may contribute to pedagogical practices and teaching processes to boost and improve first language (L1) and second language (L2) learners' vocabulary learning processes. This article addresses the research gap just mentioned by exploring the possible influence of verbal fluency capacity in adult vocabulary learning and it is divided into eight sections. The first section is the introduction, including the aims, research questions and hypothesis of the study, followed by a theoretical framework (second section). The third section describes the methodology, materials and instruments, the population that took part in the study, procedures, and how the data was processed and analyzed. Then, the fourth section consists of results and discussion, and the fifth the conclusion. The sixth section includes acknowledgements while the seventh references. The article finishes in the eighth section with the appendices.

1.2 Aims

The main aim of this article is to explore the possible influence of verbal fluency capacity in L1 and L2 vocabulary learning in university students studying in L1 English and in English as a medium of instruction (EMI) at a university in the United Kingdom. However, the article also aims to suggest pedagogical implications of the individual difference of verbal fluency capacity and its possible effect in L2 vocabulary learning.

1.3 Research Questions and Hypotheses

Two main research questions guided this study:

RQ1: Does verbal fluency capacity influence recognition scores of recently learned nouns for L1 and L2 learners?

H1: Due to the fact that verbal fluency is related to semantic memory (^{Reverberi et al., 2014}), participants with higher verbal fluency capacity will have higher recognition scores of novel nouns.

RQ2: Does verbal fluency capacity influence recall scores of recently learned nouns for L1 and L2 learners?

H1: Given that verbal fluency is related to semantic memory (^{Reverberi et al., 2014})

and that word recall challenges human cognition (^{Rutherford, 2006}), higher verbal fluency capacity facilitates recall of novel nouns.

2.1 Cognitive Individual Differences

Cognitive individual differences can be defined as “aspects of mental functioning, such as memorization and remembering; inhibiting and focusing attention; speed of information processing; and spatial and causal reasoning” (^{Robinson, 2012}, p.17). Researching L2 learners' cognitive individual differences can provide us with a more accurate understanding of learners' cognitive strengths that can be paired with targeted pedagogical practices (^{Granena et al., 2016}) to enhance their learning process. Thus, recent research is taking into account learners' individual differences in vocabulary learning. However, most research has mainly focused on the role of working memory and pre-existing vocabulary knowledge on L2 vocabulary learning. For instance, ^{Montero Perez (2020)} explored whether working memory and previous vocabulary knowledge contributed to vocabulary learning via viewing video. The author concluded that both working memory and previous vocabulary knowledge influence L2 vocabulary learning. Similarly, ^{García-Castro (2020)} researched the role of previous vocabulary knowledge on L2 word learning beyond factual memory and found out that greater vocabulary knowledge influences L2 vocabulary acquisition. ^{Hummel (2021)} explored whether phonological memory (PM) affects L2 vocabulary learning and found out that for high proficiency learners, PM does not seem to contribute as much as for beginner learners. More recently, ^{Liu et al. (2021)} highlighted that L2 vocabulary learning seems to be modulated by higher working memory capacity. A study deviating from the role of working memory and pre-existing vocabulary knowledge in L2 learning is that of ^{Marecka et al. (2020)}. They explored if auditory perception predicts word learning and found out that sensitivity to fast amplitude rise time contributes to word learning. The authors concluded that auditory perception may be a predictor of word learning.

The studies just mentioned highlight that there is a growing body of literature that recognises that cognitive individual differences are relevant on word learning processes. However, they also suggest that there is a lack of research on how, or if, verbal fluency capacity contributes to L2 vocabulary learning in adults.

2.2 Verbal Fluency

Verbal fluency is the cognitive capacity to retrieve information from memory (^{Patterson, 2011}) related to verbal ability (^{Lu and Zhang, 2007}). It demonstrates verbal ability (^{Wang et al., 2021}) and is associated with one's capacity to semantically categorise and retrieve concepts (e.g. vocabulary) from one's semantic memory or so-called “mental encyclopedia” (^{Harley, 2014}). When trying to retrieve information from the semantic memory, the mental processor employs a clustering mechanism to retrieve words from the same semantic subcategory activating verbal memory and word storage processes (^{Reverberi et al., 2014}; ^{Troyer et al., 1997}).Thus, in order to test an individual's verbal fluency capacity, clustering tests are used to ask participants to recall as many words as possible from a semantic cluster with time constraints, usually a minute (^{Reverberi et al., 2014}). In those tests participants use their lexical knowledge (^{Shao et al., 2014}), semantic memory (^{Troyer et al., 1997}), search strategies and conflict monitoring (^{Friesen et al., 2021}) which makes them cognitively challenging (^{Reverberi et al., 2014}) and indicators of a person's semantic memory (^{Soltani et al., 2021}). Verbal fluency tests can ask participants to recall words from the same semantic category (e.g. flowers) or from the same letter category (e.g. words beginning with the letter C). The former taps into semantic lexical retrieval processing that can shed light on how words are semantically organized in the mental lexicon. The latter also taps into retrieval processing but it might be more cognitive demanding as participants need to suppress responses related to semantic categorization to be able to only recall letter clusters (^{Friesen et al., 2021}). Previous research has demonstrated how verbal fluency is related to word learning and processing. For instance, ^{Rommers et al. (2015)} researched how verbal fluency capacity affects language prediction in a visual eye-tracking paradigm. Their study employed a semantic category task, with clusters of animals and professions, to test for verbal fluency. Participants had a total of 60 seconds to recall as many words as possible from every cluster. Overall, their results revealed that linguistic word prediction was associated with higher verbal fluency capacity, thereby suggesting that verbal fluency capacity may facilitate language predictive processes. Similarly, ^{Luo et al. (2010)} researched verbal fluency capacity in bilingual processing. They tested verbal fluency via a verbal category test and verbal phonemic fluency tests. Their results revealed a relationship between verbal fluency and word retrieval processes. More recently, ^{Lam and Marquardt (2021)} tested L1 and L2 speakers' verbal fluency capacity when retrieving emotional and non-emotional vocabulary. They discovered that, overall, L2 learners have lower verbal productivity when compared to L1 learners when retrieving both emotional and non-emotional vocabulary. In terms of the possible differences on verbal fluency capacity between bilinguals and monolinguals, ^{Wang et al. (2021)} showed that some bilinguals seem to have lower verbal fluency capacity than monolinguals and that verbal fluency capacity in bilinguals might depend on individuals' phonological loop. Similarly, ^{Friesen et al. (2021)} researched predictors of verbal fluency in monolingual and bilingual children and found out that English receptive vocabulary and fluid intelligence may mediate bilingual verbal fluency capacity.

The studies reviewed above highlight a link between verbal fluency capacity and language processing. Additionally, the studies also demonstrated the lack of empirical evidence on whether, or how, verbal fluency capacity facilitates L1 and L2 novel word learning in adult learners. Thus, this study aims to fill that gap by exploring the possible effects of the individual difference of verbal fluency in novel L1 and L2 noun learning in adults.

2.3 Verbal Fluency and Word Retrieval

Verbal fluency tasks consist of asking participants to retrieve words from their mental lexicon. Therefore, for the purposes of this research it becomes necessary to briefly describe word retrieval processing.

In retrieval processes, individuals find, activate, and process memory traces in their mental lexicon to retrieve words (^{Rutherford, 2006}). Similarly, in verbal fluency tasks participants activate words that are already established in their mental lexicon and make them ready to be used. However, in verbal fluency tests, participants may retrieve only semantic information even when the mental lexicon holds semantic, syntactic, and phonological components of the words that comprised it (^{Levelt, 1998}).

Multiple authors (^{Caramazza, 1997}; ^{Levelt, 1992}; ^{Levelt et al., 1999}; ^{Roelofs, 1997}, ²⁰⁰⁴) have researched how words are retrieved from the mental lexicon and they have mentioned that in word retrieval processes morphosyntactic, syntactic, phonological, and semantic aspects of the words can be retrieved. The aim of this article is not to discuss different word retrieval models proposed up to today; however, a brief description of ^{Levelt et al. (1999)} model will follow, given that it taps into speech production like verbal fluency tests do.

2.3.1 ^{Levelt et al.'s (1999)} Word Retrieval Model

^{Levelt et al. (1999)} model is based on speech production and retrieval, and it sheds light on general linguistic retrieval mechanisms. The model proposes a stage process that starts at conceptual preparation and ends in the initiation of the word's articulation (^{Levelt et al., 1999}). Speakers first select a syntactic unit, which leads them to a morphological unit and, finally, to a phonological unit that will come out as speech articulation. The first step in this model corresponds to conceptual activation tapping into lemma activation which in turn leads to lexical competition between semantically related items. For instance, a concept such as “dog” will activate the lemma “dog”, but it may also activate other lemmas such as “chihuahua”. In the end, only one lemma is selected. In order to select a lemma, the lexical concept activates its lemma node (^{Levelt et al., 1999}) and once a lemma is selected, its morphological characteristics (e.g. tense, mood, or number) are activated (e.g. plural form “dogs”), which is called morphological encoding. After this, the phonological properties start to activate also in relation to its morphological characteristics and an online phonological syllabification process starts. This process accounts for how the phonological properties convert into a phonological word. In order to be able to retrieve the phonological properties of the word (in speech production) the speaker has to undergo a series of processes to finally end up in production itself. Figure 1 illustrates ^{Levelt et al. (1999)} model of word retrieval.

Source:^{Levelt et al. (1999}, p. 3).

Figure 1  ^{Levelt et al. (1999)} model of word retrieval 

As shown in Figure 1, word retrieval processing is a complex and demanding task involving different processes that lead to spoken word retrieval.

3.1 Approach

This is an experimental quantitative research project, conducted from March to May 2018 that explored the possible influence of verbal fluency capacity on L1 and L2 learning of novel nouns through sentence reading. Learners' verbal fluency capacity was tested via a semantic categorization task (^{Rommers et al., 2015}). Offline vocabulary post-tests on meaning recognition and recall of the novel nouns were used to account for participants' word learning.

3.2 Participants

Twenty-nine L2 adult advanced learners of English (female=15; M age 30.82; SD=6.6) and 29 English native speakers took part in the study (female=20; M age 22.46; SD=5.6). Participants were studying at a university in the United Kingdom and they had various nationalities. L1 participants had British nationality while L2 participants' nationalities were all from Latin American countries. They were all recruited by email in which L1 and L2 learners were asked to take part in the study. L2 participants were selected based on their L1 (Latin American Spanish was a must). Participants' overall L2 proficiency was high (M=7.23) as reported in their IELTS scores (M=7.23; SD=0.45) which placed them in an advanced proficiency band (i.e. C1) according to the Common European Framework of Reference (CEFR). L1 participants had to be monolinguals in order to be chosen for the study as they acted as a comparison group (^{Pellicer-Sánchez, 2015}).

Participants agreed to take part in the study by signing an informed consent containing the study's aims, implications, and ethical considerations. None of the participants reported having learning difficulties. Participants took a verbal fluency test (Table 1) to determine their verbal fluency capacity. No comparisons between the learners' scores were made and none of them were excluded due to the tests' results or their cognitive capacities.

3.3 Materials and Instruments

A set of seven novel nouns (Appendix 1 ) served as linguistic stimuli for the study. Participants read English sentences (n=168), each of which contained existing English words and one novel noun. Each novel noun was used in 24 sentences. Learning of the novel nouns was tested via offline meaning recognition and meaning recall vocabulary post-tests based on ^{García-Castro (2015)} and ^{Webb (2008)}. Meaning recognition tests were used to determine whether participants were able to recognise the novel nouns in decontextualized contexts. The test included the seven target nouns and filler items. A meaning recall test was administered to find out if participants had sufficient knowledge of the novel nouns. The test included the seven target nouns and filler items. Offline recognition and recall tests have been previously employed in L1 and L2 vocabulary studies (^{García-Castro, 2015}, ²⁰²⁰; ^{Pellicer-Sánchez, 2016}; ^{Pellicer-Sánchez, and Schmitt, 2010}; ^{Webb, 2007}, ²⁰⁰⁸); thus, they can be suitable for this study. The recall test was administered first followed by the recognition test.

The verbal fluency task was based on ^{Rommers et al. (2015)}. Participants had one minute to recall, in their L1, as many words from the semantic cluster of animals as they could. L2 participants were tested in their L1 given that bilinguals are likely to have less L2 receptive vocabulary than L1 receptive vocabulary and this is likely to interfere in their performance in the verbal fluency task (^{Friesen et al., 2021}). Even though verbal fluency tasks are not difficult to administer and they usually last one minute, they are effective to test semantic fluency and they are cognitive demanding (^{Reverberi et al., 2014}). For this study, a semantic category task was administered in line with the study's aims. Participants' answers were analysed with the help of two adult Spanish native speakers and two English native speakers who did not take part in the study. Words that were repeated and that did not belong to the animal cluster were excluded from the analysis. Inter-rater reliability was calculated in R using the irr package (Gamer, et al., ²⁰¹⁹). Reliability had an almost perfect agreement with Cohen's Kappa of 0.97 for the L2 verbal fluency scores and high with Cohen's Kappa of 0.98 for the L1 verbal fluency scores.

3.4 Data Processing and Analysis

Data was processed and analysed using the statistical package R (^{R Studio Development Core Team, 2020}). First, normality checks (i.e. skew and kurtosis) were run to test the variables' normality. Given that all variables had normal distribution, two general linear models were created and tested with the “glm” package in the R environment. Two models were created for the analysis, model 1 included recognition scores as the outcome variable L1 group (English/Spanish), and verbal fluency as predictors, and model 2 included recall scores as the outcome variable L1 group (English/Spanish), and verbal fluency as predictors. An interaction between L1 and verbal fluency was specified in each module to explore whether verbal fluency had a direct influence on vocabulary learning and if it was modulated by language group.

4.1 Verbal Fluency Task

Results of the verbal fluency task show that L1 learners' verbal fluency capacity was slightly higher (M=27; SD=8) than that of L2 learners (M=25; SD=6); however, comparisons between the two groups were not made in this respect as it was not the purpose of the study. Additionally, monolinguals are more likely to outperform bilinguals in verbal fluency tasks (^{Friesen et al., 2021}) as bilinguals activate both languages when processing either of them. Bilingual cross-language interaction is a well documented phenomena highlighting that lexical representations of both languages simultaneously activate while reading, listening, speaking and/or writing either in the L1 or the L2 (^{Kroll and de Groot, 2005}). Thus, having slightly higher L1 results is not surprising.

4.2 Offline Vocabulary Post-tests

Results of the offline vocabulary post-tests are shown in Table 3.

As can be seen in Table 3, L1 learners performed better than L2 learners in both vocabulary post-tests. This observation was confirmed in the general linear models. A significant effect of first language (β =-31.54, SE= 9.38, Z=-3.362, p<0.001), and an interaction between verbal fluency and L1 (β = 1.14, SE=0.36 , Z=3.178, p<0.001) were found on the recognition scores. L1 learners (M=0.80; SD=0.40) outperformed L2 learners in their recognition scores (M=0.66; SD=0.50). In order to illustrate the result, two verbal fluency capacity groups were created based on a median split of participants' verbal fluency scores (L1 Mdn=29; L2 Mdn=25). The L1 higher verbal fluency group scored 4% higher than the lower group in their recognition scores (M=0.82; SD=0.39 vs. M=0.78; SD=0.41). The opposite effect was found on the higher L2 verbal fluency group as it scored 7% less than the lower group in their recognition scores (M=0.62; SD=0.49 vs. M= 0.69; SD=0.47) (Figure 2).

Source: Own elaboration based on participants' verbal fluency and meaning recognition scores, 2021

Figure 2 Verbal Fluency Capacity and Recognition Scores of Participants' Vocabulary Post-tests, 2018 

Recall scores showed a significant main effect of first language (β = -25.50, SE= 8.69, Z=-2.933, p<0.003) as L1 learners (M=0.51, SD=0.40) scored higher than L2 learners in the recall scores (M=0.37, SD=0.50), and a main interaction between verbal fluency and first language (β =0.83 , SE=0.33 , Z=2.511, p<0.012). Once again, participants were divided into high and low verbal fluency groups. It was found that L1 participants with higher verbal fluency capacity scored higher than those with lower capacities (M=0.53; SD=0.50; vs. M=0.47; SD=0.50). Likewise, L2 participants with higher verbal fluency capacity slightly performed better than those with lower capacities (M=0.39; SD=0.49; vs. M=0.36, SD=0.48) (Figure 3). Full list of results can be found in Appendix 2.

Source: Own elaboration based on participants' verbal fluency and meaning recall scores, 2021

Figure 3 Verbal Fluency Capacity and Recall Scores of Participants' Vocabulary Post-test, 2018 

4.3 Discussion

The empirical findings showed that L1 learners with higher verbal fluency capacity were better able to recognise the meaning of novel nouns than those with lower capacity. Similarly, L1 and L2 learners with higher verbal fluency capacity showed higher recall scores than those with lower capacities. Therefore, it can be suggested that verbal fluency capacity may influence L1 offline meaning recognition and L1 and L2 offline meaning recall of recently learned novel nouns.

In terms of the first research question of this study (i.e. Does verbal fluency capacity influence recognition scores of recently learned nouns for L1 and L2 learners?) it was found that on the recognition scores, L1 participants with higher verbal fluency capacity outperformed the lower group only by 4% (M=0.82; SD=0.39 vs. M=0.78; SD=0.41); thus, higher L1 verbal fluency capacity might be a predictor of L1 offline meaning recognition. However, due to the small difference between verbal fluency capacity groups this result has to be carefully considered. The possible relationship between L1 verbal fluency capacity and meaning recognition of novel words can be explained by the fact that processing semantic information activates the semantic lexical storage in the mental lexicon. Additionally, when one recognises a word while reading, semantic recognition processes start to activate whilst the mental processor matches what is being recognised to existing orthographic representations in the mental lexicon (^{de Groot, 2011}). Thus, those with higher capacities to process such information might be better able to activate and match semantic representations of existing words with those of novel words, resulting in more semantic recognition in vocabulary post-tests. It can be argued, then, that the recently learned novel nouns formed semantic memory traces in the mental lexicon (^{Bordag et al., 2017}), given that participants were able to not only recognise them in vocabulary post-tests but also to show greater recognition when they possessed higher verbal fluency capacity. In light of the results, the hypothesis that higher verbal fluency capacity facilitates recognition scores of the novel nouns is confirmed only for L1 learners. However, more research is needed to determine the full extent of verbal fluency capacity in novel word learning.

The second research question of this study asked if verbal fluency capacity influences recall scores of recently learned nouns for L1 and L2 learners. The results on the recall scores showed that L1 participants with higher verbal fluency capacity outperformed those with lower capacities by 6% (M=0.53; SD=0.50; vs. M=0.47; SD=0.50) on the recall test. Similarly, L2 participants with higher verbal fluency capacity slightly outperformed those with lower capacities by 3% (M=0.39; SD=0.49; vs. M=0.36, SD=0.48) on their recall scores. The findings indicate then that having higher verbal fluency capacity may facilitate L1 and L2 meaning recall of novel nouns. The strength of the representation of the novel words may have been sufficient to activate and/or select their meaning, making them ready for recall (^{Bordag et al., 2017}). Thus, participants with higher capacity to semantically recall words from the same semantic category may be better able to recall the meaning of the novel nouns. In meaning recall participants employ skills-based productive knowledge (^{Schmitt, 2010}) in either oral or written tasks. Possibly, those with higher verbal fluency capacity consolidated the semantic knowledge of the novel nouns more robustly than those with lower capacities, thus, they were more able to entail communicative productive knowledge than those with lower verbal fluency capacity. The mental lexicon holds not only general semantic, syntactic, phonological, and morphological information about words (^{Levelt, 1998}), but also specific information (^{Eysenck and Keane, 2015}). Therefore, one can argue that having greater verbal fluency capacity might possess the cognitive advantage of being able to store and activate semantic information of already established words in the mental lexicon more easily. For these reasons, participants with higher verbal fluency capacity might be better able to recall the meaning of novel words when compared to those with lower capacities. Additionally, word retrieval is a stage process that involves conceptual preparation that activates lemmas tapping into semantic activation, selection of a syntactic and of a morphological unit ending up into the phonological properties of those units which then lead to speech production (^{Levelt, 1998}). Hence, having greater verbal fluency capacity might be an advantage when activating and processing the semantic, syntactic, morphological, and phonological properties of words leading to more word recall. Overall, the hypothesis that higher verbal fluency capacity facilitates recall of novel nouns is confirmed for L1 and L2 learners.

One cannot exclude the possible effect of the type of test used in the findings of the study. Semantic clustering, which was used in this study, is less cognitive demanding that phonemic clustering (^{Friesen et al., 2021}); however, the aim of this study was to explore the possible effects of semantic memory on L2 vocabulary learning and semantic clustering is an indicator of semantic memory (^{Soltani et al., 2021}).