2.1. The pioneering study on WTC

MacIntyre et al. (1998) gave WTC research a new impulse. The authors argued that WTC in an L2 involves many uncertainties and complexities which are absent in WTC in an L1, and proposed a heuristic model to capture the range of potential variables affecting the volitional process to initiate communication in the L2, in a particular situation, with a particular person, at a particular moment in time (see Figure 1). The predictors of WTC include linguistic, psychological, and social variables, organized in the form of a pyramid, with L2 use at the top and WTC just underneath. Variables at the bottom of the pyramid are more stable and enduring than those higher up that focus on the here and now. Also, variables to the left of the bottom layers are more related to a macro intergroup perspective, while those to the right are more focused on the psychology of individual L2 learners and users.

The basis of the pyramid (Layer VI) captures enduring intergroup and genetic influences.³ Individuals have little influence over these factors, and generally they play a somewhat indirect role in language behaviour (MacIntyre, 2007, p. 567). The next layer (V) of the pyramid refers to the individual’s typical affective and cognitive context. MacIntyre explains that this layer sets “the tone for motivation to learn the L2”, namely “the tension between a desire to approach the target language group and a sense of hesitation or fear of the implications of doing so” (p. 567). It could be argued that this context is shaped within schools and FL classrooms in particular. The next level (Layer IV) “Motivational Propensities” marks a transition from more stable contextual characteristics to more variable and individual ones:

Roles and motives combine with L2 self-confidence; perceptions of communicative competence coupled with a lack of anxiety (p. 568).

The mention of “lack of anxiety” suggests that this is the level where the emotions are situated. Layer III of the pyramid refers to the individual in specific situations: “At this level of the pyramid model is the desire to communicate with a specific person as well as a state of self-confidence” (p. 568). Emotions are present at this level too (self-confidence in a specific situation). The next layer (II) refers to the behavioural intention to speak if one has the opportunity, or to remain quiet. Layer II, the WTC layer, “represents the final psychological step in preparation for L2 communication” (p. 568), while the top of the pyramid represents the actual communication in the L2. The pyramid model has been extremely influential in applied linguistics.

Since research in this area relies on self-report, WTC values reflect subjective self-evaluations of L2 users’ ability to communicate rather than objective measures (Gregersen & MacIntyre, 2014, p. 214).

2.2. Recent WTC research (from 2010)

Recent research is characterized by an interest in the stable and dynamic characteristics of WTC, and the complex interactions of the independent linguistic, situational and psychological variables. Interestingly, it often leaves out macro interethnic variables such as attitudes towards the FL or its speakers. The studies included in this section have been grouped in two categories: firstly, studies focused on identifying multiple sources of individual differences in WTC, organised chronologically, and, secondly, intervention studies where the aim was to test pedagogical practices that boost WTC.

Peng and Woodrow (2010) adopted the pyramid model as their theoretical basis in their investigation of WTC among 579 Chinese EFL learners. They used structural equation modelling to identify the best predictors of WTC. Predictor variables included communication anxiety in English, perceived communication competence in English, intrinsic and extrinsic motivation to learn English, learner beliefs, and classroom environment. In other words, they did not include variables linked to the bottom layer of the pyramid and focused more on the right side of the pyramid, and did not include attitudinal variables. Communication confidence (a combination of perceived competence and a lack of anxiety) was the strongest predictor of variance in WTC (54%), followed by motivation (14%) and learner beliefs (11%), which was itself predicted by classroom environment (p. 858).

MacIntyre and Doucette (2010) adopted Kuhl’s (1987) theory of action control. The authors adopted Kuhl’s three key concepts, preoccupation, volatility, and hesitation, arguing that these form part of the base from which L2 WTC is built. They investigated WTC, both inside and outside of the FL class, of 238 Anglo-Canadian high school learners of French. Using path analysis, they found that WTC inside class was negatively predicted by language anxiety and positively predicted by perceived competence. A negative third predictor emerged, volatility, characterized as a tendency to abandon tasks and remain silent. Less volatile learners typically engaged in tasks with more persistence and were more willing to speak in the L2.

Cao (2011) took the MacIntyre et al. (1998) paper as a starting point to develop an “ecological perspective” in WTC research, where the notion of context is central, namely “a web of intertwining relationships between students, teachers and their surrounding micro classroom contexts and macro institutional environments” (p. 469). In her multiple case study approach, she observed a total of 18 advanced Chinese and Korean EFL students in New Zealand during classroom interactions, complemented by stimulated-recall interviews and data from reflective journals. She lists “self-confidence, personality, emotion and perceived opportunity to communicate, classroom environmental conditions such as topic, task, interlocutor, teacher and group size, together with linguistic factors” as affecting her participants’ WTC (p. 468). Her conclusion is that while all these factors can contribute to WTC, their influence depends on the occasion as some may override others at certain times during classroom interactions.

Denies, Yashima and Janssen (2015) compared the predictors of classroom WTC with those of societal WTC in a representative sample of 1117 Dutch L1 Grade 12 students studying French L2 in Flanders. Using structural equation modelling, the authors found that classroom WTC is a strong predictor of WTC outside the classroom. Integrativeness and perceived competence emerged as positive predictors of classroom WTC while anxiety had a small negative impact (p. 730). Similar patterns were found for WTC outside the classroom, with a smaller role for integrativeness. The authors conclude that MacIntyre et al.’s (1998) WTC model seems “to be generalizable to the specific context of French L2 learning in Flemish secondary schools” (p. 734). They encourage FL teachers to bolster students’ sense of perceived competence in French and reduce their anxiety in using the language.

Teimouri (2017) expanded the range of WTC research by including concepts from motivation research and from emotion in SLA studies. Combining factor analysis, partial correlational analyses and multiple regression analyses, he investigated the WTC of 524 Iranian EFL high school students, linking it to L2 selves and anxiety, joy and shame. He considered WTC and intended effort as two qualitatively different forms of motivation (p. 6). The analyses revealed that Ideal L2 self (“the attributes that someone would ideally like to possess”, Dörnyei & Ryan, 2015, p. 87) predicted WTC. However the Ought-to-L2 self (“the attributes that one believes one ought to possess”, Dörnyei & Ryan, 2015, p. 87) was not a predictor. Ideal L2 self and Ought-to-L2 self predicted joy but only Ought-to-L2 self predicted anxiety and shame. Teimouri argued that learners with strong Ideal L2 self have a more preventional focus which plays a facilitative role as it keeps them alert to the presence of possible negative outcomes. However, these same learners have a predominant promotional focus and are very much aware of positive outcomes. Joy best fits their motivational orientation while anxiety represents a misfit, as it weakens motivation.

Khajavy, MacIntyre and Barabadi (2017) did not include motivation concepts in their research design to investigate WTC, but did include emotions and classroom environment. They used the advanced quantitative procedure of doubly latent multilevel analysis. Their participants were 1528 secondary school students (aged 12–18) in Iran who reported moderate levels of WTC and higher levels of enjoyment than anxiety. The researchers found that positive classroom environment (teachers’ support, students’ cohesiveness, and task orientation) was positively linked to WTC and enjoyment, and led to reduced anxiety among students. Moreover, enjoyment was found to be an important factor in increasing WTC at both student level and classroom level, while anxiety reduced WTC only at the student level (p. 14). Girls were found to be slightly more anxious and more willing to communicate in English.

The effect of the classroom social climate on WTC was also central in the study by Joe, Hiver and Al-Hoorie (2017) on 381 Korean learners of English L2. The researchers used structural equation modelling to consider the link between classroom social climate (teacher academic support, teacher emotional support, classroom mutual respect), the satisfaction of learners’ basic psychological needs (autonomy, competence, and relatedness), autonomous forms of motivation (identified regulation and intrinsic motivation) and their combined effect on levels of WTC (with friends, acquaintances and strangers) and achievement (exam results). WTC was found to be predicted directly and indirectly by satisfaction of basic psychological needs, followed by classroom social climate, prior achievement and identified regulation, but not by intrinsic motivation (p. 139). The authors conclude that “WTC and L2 achievement are determined both individually (…) as well as situationally (…)” (p. 140).

The two final studies are intervention studies based in Japan that sought to boost learners’ WTC in English. Munezane (2015) recruited 373 Japanese university EFL learners, which were divided into a control group that received no treatment, and two experimental groups. The first experimental group combined visualization (i.e. seeing themselves as future specialists in their field) with goal setting (i.e. articulation of speaking goals in each class). The second experimental group engaged in the visualization exercises without goal setting. The three groups had the same amount of communicative activities. The combination of visualisation and goal setting group had significantly higher levels of WTC.

Yashima, MacIntyre and Ikeda (2018) carried out an interventional study over a semester with 21 Japanese EFL students in Japan. The intervention consisted of avoiding Initiation–Response–Feedback (IRF) patterns during discussion sessions in English in order to encourage students to initiate communication. The analysis of data from three participants showed how the frequency of self-initiated turns was linked to personality, proficiency, and group behavior.

2.3. Studies on Foreign Language Enjoyment and Anxiety

Dewaele and MacIntyre (2014) introduced the concept of FLE, which they combined with FLCA in order to catch both positive and negative emotions of a heterogeneous group of 1740 FL learners about the learning experience, peers and teacher. Higher levels of FLE and lower levels of FLCA were associated with older learners, a high level of multilingualism, FL proficiency, self-confidence, and education level. Female participants reported significantly more FLE and FLCA. Narrative feedback pointed to the crucial role of the classroom environment, including peers and teachers.

Delving into the same dataset, Dewaele and MacIntyre (2016) used a Principal Components Analysis to reveal three dimensions explaining nearly half of the variance, with two separate dimensions for the 21 FLE items, namely social and private FLE.

A final study on the same database looked more specifically at the gender differences in FLE and FLCA at item-level (Dewaele, MacIntyre, Boudreau & Dewaele, 2016). Female participants reported having significantly more fun in the FL class, agreed more strongly that they learned interesting things, and were prouder than the male peers of their FL performance. However, they worried significantly more than their male peers about mistakes and lacked confidence in using the FL.

In order to compare the influence of learner-internal and learner-external variables on levels of FLCA and FLE within a single national and educational context, Dewaele, Witney, Saito and Dewaele (2017) collected data from 189 secondary school pupils in London (the same corpus on which the present study is based) who were mostly studying French as a FL. Older pupils reported more FLE than younger ones. Female pupils scored higher on both FLE and FLCA. FLCA turned out to be much less related to teacher and teacher practices than FLE. Indeed, lower levels of FLCA were linked to positive attitudes towards the FL, higher relative standing among peers in the FL, and being more advanced in the FL. In contrast, higher levels of FLE were linked to more positive attitudes towards the FL, the FL teacher, frequent use of the FL by the teacher, more time spent by pupils on speaking, a higher relative standing among peers and being more advanced in the FL. The conclusion was that an effective teacher needs to focus on learners’ enthusiasm and enjoyment while creating a friendly low-anxiety classroom atmosphere.

In a final paper using the same database, Dewaele and Dewaele (2017) used a pseudo-longitudinal design to investigate how FLE and FLCA of FL pupils evolved over time. A comparison of FLE and FLCA values among 12–13 year olds, 14–15 year olds and 16–18 year olds showed little change in FLCA and a slight increase in FLE. Multiple regression analyses revealed that fewer independent variables predicted FLE and FLCA at the start and at the end of the secondary education compared to the middle phase. Levels of FLE and FLCA in the youngest group were linked to relative standing in the group and the level of mastery in the FL. In the middle group, FLE was predicted by attitude towards the FL, attitude towards the teacher, the teacher’s predictability (a negative predictor) and the number of languages known by the pupils. FLCA was predicted by relative standing in the group and the level of mastery in the FL. In the oldest group, FLE was only predicted by attitude towards the teacher while FLCA was predicted by relative standing in the group (a negative predictor) and teacher’s predictability. It thus seems that the causes of positive and negative emotions in the FL classroom are dynamic and change over time.

What emerges from the literature review is that there are many sources of WTC that can be situated at various levels in MacIntyre et al.’s (1998) multi-layered “pyramid”. The relative influence of the different layers may differ according to the geographical, political and historical contexts. The view of WTC as a stable personality trait has shifted to acknowledge its dynamic character, namely the interaction between learner-internal and learner-external variables. The role of learners’ emotions on WTC is also increasingly being investigated. No research so far has linked the three concepts of private FLE, social FLE, and FLCA to WTC, and looked at the combined effects of learner-internal and learner-external variables. This is exactly what we aim to do in the present study.

4.1. Foreign languages in UK secondary schools

The study of a FL is compulsory for 11 to 14 year olds in UK public-sector-maintained schools. For older pupils, FL classes have to be offered by the school but are no longer compulsory (https://www.gov.uk/national-curriculum/key-stage-3-and-4). FL pupils in British secondary school system take the national General Certificate of Secondary Education (GCSE) at ages 15 and 16. Good results are required for admission into the Sixth Form (final two years of school) and universities. Students who wish to pursue the study of FLs can do so at Advanced Level (A-Level), as part of the three or four elected subjects in the last two years of school, for which they usually sit the national A-level exams at the end of their schooling, aged 18. Universities typically make conditional offers to secondary school students who are in their final year, based on students’ personal statement, GCSE results, predicted A-level results and sometimes university entrance tests and interview performance. The competition is fierce as students aim for the most prestigious universities. A conditional offer for a language or linguistics degree from a high-calibre university is typically “AAA” or “A*AA”.

4.2. Participants

A total of 189 secondary school pupils (49 females, 140 males) participated in the study in 2015. They came from two schools based in Greater London: 63 pupils were from Dame Alice Owen’s, a semi-selective state school in Potters Bar, and 126 pupils were from Westminster School, an independent boarding and day school within the precincts of Westminster Abbey, which is selective and fee-paying. Both schools are amongst the top performing schools in the UK.

A majority of pupils were British (n = 156), often with double nationalities. One hundred and sixty-nine pupils reported having English as a first language (L1), which was often combined with other L1s. Close to a third of participants (n = 57) reported growing up with more than one language from birth.

Most participants were studying French as a FL (n = 144, 68%), while others were studying Spanish (n = 21), German (n = 15), with smaller numbers studying other FLs.

Participants were also asked about the point they had reached in their FL journey. These included “low intermediate” (n = 22), “intermediate”, (n = 76), “high intermediate” (n = 71), and “advanced” (n = 20).

Students compared their own FL performance with that of their peers in their FL class (ranging from “far below average” (n = 1), “below average” (n = 12), “average” (n = 58), “above average” (n = 92), and “far above average” (n = 26). The categories of “far below average and “below average” were merged. These values were positively correlated with self-reported results on their last major FL test (r (187) = .50, p < .0001). These test scores ranged from 49% to 100%, with a mean of 87.7% (SD = 10). In other words, these were excellent FL students.

4.3. The instrument

The questionnaire opened with a demographics section from which the above information was retrieved. Following this, participants were asked to respond to an item on their attitude towards their first modern FL (as some pupils learned two FLs simultaneously), on a 5-point Likert scale. Because very few reported “very unfavourable” attitudes, this level was merged with the next level, i.e. “unfavourable” attitudes (2), followed by “neutral” (3), “favourable” (4) and “very favourable” (5) attitudes. Mean score was 3.1 (SD = 0.9).

The next question asked about attitudes towards the FL teacher using a 5-point Likert scale (ranging from “very unfavourable” (1), “unfavourable” (2), “neutral” (3), “favourable” (4), to “very favourable” (5) attitudes. Mean score was 4.0 (SD = 1.0).

The following question focused on frequency of use of the FL in class by the FL teacher. Answers ranged from “Hardly ever” (1) to “Not very often” (2), “Sometimes” (3), “Usually” (4) and “All the time” (5). Mean score was 3.9 (SD = 1.0).

Pupils were then invited to complete 10 items, which were extracted from the FLE questionnaire (Dewaele & MacIntyre, 2014) (see appendix). They were chosen to capture the reliability of the original scale without sacrificing the reliability of the measurement. They included items reflecting the two FLE dimensions: Social FLE and Private FLE (Dewaele & MacIntyre, 2016). They were based on standard 5-point Likert scales with the anchors “Strongly disagree” = 1, “Disagree” = 2, “Neither agree nor disagree” = 3, “Agree” = 4, “Strongly agree” = 5. All items were positively phrased. A scale analysis of the Social FLE and Private FLE scales revealed high internal consistency (Cronbach alpha = 0.81 and 0.78 respectively). Mean score was 4.0 (SD = 0.6) for Social FLE and 3.8 (SD = 0.7) for Private FLE.

Another 8 items were extracted from the FLCAS and reflected physical symptoms of anxiety, nervousness and lack of confidence (Horwitz, Horwitz & Cope, 1986) (see appendix). They also captured the reliability of the original scale (Dewaele & MacIntyre, 2014). Two FLCA items were phrased to indicate low anxiety and six were phrased to indicate high anxiety. The low anxiety items were reverse-coded so that high scores reflect high anxiety for all items on this measure. A scale analysis of the whole dataset revealed high internal consistency (Cronbach alpha = 0.85). Mean score was 2.4 (SD = 0.8).

A final section contained 8 items about the likelihood to start a conversation in the LX in hypothetical scenarios with known and unknown interlocutors in a range of different situations (see appendix).⁴ Likert scales had anchors ranging from “Almost never willing” = 1, “Sometimes willing” = 2, “Willing half of the time” = 3, “Usually Willing” = 4, to “Almost always willing” = 5. Internal reliability was good (Cronbach alpha = .80). The mean score was 2.9 (SD = 0.6). Although a one-sample Kolmogorov-Smirnov test showed that the distribution was not normal because of slight skew toward the high end of the scale (KS Z = .074, p < 0.013), a Q-Q plot (quantile-quantile plot) showed a distribution that was close enough to normal (see Figure 2). We thus decided to use the more powerful parametric statistics.

The questionnaire was totally anonymous. After the research design and questionnaire obtained ethical approval from the first author’s institution, the headmasters of the two schools were contacted to obtain their approval. Consent was obtained in two stages: parents were contacted by the school to explain that their children would be invited to participate in a survey on affective variables in the FL classroom. They were invited to contact the researchers to obtain extra information. A couple of parents did so, and none opted out of the survey. Next, the parents received an email in which they were asked to invite their child to participate in the study. The pupils’ individual consent was obtained at the start of the survey. The questionnaire was posted online using Googledocs.