3.1 Public attitudes towards an early start of FL learning

Overall, most publications on pre-primary FL programs document enthusiasm (Ioannou-Georgio, Reference Ioannou-Georgiou, Mourāo and Lourenço2015), especially on the part of parents (e.g., Mourão, Reference Mourão, Garton and Copland2019). Their favorable attitudes towards early start programs are responsible for the mushrooming of pre-primary English programs in Europe, Asia, South America, and other places. Parents tend to believe that early stimulation is conducive to learning English and the earlier their offspring starts, the better they will be over time. Parental pressure can, therefore, oblige preschools to offer such courses to meet these expectations. Decision makers tend to accede to these demands although teacher shortage is often an issue. The dynamics play out in specific contexts in different ways. In South Korea, for example, Kiaer et al. (Reference Kiaer, Morgan-Brown and Choi2021) found positive relationships between how much English was used in two types of kindergarten programs and the children's level of anxiety. Parents with higher ambitions placed their children in more intensive (immersion-like) programs. The children tended to be more anxious than their peers in programs with less exposure to and insistence on “English only”. In Poland, Rokita-Jaśkow (Reference Rokita-Jaśkow2013, Reference Rokita-Jaśkow2015) also revealed a complex picture resulting from interactions across parents’ aspirations, their choice of public or private preschools, English teachers’ beliefs and competences, and differences in children's learning opportunities. Parents’ previous language learning experiences impacted their decisions as to how early they wanted their offspring to start English and in what program: the less successful as L2 learners the parents themselves were, the more they believed in earlier and more intensive programs (Rokita-Jaśkow, Reference Rokita-Jaśkow2015).

3.2 Pre-primary programs

Despite the large number of publications on early FL teaching and learning, little is known about preschool programs. As the headings in this review on the emerging themes in studies indicate, authors offered many insights into narrow aspects, but we found no program evaluation or studies offering a clear picture of what the curricular aims were, how and to what extent they were achieved based on criteria. Owing to these reasons, few studies are replicable or offer details of programs for adoption. Some examples of good practice were found. For example, valuable documents are shared by the Goethe Institut on early German as a FL: detailed recommendations as to how programs should be organized (Widlok et al., Reference Widlok, Petravić, Org and Romcea2011) and freely available age-appropriate teaching materials can be downloaded from their websites (https://www.goethe.de/ins/au/en/spr/unt/kum/dfk/fuf.html; https://www.goethe.de/ins/pl/de/spr/unt/kum/dfk/dms.html). No empirical study was, however, found on how early German or other L2 programs work over time.

A few publications compare children's development in immersion and FL programs (e.g., Kiaer et al., Reference Kiaer, Morgan-Brown and Choi2021). Others offer insights into projects aiming to raise children's awareness of additional languages (e.g., Lourenço & Andrade, Reference Lourenço and Andrade2013), and examine how children's attitudes to English and their grades were impacted by preschool programs (Chen et al., Reference Chen, Zhao, de Ruiter, Zhou and Huang2020). Such studies are discussed under the specific focal points below, but first we present problems related to transition from pre-primary to primary school.

3.3. Transition issues

Preschool education has created a new transition in children's lives. The need to investigate transitional processes is often stressed (e.g., Nikolov, Reference Nikolov and Nikolov2009), but such studies are still rare. This early transition occurs at different learner ages and raises issues at different levels, which are solved in different ways in different contexts (OECD, 2017). General educational projects (e.g., Carida, Reference Carida2011; O'Kane, Reference O'Kane2016) call for national transition policies, cooperation between preschools and primary schools, introduction of children's portfolios with information on their transition experiences, and suggest that widespread problems of inconsistency and discontinuity can be solved only by involving all stakeholders.

In the FL education field, problems in the preschool to primary school transition emerge in domains of FL policies, school and curriculum, classroom, teacher, parents, and young learners themselves (Mihaljević Djigunović & Letica Krevelj, Reference Mihaljević Djigunović, Letica Krevelj and Schwartz2022). European documents (e.g., European Commission, 2011) emphasize the key importance of continuity and coherence. However, there are many challenges regarding these. In some contexts (e.g., parts of Germany, Estonia), securing continuity may be difficult because the FL learnt in preschool may not be offered in a particular primary school, or it may not be offered from primary grade one (European Commission, 2017).

Coherence, too, often emerges as a problem where teaching approaches at the two levels are not consistent and logical to children. Suggested ways of ensuring coherence are structural integration (with preschool and primary school as part of the same educational unit), conceptual integration (basing the two curricula on consistent knowledge and values) and parent–teacher partnership integration.

McElwee (Reference McElwee, Mourão and Lourenço2015) found that preschool learners of French in England, who were exposed to the Narrative Format teaching approach using a series of stories focusing on children's common experiences, developed language learning strategies that led to higher metalinguistic awareness in grade three of primary school; such level of awareness was very rare in FL beginners at that age. He concluded that introducing FL at preprimary could secure continuity and progression during transitioning to primary school. A Chinese study (Chen et al., Reference Chen, Zhao, de Ruiter, Zhou and Huang2020) into effects of preschool English as a foreign language (EFL) on achievement and attitudes to EFL and L1 Chinese in primary grades one and three found a positive relationship between the included variables. In their comparative study of Croatian primary EFL learners who had learned EFL at preschool and joined classmates who were total beginners, Letica Krevelj and Mihaljević Djigunović (Reference Letica Krevelj and Mihaljević Djigunović2021) found that their attitudes and motivation differed. While the EFL teacher was the main motivator in the case of total beginners, this was not the case for non-beginners: in their case the teacher's motivating role was less prominent and they were more motivated by a sense of progress. However, some non-beginners reported being demotivated by the mismatch of their self-perceived EFL proficiency and the input they were exposed to in the primary classroom.

Understanding the complexity of the pre-primary to primary transition in FL learning requires both extensive and in-depth study. A recently proposed research framework (Mihaljević Djigunović & Letica Krevelj, Reference Mihaljević Djigunović, Letica Krevelj and Schwartz2022, p. 632) involves all participating actors and their interactions, includes both macro and micro aspects, and is nested in the social context where transition is taking place. The framework implies that transition begins before children leave preschool and ends when they have settled in their new environment, which means that its duration may be different for different children.

4.1. Children's individual differences

Research suggests that preschoolers differ in many characteristics impacting FL learning behavior and achievements. Some IDs play a more crucial role in preschool FL learning than in the case of older learners. One such ID is temperament (mood, adaptability, level of responsiveness). Positive mood, higher levels of adaptation, activity, and initial reactions were found to correlate with a greater number of verbal and non-verbal reactions of Chinese EFL preschool learners; negative mood coupled with lower number of initial reactions was observed to cause teachers’ misinterpretation of children's needs, which then negatively affected learner motivation and incidental learning (Sun et al., Reference Sun, de Bot and Steinkrauss2015a, Reference Sun, Steinkrauss, Tendeiro and de Bot2015b).

Aptitude emerged as an important ID in vocabulary and grammar acquisition. Greek young learners’ scores on vocabulary tests varied significantly depending on their cognitive (analytic reasoning ability, short term memory) and phonological skills (phonetic repetition, distinction of non-words) (Alexiou, Reference Alexiou and Nikolov2009). Chinese preschool learners’ phonological short-term working memory was found to be a significant predictor of their vocabulary scores (Unsworth et al., Reference Unsworth, Persson, Prins and de Bot2014). Analytical reasoning correlated with receptive grammar, while short-term memory correlated with productive vocabulary learning (Sun et al., Reference Sun, Steinkrauss, van der Steen, Cox and de Bot2016).

Motivation is a key factor in FL learning. Preschoolers’ motivation is based on their positive attitudes to FL (Griva & Sivropoulou, Reference Griva and Sivropoulou2009), their FL teacher (Ahn & West, Reference Ahn and West2018), and they are influenced by their parents and peers (Jin et al., Reference Jin, Zhou, Hu, Yang, Sun and Zhao2016). Children are intrinsically motivated for FL learning. Maintaining their intrinsic motivation calls for a predictable learning environment, moderately challenging tasks, instructional support, assessment focusing on self-improvement, and learner autonomy in deciding on FL content and outcomes (Wu, Reference Wu2003). Pre-primary FL learners can develop self-motivation if their teacher fosters their self-esteem, self-confidence, co-operation with other learners, and praises their efforts (Brumen, Reference Brumen2011).

Vičević Ivanović et al. (Reference Vićević Ivanović, Košuta and Patekar2021) studied FL learning strategies of Croatian very young learners of French, German, and Italian. They used the projection technique to elicit information on children's strategies: children were asked how they would help their plush toy to say particular words and structures in the L2 (e.g., How would you help the elephant to learn to say ‘der Apfel’?). The findings pointed to high frequency of memory (e.g., model repetition, listening to the interlocutor) and informal strategies (e.g., showing and naming the object, exposure to media). Frequency of some of the identified strategies varied with the intensity of the participants’ FL exposure; thus, children with a more intensive L2 exposure tended to use informal strategies (learning through songs and rhymes) and social strategies (contact with a native speaker). Some formal strategies (e.g., learning by drawing or writing words) were not connected with intensity of L2 exposure.

When the importance of positive learner attitudes and motivation is ignored, children may develop FL anxiety. Insistence on rote memorization or native speaker pronunciation are some of the sources of preschool EFL learners’ anxiety found in the Korean context (Kiaer et al., Reference Kiaer, Morgan-Brown and Choi2021). Using age-appropriate teaching practices and developing FL enjoyment are suggested as ways of creating anxiety-free classrooms. In another example on low anxiety, kindergarten children were assessed with a puppet theatre twice a year in Serbia (Prošić-Santovac & Navratil, Reference Prošić-Santovac, Navratil, Prošić-Santovac and Rixon2019). Analyzing videotaped observations and interviews, the authors found that children happily engaged in the tasks administered in English and Serbian, they were not anxious, and were willing to interact. How the children performed the tasks in English was not analyzed; thus, it is unclear how detailed, valid, and reliable the assessments were.

Findings on family socio-economic status (SES) indicate that SES is both a mediating cognitive factor (securing access to FL learning resources) and a mediating affective factor (through parental involvement and encouragement) (Rokita-Jaśkow, Reference Rokita-Jaśkow2013). SES emerged as a good predictor of productive vocabulary development (Sun et al., Reference Sun, Steinkrauss, van der Steen, Cox and de Bot2016); however, SES did not play a key role in studies on the youngest age group (Ferjan Ramírez et al., Reference Ferjan Ramírez, Sheth and Kuhl2021).

4.2. Children developing awareness about language, sounds, and pragmatics

Studies on preschool FL learners’ perception of the new language they are learning focused on phonological, phonemic, and pragmatic awareness. Kearney and Ahn (Reference Kearney and Ahn2013) studied children engaging with multiple languages (Chinese, Korean, and Spanish) in minimal input preschool classrooms in the United States. They observed how children's language awareness developed in teacher-led lexically focused activities. Kearney and Barbour (Reference Kearney and Barbour2015) used a socialization approach: they observed how children became aware of languages different from their own, how they related to and joined activities, and how they saw themselves as children who knew a few words in multiple languages. In other words, they documented how children started on the journey of becoming language learners by participating in, thereby socializing into, language-related activities. The authors analyzed how children's language awareness developed through engagement with the three languages, and they documented their non-linear process of developing awareness and acceptance of differences and diversity in languages.

In their experimental study, Lourenço and Andrade (Reference Lourenço and Andrade2013) used the ‘awakening to languages approach’ (exposing children to a variety of languages to discover and explore new sounds and writing systems) with Portuguese children (age three–six) and found that the approach helped increase preschoolers’ phonological awareness in their L1: the children developed interest in observing and reflecting about sounds and language. Being able to think about and analyze units of speech in L1 is considered a steppingstone for successful development of vocabulary, reading, and writing in a FL, too.

Phonemic awareness was the focus of a study by Martinez et al. (Reference Martinez, Coyle, De Larios, Murphy and Evangelou2015), in which they tested effects of explicit instruction of English sibilant phonemes on Spanish preschoolers. Their findings showed a positive effect of explicit phonics teaching. Interestingly, using actions accompanying pronunciation of included words in the experimental group did not lead to higher acquisition of the sibilants. The authors suggest that actions need to be linked to the sounds in an explicit and unambiguous way to be a good enough support. Immigrant children for whom English was a third language were more successful than Spanish L1 children: their larger experience of L2 learning raised their phonemic awareness.

Portóles Falomir (Reference Portóles Falomir2015) investigated pragmatic awareness of preschool and primary school children in Spain who were learning Catalan, Spanish, and EFL. She found that by the age of four these multilingual preschoolers had already developed a certain level of pragmatic awareness (though lower than primary school learners). Their pragmatic awareness was not connected to language proficiency in any particular language, but to their multilingual proficiency (comprising interaction of the three language systems, cross-linguistic interaction, and the specific linguistic and cognitive skills exclusive to multilingual learning).

4.3. The acquisition of L2 morphological and morphosyntactic elements

Morphological and morphosyntactic studies focusing on preprimary FL learners are rather rare; only two publications examined these areas. Hillyard (Reference Hillyard, Murphy and Evangelou2015) carried out an action research project with four-year-old EFL learners in Argentina. Her aim was to see if using rhythmic patterns in stories would help her learners acquire and use correctly word order of adjective + noun constructions, which cause interference with Spanish word order. The four stories used in the project involved both rhythmic patterns and pictures, or only one of them, or none. The findings showed that the higher number of correct word order was related to stories which did not include rhythmic patterns. The author attributed this finding to the stronger impact of substituting vocabulary, personalizing the story and using language creatively.

McElwee (Reference McElwee, Mourão and Lourenço2015) investigated effects of using short narratives in teaching French to preschoolers in England. His findings showed that after listening to and watching stories, the participants were generally able to produce very short sentences and phrases, and that everyone produced more complete than incomplete utterances. Overall, the young learners’ morphosyntactic knowledge and skill became more sophisticated with increasing learner age as a result of working with stories in French.

4.4 Studies on vocabulary development

In very early FL learning, vocabulary is often considered as the foundation of children's FL competence in line with theories (see section 2). In this section, we overview what aspects of vocabulary studies have been inquired into. Focusing on lexical development of preschool EFL learners (age range 1; four–five years), Rokita (Reference Rokita2007) carried out two extensive studies, a longitudinal one and a cross-sectional one. Her longitudinal findings showed that Polish preschoolers could learn not more than 200 words in English during a two-year period, their vocabulary reflected the input they had been exposed to, and what they knew was unstable and prone to fast forgetting. Based on the cross-sectional data, the author suggested there might be a threshold of FL lexical knowledge that facilitates acquisition of further items. Four-year-olds progressed significantly faster than younger learners, and Rokita attributed this to their larger short-term memory capacity. The production level of all preschoolers was characterized by mechanical use of items learned in class, with children who had learned longer incorporating EFL words in L1 utterances. Spontaneous or creative use of single words or formulaic expressions was extremely rare. Best performances were related to high and persistent parental involvement in children's preschool FL learning in the form of, for example, practicing words and reading at home.

A study on Spanish children focused on which activities contribute to vocabulary knowledge. Albaladejo Albaladejo et al. (Reference Albaladejo Albaladejo, Coyle and de Larios2018) inquired into the benefits of stories, songs, and a combination of the two for vocabulary learning by preschool EFL learners (aged two–three). They found that stories were the most and songs the least conducive to vocabulary development and that the children who behaved in a ‘quietly attentive’ manner during the activities performed best.

Drawing attention to a methodological concern in studies claiming that songs contribute to vocabulary acquisition, Davis and Fan (Reference Davis and Fan2016) remarked that such studies do not include true control groups against which to compare results. Thus, their study involved three EFL classes of Chinese preschoolers aged four–five years. They were exposed to 15 selected vocabulary items with five out of the 15 items in each of the three conditions: song, choral repetition, or neither of the two as a control condition. The authors concluded that both songs and choral repetition have important pedagogical value in preschool children's FL vocabulary acquisition.

Authentic cartoons were used in two projects in Greece. Greek children's implicit vocabulary uptake of 21 nouns, including cognates, was assessed after watching four episodes of the popular British cartoon series Peppa Pig multiple times over a month followed by discussions in their mother tongue (Alexiou, Reference Alexiou, Gitsaki and Alexiou2015). Peppa Pig used slow, clear, child-directed speech and was easy to follow. Children could recall about half of the words, indicating that implicit learning worked well in highly contextualized age-appropriate activities in which meaning-making was supported both visually and in L1.

Explicit versus implicit teaching effects were the focus of Kokla's (Reference Kokla and Alexiou2021) experimental study with Greek preschool EFL learners. All children watched episodes of the Peppa Pig cartoon. The control group was tested after viewing, while the experimental group received instruction after viewing and was then tested on formulaic language used in the cartoon. The findings indicated that mere viewing was beneficial and raised the control group preschoolers’ test results. The even higher scores in the experimental group proved that explicit teaching of formulaic expressions reinforced their learning.

Yeung et al. (Reference Yeung, Ng and Qiao2020) compared effects of implicit and explicit teaching of EFL vocabulary and phonemic awareness on Chinese children between the ages of four and six in Hong Kong. The findings of their experimental study showed that explicit vocabulary teaching was more effective than implicit instruction in acquiring both productive and receptive vocabulary, with a stronger effect on productive vocabulary. They attributed the impact of explicit vocabulary teaching to learners’ increased attention and engagement with the target words. Applying a CLIL-inspired pedagogic approach to teaching EFL to Portuguese preschoolers aged three–five years, Lucas et al. (Reference Lucas, Hood and Coyle2020) found the approach extremely useful for acquiring both receptive and productive vocabulary. The project children were consistently successful at contextualized word production and were very motivated to learn how to write the words learned.

Looking into the potential of task-related physical activity and gestures for promoting vocabulary learning in preschool children, Toumpaniari et al. (Reference Toumpaniari, Loyens, Myrto-Foteini Mavilidi and Paas2015) carried out an experimental study with four-year-old Greek learners of EFL. Three learning conditions were examined: learning vocabulary without physical activity or gestures, learning with gestures only, and learning with both gestures and physical activity. The results of the study showed that a combination of task-related physical activity and gestures during vocabulary learning was related to the highest learning outcomes.

An important emerging topic awaiting systematic research concerns measuring the process and results of preschoolers’ FL vocabulary learning over time. In most of the above studies, the amount of vocabulary the children were able to comprehend or produce was very small. The reason that so many studies in the dataset want to quantify outcomes indicates that researchers think in terms of models characterizing older learners: they want to measure how many words children can comprehend and use actively, but the outcomes are very modest, indicating the slow progress children make (Sun et al., Reference Sun, Steinkrauss, Wieling and de Bot2018; Yeung et al, Reference Yeung, Ng and King2016).

4.5. Research on FL teachers and parents

A range of studies have examined perceptions and beliefs of pre-service trainee teachers of FLs as well as practicing preschool FL teachers about teaching very young learners (YLs) and how children define the good language teacher. Who would be best equipped to teach FLs to very YLs, FL specialist teachers or pre-primary practitioners, native speaker teachers or non-native speaker teachers, are also recurring topics. The section ends by considering FL teachers’ interactions with learners and parents as well as key challenges in teacher education.

4.6 Inside the preschool FL classroom

Research studies on how teaching processes and uses of languages impact children's engagement and L2 learning are highly valuable contributions to the field. As they zoom in on what happens in various activities in specific contexts, they offer insights into participants’ emic perspectives and the tiny steps in children's development. Findings show how children spontaneously draw on all their languages, including body language, in their repertoire.

5.1 How preschool children, their teachers, and parents are researched: Approaches to research design

Several authors discussed how very young learners of FLs, their parents, and teachers were studied. Mourão (Reference Mourão, Garton and Copland2019) stressed that preschool FL learning and teaching is a largely under-researched area to the extent that systematic research of all its aspects is needed. She calls for compiling robust enough data so that FL learning policies can be based on valid and reliable research evidence. This kind of evidence, in our view, is missing at this point for three reasons: (1) lack of model and clear curricular aims, (2) research design issues, and (3) the research questions researchers aim to answer. These points are interrelated, but we discuss them one by one.

There is no widely accepted model of pre-primary FL learning integrating how children's IDs and their L1 and L2 development interact with aspects such as their teachers’ competences in L1 and L2, togerther with age-appropriate classroom methodology, classroom-related variables, and extra-curricular learning opportunities (Nikolov & Lugossy, Reference Nikolov, Lugossy and Schwartz2022, pp. 27–28). FL programs which fail to specifiy aims (discussed in 5.2) pose an additional challenge. A model would allow researchers to test how variables work in specific contexts by using a quantitative approach: (quasi)-experimental or correlational or survey studies (Creswell, Reference Creswell2012, p. 12). Such a model could be proposed based on exploratory studies working towards a unified theory.

The 74 publications included 24 quantitative, 28 qualitative, and 22 mixed methods inquiries, showing that all three approaches to research design are present. The studies using a quantitative approach tested relationships between variables on constructs assumed to be known. A complex example is Unsworth et al. (Reference Unsworth, Persson, Prins and de Bot2014): this longitudinal experimental study involving 168 Dutch children learning English in 17 groups measured the children's aptitude and L2 progress, their teachers’ English proficiency, and the amount of input received. They found that children already had measurable receptive vocabulary at initial data-gathering (measured with a validated test for L1 English learners), the control group also improved, and increase in vocabulary in all groups was minimal. No information was available on what happened in the 17 classes during the two years. Thus, applying a reasonable model and instruments validated for older FL learners or native speakers offers less valid and reliable results with very young children without analyzing classroom data through a small lens.

As for the research design used with different stakeholders, teachers and parents were either surveyed or researchers elicited qualitative data on teachers’ views, beliefs, and practices using a range of open-ended techniques: e.g., reflections, focus groups, interviews, and video-taped observations. Such data are often collected from adults with validated instruments, and they tend to allow researchers to return with valid and reliable, or credible and trustworthy findings. The most challenging issue is how the interaction between teachers and children is studied in age-appropriate ways, for example, in teacher-led activities and in free play as children interact (Mourão, Reference Mourão and Schwartz2018; Robinson et al., Reference Robinson, Mourão and Kang2015). As pre-primary FL learning is still unchartered territory in terms of research, exploratory qualitative inquiries are conducted. These are valuable and necessary for building models to allow researchers to work towards the larger picture of what variables contribute to success in pre-primary FL programs. How success is defined, however, is a recurring issue. We argue that curricular aims should include attitudinal, emotional, cognitive, social, and L2 targets (Johnstone, Reference Johnstone, Garton and Copland2019) in line with what children in this age group are expected to be able to do in their L1(s) (see 5.2). This latter aspect is typically only implied rather than explicitly stated in research.

The research questions posed have important implications for how narrowly focused studies are, what they add to current knowledge, how many participants are involved, how datasets can be triangulated, and whether the findings can or cannot be generalized to other populations or replicated. As Nikolov and Lugossy (Reference Nikolov, Lugossy and Schwartz2022) pointed out, research on preschool FL programs tend to focus on either one or a few variables (e.g., vocabulary, explicit learning, repetition, motivation, learning area, teachers’ competences in L2 and teaching methodology). Few studies are longitudinal over years, and they fail to add to the formulation of realistic aims, optimal teaching conditions, patterns in implicit learning, or address how children with learning difficulties cope with challenges in FL learning. An additional problem, as Alstad and Mourão (Reference Alstad and Mourão2021) pointed out, is a lack of interdisciplinary research.

Although varied research methods were used in the studies in our database, as Nikolov and Lugossy (Reference Nikolov, Lugossy and Schwartz2022) observed, most are small-scale and some of the instruments applied are poorly designed. For example, Yeung et al. (Reference Yeung, Siegel and Chan2013) administered an explicit language-enriched phonological awareness program over 12 weeks, measured the outcomes on 12 tests unrelated to the treatment, and found that the results were not valid. Coyle and Gómez Gracia (Reference Coyle and Gómez Gracia2014) taught five nouns in 3 × 30 minutes in a song; as for how many words children could recall, the mean score was 1 after treatment. Jin et al. (Reference Jin, Zhou, Hu, Yang, Sun and Zhao2016) elicited metaphors from Chinese children on what learning English was like. They did not realize that metaphors were beyond the children's cognitive abilities and the visuals they used misled them. Furthermore, many studies involved privileged preschool learners of English of educated elite families (e.g., Keydeniers et al., Reference Keydeniers, Aalberse, Andringa and Kuiken2021), whose parents could afford to pay for tuition in private programs. No study involved children with learning difficulties or behavioral problems or low SES.

Teachers’ roles are of paramount importance, but studies have mostly focused on formal aspects of their education, general attitudes to very early FL learning, and collaboration with homeroom and other FL teachers and parents. There has been hardly any in-depth research on teacher strategies, cognition and beliefs, their attitudes to and motivation for very early FL teaching, self-perception of efficacy, and professional and personal development with increasing teaching experience. Teacher studies have mostly relied on questionnaires, interviews, observations, and teacher logs. Qualitative research and longitudinal studies could offer a deeper understanding of FL teachers’ roles in preschool FL learning (Nikolov & Lugossy, Reference Nikolov, Lugossy and Schwartz2022; Thieme et al., Reference Thieme, Hanekamp, Andringa, Verhagen and Kuiken2021).

Studies on parents focused on their beliefs about FL learning, expectations, and aspirations concerning their child's L2 development, involvement in the child's learning at home, and the impact of family SES on these factors. Some attention has also been paid to parental pressure on preschool institutions to introduce FL learning and to insist on an English-only approach. Questionnaires and interviews have been the main instruments used in these studies. For a deeper understanding of parents’ roles, we need more in-depth longitudinal research on parental influence on child FL learners’ attitudes, involvement in their L2 learning, collaboration with teachers as well as other parents, on their awareness about FL teaching and learning processes, and ways they can help their child with emerging FL learning problems (e.g., anxiety, feeling of failure) (Nikolov & Lugossy, Reference Nikolov, Lugossy and Schwartz2022).

5.2 Assessment of preschool FL learners: Is it all about learning words?

None of the language policy documents and curricula specified how pre-primary children's attitudes and openness to languages are expected to develop, what very young learners are expected to be able to do in their new language at different stages in pre-primary FL programs, how those are aligned with their cognitive and L1 development, and how teachers should diagnose to what extent children have achieved those targets. Therefore, it is unclear what the constructs are and how they can be operationalized (Nikolov & Timpe-Laughlin, Reference Nikolov and Timpe-Laughlin2021). However, many projects assessed how children progressed in their FL learning and what they could do in a FL over a relatively short time (typically weeks). These are problematic points in the studies on pre-primary FL learners for multiple reasons. As their development is very slow (see section 2), realistically, only minimal development is to be expected, and this is what studies on vocabulary have found. At what points and to what extent it is possible and ethical to measure this slow development are worth considering. What happens with the results, who uses them and for what purposes, apart from publishing them, are also in need of discussion. We found hardly any data on how assessment may impact the children themselves (for a positive example, see Prošić-Santovac & Navratil, Reference Prošić-Santovac, Navratil, Prošić-Santovac and Rixon2019, for a negative example where testing caused sickness see Jin et al., Reference Jin, Zhou, Hu, Yang, Sun and Zhao2016). Concerns remain over what can realistically be concluded or generalized from such assessments using validated tests for L1 learners and researcher-designed tests. Nikolov and Timpe-Laughlin (Reference Nikolov and Timpe-Laughlin2021) argued for redefining the aims of early FL assessment and replacing language tests with longitudinal observation that goes beyond linguistic results and includes the child's general cognitive, social, intercultural, and other development, together with their awareness of languages and other important values in life.

In our dataset, researchers collected assessment data on pre-primary FL learners’ progress and achievement along the lines they defined. Out of the 74 empirical studies, 32 used one or more tests, including 28 publications reporting L2 vocabulary test results. The number of items in receptive vocabulary tests, the most widely used test type, ranged between 5 and 35. Two studies included tests of L1 (Polish and Chinese); all other projects assessed children only in their L2 and some in other domains also (e.g., aptitude). This sharp focus on vocabulary (Alexiou et al., Reference Alexiou, Roghani, Milton, Prošić-Santovac and Rixon2019) would be in line with theories on how children learn languages (section 2) if all researchers took into consideration what task types were familiar enough to make sure that the tests were valid and reliable (Butler, Reference Butler2019) and they measured what this age group is good at (i.e, implicit rather than explicit learning). For example, Davis and Fan (Reference Davis and Fan2016) taught 64 Chinese children 15 short sentences in songs or in choral repetition in 15 lessons of 40 minutes over seven weeks and tested their productive vocabulary by asking them to say what they could see in 15 visual prompts. They reported results in mean length of utterance, as they expected multiword answers. Results were similar in both conditions: most children said either nothing or a single word. Neither the teaching approach nor the assessment were appropriate, although implicit chunk learning would have been a legitimate focus, missing from the dataset. In another quasi-experimental study, Albaladejo Albaladejo et al. (Reference Albaladejo Albaladejo, Coyle and de Larios2018) tested how 17 Spanish learners of English learnt five nouns in three types of activities: by listening to stories or songs or both. Test results showed that the story was most conducive to learning; however, three of the five words in the story were cognates and none in the other conditions.

An additional point concerns the content and cognitive validity of vocabulary tests used with very young learners. As was mentioned earlier, Unsworth et al. (Reference Unsworth, Persson, Prins and de Bot2014 and other studies) used a vocabulary test validated for English native speakers to compare how much progress Dutch children made over two years. Winke et al. (Reference Winke, Jieun, Ahn, Choi, Cui and Yoon2018) analyzed the cognitive validity of English tests for young learners by comparing results of L1 and L2 English test takers; they raised the question not only of how children using the L2 as their L1 would perform on these tests, but also how the children taking the L2 tests would manage to do such tests in their own L1, and if they ever do so.

The recurring theme of comparing implicit vs. explicit learning, and not conceptualizing them along a continuum, prevents researchers from understanding how children develop in these programs. No study looked at both dimensions of receptive and productive vocabulary and formulaic language or chunks as continua. The very short time devoted to (quasi)experimental projects indicates that researchers tend to bear in mind either children with high aptitude who can learn fast and their results can be measured in short treatment studies, or they think in terms of what assumptions are reasonable with older learners.

Furthermore, as discussed in section 4, empirical studies on children's IDs, language awareness, and learning of various aspects of communicative competence, used certain tests to tap into either explicit or implicit learning or both, although they were not all framed as assessment. The issues related to these publications concern why and how children are assessed, how familiar they are with the tasks used for testing, how valid and reliable the tests are, how they are aligned with curricula and how children learn new languages, and in what ways tests are conducive to learning (Nikolov & Timpe-Laughlin, Reference Nikolov and Timpe-Laughlin2021) and contribute to children's wellbeing in more general terms.

Finally, it would be helpful to propose a systematic and age-appropriate framework for developing and diagnosing pre-primary children's milestones along which they can be expected to develop over the years. Such a document should not only be aligned with existing frameworks for older FL learners (e.g., Alexiou & Stathopoulou, Reference Alexiou, Stathopoulou and Alexiou2021), but also should integrate age-appropriate task types (Nikolov, Reference Nikolov and Nikolov2016) tailored to the children's physical, cognitive, emotional, and social characteristics in collaboration with ECEC teachers and experts. This idea, however, does not mean that we propose assessing children in pre-primary FL programs. On the contrary, we suggest leaving them alone to enjoy activities, play as they usually do, and develop at their own pace. Therefore, FL researchers should align their practices with those of ECEC experts. Observation is a much better but a lot more time-consuming way to collect data than using tests. It would be a lot more reasonable to find out in later years what children think they have learnt in their preschool years, how they reflect upon their very first experiences by using retrospection, and how what they could do in the FL was built upon later. Findings on primary FL programs indicate that later starters often outperform earlier starters. Therefore, we assume that the benefits of early exposure in pre-primary are not in the L2 domain and results of assessments should be taken with a pinch of salt.

6.1 Summary of key findings

Although beliefs in very early FL learning are strong, and pre-primary courses have spread significantly, our review of published studies suggests that - in terms of research - this field is still in its infancy. This statement is true despite the exponential growth in the number of publications (see Figures 1 and 2 in section 1.2) unevenly distributed in 25 countries. Where research was done, however, does not reflect in which countries preschool FL programs are widely spread and if they are available to children only in a few preschools. Most probably, individual researchers’ interest in very young learners and in their teachers motivate publications. This could be one of the reasons why most authors are enthusiastic about pre-primary FL programs and frame their findings in positive terms.

We found a mismatch between what theories, policy documents, and handbooks for teachers promote (sections 2 and 3) and what many studies examined. Pre-primary FL programs should allow children to develop positive attitudes towards the L2 (Andúgar & Cortina-Pérez, Reference Andúgar, Cortina-Pérez and Schwartz2018) and awareness of languages (Johnstone, Reference Johnstone, Garton and Copland2019), and to acquire the new language spontaneously rather than learn it consciously (European Commission, 2011, p. 17). Programs “should be integrated into contexts in which the language is meaningful and useful, such as in everyday or playful situations, since play is the child's natural medium of learning in pre-primary’ (p. 14). As studies in sections 3–4 show, evidence is often missing on how well programs are integrated into ECEC curricula. Many authors focus on explicit teaching and learning and expect short-term measurable outcomes instead of analyzing implicit contextualized acquisition and changes in children's attitudes, awareness, and spontaneous behavior and uses of multiple languages. In other words, many researchers lost sight of the whole child embedded in their preschool lives and their needs and abilities to elicit data for a publishable paper rather than move the field forward in meaningful ways. Yet another mismatch between language policy and reality emerging from the studies is that English is the most widely researched target language, and it dominates the field. Studies on other languages are very much needed.

As a result, the overall picture emerging from this review is fragmented; many researchers formulate their research questions based on how older students learn languages or what they see as feasible or measurable rather than using age-appropriate approaches to how children interacting with their teachers and peers acquire new skills and socialize into learning and using their languages (Alstad & Tkachenko, Reference Alstad, Tkachenko and Schwartz2018). On the bright side, many good practices are also explored (section 4.6): they offer insights into FL activities conducive to children's development (e.g., songs and rhymes accompanied by physical movement, role plays, listening to stories, watching cartoons). Such small-scale small-lens studies document the slow steps children take with their teachers’ support on their FL learning journey. These steps are tiny indeed: findings on primary-school FL learners tend to show that the younger the children are, the slower their FL development is in these empirical studies (Boyson et al., Reference Boyson, Semmer, Thompson and Rosenbuch2013).

Other challenges characterizing pre-primary programs also overlap with those typicaly found in published research on primary FL teaching and learning: lack of continuity in transition from pre-primary to primary (section 3.3), how differences in children's aptitude, motivation, anxiety, strategies, and SES impact their trajectories as their awareness and L2 competences develop (sections 4.1–4.4), what competences ideal teachers should have versus what they have, and a shortage of appropriately trained teachers (section 4.5.1), parental impact (section 4.5.2, mediated through children's SES) on language choice (English vs. other FLs) and public versus private, intensive versus minimal input programs.

Finally, let us summarize what evidence the studies offered to underpin arguments for the existence of pre-primary FL programs, what the key conditions include, what the main outcomes and challenges are, and where further research is needed.

Studies document that key conditions include:

1. 1. All stakeholders’ positive attitudes towards pre-primary FL programs in general, and the target language in particular.

2. 2. Age-appropriate curricula and materials including indicators of children's attitudes, language awareness, and target language objectives in harmony with other domains in the pre-primary curriculum (i.e., L1, cognitive, emotional, social, physical development and wellbeing).

3. 3. Teachers who are well-versed in ECEC and FL teaching methodology as well as their ability in the children's L1 and the target language.

4. 4. Classrooms which are familiar to children, equipped with attractive toys, games, picturebooks, and so forth, and suitably timetabled FL activities during the day.

5. 5. Parents who support their child's FL learning, cooperate with the FL teacher, and are involved in their child's FL learning at home.

6. 6. Primary programs should build on and integrate what children know and can do, so that they feel they are appreciated as individuals and their efforts and achievements can be valued.

6.2. Where future research is necessary

Our understanding is that the field is still at a stage where more exploratory studies are necessary to offer insights into what it is like to work with pre-primary FL learners at public and private preschools, and how FL and ECEC teachers’ expertise can contribute to best practices. Policy makers and curriculum designers should ensure that the key conditions are in place and learning a new language is an integral part of enriching many children's preschool life, not only those of the elite, and that learning a new language contributes to their general wellbeing (Caporal-Ebersold & Young, Reference Caporal-Ebersold and Young2016). How these ideas are put into practice should be the focus of future studies.

Many researchers whose work we overviewed are enthusiastic and positively-biased, but critical reflections on findings are rare, so we need more accounts of what does not work as planned, and why not. Similar trends have also been typical in publications on primary-school FL programs. Although challenging overly positive accounts took a long time to emerge, in our view, the wrong yes/no questions were asked: Is an early start worth it? Are L2 outcomes in the long run better than in later start FL programs? The studies we overviewed offer no answer, but it would probably be “no”. Researchers need to formulate questions in line with the aims of pre-primary FL learning and preschool education: How can we characterize children's awareness of languages and cultures, their attitudes towards learning new languages, engagement in FL activities, their self-concept, self-confidence, mindset and openness, and their abilities in L1, L2, and L3, and so forth, not only in ECEC but also at later stages in life? Such questions would allow researchers to widen the scope of inquiries, move away from narrowly focusing on FL and assessment to studying children and their caretakers in their daily lives.

More research is needed into what goes on in preschool FL lessons not only during teacher-led activities, but also in child-initiated role play, games, and free play (Mourão, Reference Mourão and Schwartz2018; Robinson et al., Reference Robinson, Mourão and Kang2015; Waddington et al., Reference Waddington, Coto Bernal and Siqués Jofré2018). Future studies should further explore how playful meaning-focused activities such as storytelling, picturebook sharing, uses of ICT and websites (Alexiou & Vitoulis, Reference Alexiou, Vitoulis, Enever, Lindgren and Ivanov2014; Giannakopoulou, Reference Giannakopoulou and Alexiou2021; Korosidou & Griva, Reference Korosidou, Griva and Alexiou2021), repetition (Roh & Lee, Reference Roh and Lee2018), combining gestures and other physical activities, and using objects such as toys (Güngör, Reference Güngör2018; Padial-Ruz et al. Reference Padial-Ruz, García-Molina and Puga-González2019; Kiaer et al., Reference Kiaer, Morgan-Brown and Choi2021) increase and maintain children's interest and engagement.

Pre-primary FL learning and teaching are clearly transdisciplinary (e.g., Alstad & Mourão, Reference Alstad and Mourão2021; Thieme et al., Reference Thieme, Hanekamp, Andringa, Verhagen and Kuiken2021), integrating expertise in ECEC and language pedagogy (Cortina-Pérez & Andúgar, Reference Cortina-Pérez, Andùgar and Alexiou2021). However, most publications use a narrow linguistic lens rather than focusing on the whole learner and teacher with all their competences and characteristics. The studies in our database were framed and conducted by experts on FL learning, whose ownership of the field is overwhelming, but who have little or no formal expertise in ECEC. Some projects were conducted by researchers who taught some lessons and collected data in preschools without professional investments in the field.

Recent studies involving children as co-researchers (Pinter Reference Pinter, Garton and Copland2019) and growing awareness of ethical issues involved in researching children (Pinter, Reference Pinter, Garton and Copland2019; Pinter & Kuchah, Reference Pinter and Kuchah2021), coupled with a focus on children's wellbeing as a prominent variable in understanding their FL learning and interpreting the findings (Thieme et al., Reference Thieme, Hanekamp, Andringa, Verhagen and Kuiken2021), signify novel approaches and new focal points to research in the field. However, as has been documented, teachers are not equal partners in most studies, and none was conducted by practicing pre-primary FL or ECEC teachers. We need studies conducted by practicing teachers. This is the most obvious area missing from the studies in the dataset. Researchers either come in, experiment with their ideas, or observe practices, and then analyze their data or they inquire into teachers’ practices, beliefs, and experiences. However, we need more studies initiated by teachers where their voices can be heard, and their concerns can be understood.

More systematic research is needed on teacher education (Kubanek, Reference Kubanek, Wilden and Porsch2018). Collaboration among ECEC and FL teachers should be a priority: how they benefit from one another's expertise, what they agree and disagree on, should be insightful. In our view, framing studies from the FL perspective limits what studies reveal about what it means to be a young child participating in FL activities, and what it is like to try to teach them something in a new language. Teachers and researchers should focus on enabling child-initiated play in FL contexts. They should not capture FL learning only and lose sight of the whole child, their general wellbeing, and their need to play freely - a key characteristic of children's daily lives. Studies should integrate child-initiated play as an important aspect of their learning and reality, as one of our anonymous reviewers emphasized. We have observed children in many contexts whose parents wanted them to learn English; however, children refused to participate in FL activities instead of enjoying free play (e.g., Lugossy, Reference Lugossy and Schwartz2018). Preschool children are not a monolithic group (Lockiewicz et al., Reference Lockiewicz, Sarzala-Przybylska and Lipowska2018). It would be useful to consider their age-related characteristics that may determine teaching approaches most suitable for a particular age subgroup. Among those needs is that of playing to learn and learning by playing.

Meaning making in playful activities should be at the forefront of studies. We need to know more about the ways in which teachers ensure that children comprehend input by relying on their background knowledge, L1 abilities, strategies, visuals, and the children's guessing meaning in contexts (Lugossy, Reference Lugossy2012). Translation, used in many classrooms (e.g., Dikilitaş & Mumford, Reference Dikilitaş and Mumford2020; Prošić-Santovac & Radović, Reference Prošić-Santovac and Radović2018a; Song & Lee, Reference Song and Lee2018), can take out the chill out of meaning making; how other strategies work over extended periods of time should also be researched. Teachers’ code-switching can help learners’ comprehension (Prošić-Santovac & Radović, Reference Prošić-Santovac, Radović and Schwartz2018b; Scheffler & Domińska, Reference Scheffler and Domińska2018); it would be helpful to know more about how. Studies should build on and integrate research findings from the L1 literature on conceptualizing and analyzing classroom discourse. For example, it would be useful to adopt “the trifocal lens of interaction, language, and concepts” (Rowe & Snow, Reference Rowe and Snow2020, p. 16) for analyzing how teachers and children use the languages at their disposal to promote learning while bearing in mind children's age, cognitive, and language abilities.

Longitudinal studies of good practice in multiple FL languages need to be conducted, and not only in English, integrating the teachers’, children's, and parents’ perspectives over years. Some of the benefits and possible drawbacks may emerge only years after children transitioned from pre-primary FL programs. The key question characterizing studies on primary FL programs will most probably emerge after a lot of investment into preschool FL programs. It will take time to find out what the added value is over time and how children benefit from being part of such programs for years. Evaluation studies of full programs, like Dobson et al. (Reference Dobson, Pérez Murillo and Johnstone2010), for example, could offer such insights.

Questions arising