1. Context

1This research brings together scholars in two Belgian universities (Leuven and Liège), for whom Sabine De Knop has been a source of intellectual escapades, by asking intriguing questions: (1) how do we need to to consider and use in teaching the similarities and differences between L1 and L2 learners and their constructions; (2) how can constructional research result in efficient teaching practices and resources?

2In this contribution, the extended version of a paper presented at CALP3 (kindly organized by Hans Boas at the University of Texas at Austin in 2018, in the Constructionist Approaches to Language Pedagogy conference series, first organized by Sabine De Knop in Brussels in 2013), we rethink traditional teaching practices by aligning the abstract theoretical apparatus of cognitive linguistics, more specifically Ron Langacker’s cognitive grammar (CG), to the naturalistic cross-linguistic output by L1 and L2 learners of Dutch and French, from the perspective of usage-based approaches. Construction-based teaching offers an integrated approach to (communicative) language teaching, since it is not limited to constructions at sentence-level. Rather, it allows for the integration of structures at the discourse and interactional levels, while including a grammar-lexicon continuum. In previous work conducted with Sabine De Knop and Fabio Mollica (De Knop et al. 2015), lexical approaches to language teaching were shown to benefit from taking grammatical patterns into account. In this contribution, we want to explore how the traditional method for teaching verbal tenses can be complemented by authentic complex clausal patterns and lexical information for imagining the future in communicative actions. These insights will be of interest not only to students of (applied) linguistics, but also to foreign language learners in general.

3Imagining the future is a fundamental cognitive capacity for shifting consciousness: through imagination we move beyond present experience and current thought into anticipated time and space (Chafe 1994). The semantics of future time is a largely underexplored grammatical topic (Sambre 2012, 2017), especially in contrastive teaching of Romance and Germanic languages such as French and Dutch. For these languages, prevailing communicative or task-based teaching methods and grammars do not advocate explicit language comparison, but focus on discrete rules (van Aert 2019) for future verb forms and lexical items as well as limited lists of (not always authentic) examples of language use. Seen from another perspective, the focus on implicit learning in communicative and inductive teaching methods seems to hinder the integration of recent advances in linguistic theory and cognitive science (Tyler 2012). In our teaching practice, we stimulate theoretical reflections on the meaning of grammar (Achard 2008) in a bilingual Belgian French-Dutch academic setting (a contrast pair rarely studied from a constructional point of view; for an overview of recent work in applied cognitive linguistics, cf. Hijazo-Gascón & Llopis-García (2019)).

2. Objectives

4Our objectives are threefold. At the descriptive level, we use an epistemic model to investigate the range of authentic constructions for future time that are not restricted to one or more future (inflected or periphrastic) tenses. We contrast L1 and L2 speech patterns of futurity in a corpus of (the same) L1 and L2 languages. Our second aim is didactic: we want to teach L2 future constructions based on a rich array of authentic L1 production and therefore formulate guidelines that integrate temporal and modal aspects of futurity. These numbered recommendations appear in square brackets [R] in the analytical sections of the text and are explained in the conclusions (see Section 6). The third objective is implicit and has a societal relevance: we want to embed grammatical futurity in a realistic and student-centered learning case that is inspired by the educational policy of KU Leuven regarding its future disciplinary self. This self-representation is intended to provide students with an explicit orientation in their academic subdiscipline and as future professionals, and to broaden and deepen their self-reflection throughout the curriculum about their professional working field, other students (at their own and partner universities such as the University of Liège) and their scientific peers. The ‘future self’ is a conceptual construct in which current ideas about the student’s life are projected into the future and should be made explicit throughout the curriculum (Saunders-Smits et al. 2021). In operationalizing this concept of the disciplinary future self, we combine meaningful and relevant teaching about future time in different stages of the BA-MA curricula. In our view, conceptual grammar needs contextualization and is concerned with students gaining relevant knowledge (e.g. in tandem learning) about the relevance of their curriculum and about their professional and personal futures (Priniski et al. 2018). Imagining the future self requires a language (in to be shared in multilingual academic or work settings) in which students imagine themselves in real and imagined communities (Park & Schallert 2020).

3. Framework

5We embed constructional form-meaning variation in Langacker’s epistemic model, which represents reality as a growing cylinder, called the present reality, in which growth towards the future occurs, starting from the present reality, and is free to take any position on a temporal-modal scale, ranging from projected reality (Langacker 2008: 306), over modal variations on a diagonal axis, from which the world is not conceived as reality, but as potentiality, extending from present to future. Thinkable irreality, then, is the zone of the unreal or impossible: that which we do not (can or may not) know or which is unlikely to happen in the future (Langacker 1991: 242–246, 2009: 201–212; Sambre 2009).

Figure 1. Epistemic time model for futurity following Langacker (2008:306) with additions for temporal distance and epistemic modality

6Future structural assemblies combine lexical subparts and grammatical forms such as verb tenses and their morphological subparts to posit usage events in this conceptual template. Assemblies occur to varying degrees of entrenchment and can be idiosyncratic. In our discursive conception of future time, all levels and assemblies of language production, need to be included in the analysis, without a priori excluding forms based on so-called grammatical deficits or language errors.

4. Corpus

7In 20 video-recorded interviews, 10 second year bachelor students (hence BA2) of (applied) linguistics and literature (5 Dutch-speaking [NL1] in Leuven and 5 French-speaking [FR1] in Liège) with a Dutch-French language combination were asked 20 questions about their future in the two languages, first in their L2 and after a short break in their L1. These (open) questions integrate specific tenses (present, future, conditional) and nominal, adjectival or adverbial references to the future as well as temporal distance (near, neutral or remote) and modality (realis, potentialis or irrealis). These questions and their conceptual and linguistic features (YES or NO) are listed in Table 2 below (our translation).

8The transcripts of answers were filtered for future time reference, then decomposed first into intonation units (hereafter IUs) (Chafe 1987), and further into subclauses (viz. matrix verbs and complement clauses/subclauses) on the one hand and satellite clauses independent of the main verb on the other. The transcripts of the answers provided by the students compose our corpus; for this research, we did not explicitly take into account the extent to which students reactions mirrored the morpho-syntactic and lexical forms used in the questions.

9Each subclause was double-checked against a coding sheet with parameters and abbreviations. We obtained 2,252 IUs with balanced NL-FR and L1-L2 subsamples, of which 85% (N=1,913) contained a reference to the future. Our analyses are based on these 1,913 IUs.

5. Typology and discussion

10In what follows, we try to gain a better understanding of the constructions respondents used to talk about their future. We summarize our main findings, starting from simple constructions and gradually moving towards more elaborated construction templates that speakers use in their first (native) language and second language. We do this first for the time reference (5.1), then for the modal (5.2) variants.

5.1. Future time and tenses

5.1.1. Future time and future tenses

11When a single verb is used, the future tense is not the predominant verb form [R-1 (a)]. More often, the future is expressed by a (simple) present tense (N=1,271 [FR = 668; NL = 603], five times more often than the future tense (N=232 [FR = 115; NL = 117]). For the sake of clarity, we add that conditionals (7%) and infinitives (21%) are used less frequently.

12Nevertheless, infinitives appear about twice as often as the future tense. Infinitives in answers are non-finite and therefore semantically ungrounded verb forms in which the answer blurs references to time and person (example (1a-b)). There are two possible explanations for such forms: the simplification may result from a morphological deficit, but it is more likely that the temporal framework implied by the question does not require explicit repetition of the person and temporal deixis. Teaching tenses could operationalize ungrounded processes [R-2] at a communicative L2 level by using an infinitival core (as in (1)) where, instead of focusing on (standard automation exercises for) repetition of the time expressed in the question, answering fluency is improved by a (basic) infinitive as a (future) discourse pattern per se (Cappelle 2021), eventually enriching the infinitive with additional verb arguments.

13A second dimension relates to the specific use of future tenses (Table 3), and the choice between simple (absolute) future tense (Dutch periphrastic onvoltooid toekomende tijd with auxiliary zullen + infinitive [ik zal werken], French inflected indicatif futur simple [je travaillerai]), the periphrastic form with auxiliary go (FR aller, NL gaan) + infinitive [FR je vais travailler, NL ik ga werken] and future perfect (Dutch voltooid toekomende tijd and French indicatif futur antérieur, the latter indicating anteriority with respect to the main verb in the future tense or a present tense with future reference). Generally, there seems to be a balance between the simple future tense and the periphrastic go + infinitive in both French and Dutch: in Dutch, the preference for the simple future tense (55.7%) is slightly more pronounced than for the go-form (37.4%). However, this first reading hides an overuse of specific tenses among L2 speakers [R-3]. Traditionally, the Dutch variant with go is thought to establish a conceptual link between the present action context of the utterance time and the near future action or even inchoative movement leading to the future state expressed by the verb (Beheydt 2011). French L2 speakers strongly use the periphrastic go-construction, while their Dutch L2 counterparts prefer the Dutch zullen. Accordingly, a didactic differentiation is required here: the andative meaning of the Dutch go-variant (Chafe 2018) should be made explicit in Dutch L2 lessons, while the future tense variants zullen and futur simple should be reserved for non-action-related verbs or for (more isolated) remote events in both Dutch L2 and French L2.

5.1.2. Tense combinations

14Apart from the isolated tenses traditionally emphasized in grammar teaching, we observed frequent combinations of tenses in complex patterns for future events, at the intersection of present (P), future (F), conditional (C) and infinitive (I). In table 4, we represent these combinations as sequences (with 0 for empty elements), such as P00I (present tense and infinitive (2)) or PF00 (present tense and future tense (3)), as in the following examples:

15Apart from PFCI, all combinations of two or three elements occur, with different relative frequencies. The already mentioned basic types P000 0F00 00C0 000I are relatively important (N=1,133 or 59.2%). It is striking that other patterns with two time elements appear, such as P00I, P0C0 and PF00 (N=396 or 20.7%), not to mention the complex PF0I or P0CI (N=19 or 1%). Tense combinations, viz. finite (present, future, conditional) or non-finite (infinitive) forms, represent an interesting category for the expression of future time which is ignored by (construction) grammars and their traditional focus on isolated tenses in simple clauses. Combined tenses link a present expectation (about what is to come) or a present thought with a projected possibility (being able to teach (1) as a continuous summary infinitive) or a future change (become a translator (2)), expressed by the sequential change implied by become. Tense combinations reveal complex conceptual operations in thinking and speaking about an imagined future that are not limited to a specific moment, but functionally link the conceptualizer’s present to his or her imagined future. Constructional futurity in this respect is not limited to a single morphological class (one tense) but becomes part of a larger clausal and elaborated sentence syntax for a split self, namely a complementation structure that combines two tenses, a present and an explicitly marked (future tense) or implied future (infinitive) in the direct object of this first tense [R-1 (b)], sometimes even subclauses use a past tense to mark anteriority with respect to a future main clause (a common trend in spoken Dutch and French, instead of the future perfect, i.e. French futur antérieur or Dutch voltooid toekomstige tijd). This spoken register variant is rarely taught in L2 learner grammars.

5.1.3. Futurity beyond tense: FUT N, Adj, Adv

16A less frequent but interesting phenomenon is the reference to the future in nouns (e.g. the future), adjectives (e.g. my future ambition) and adverbs (next or satellites of time, e.g. in the following years). We examined how often a verb coincides (or not) with such non-verbal markers. Obviously, most constructions (81.93%) contain only one (future) verb. Non-verbal markers can occur as adjuncts, with a preference for (Dutch) nouns (2.71%) and adverbs (0.39%). In both L1 and L2, combinations of tenses and non-verbal tense markers open up an interesting perspective. Think of L1 constructions such as

17On the other side of the constructional array, in elliptic VPs (13.91%), L1 offers more compact answers, without nouns (8), adjectives or adverbs, which require additional syntactic (conversational) competence.

18Apart from teaching explicit grammatical rules and useful words in doing futurity in interaction, conventionally efficient (elliptic) forms for recurrent conventional socio-pragmatic patterns should consider structures beyond individual speakers’ sentences, and the simultaneous omission of future verb fillers (or even noun fillers, as in (8)) in particular interactive genres (Nikiforidou & Fischer 2015) [R-4].

5.2. Modality

19Besides verbal and non-verbal constructions of futurity, two dimensions of modality emerged from our corpus. The first dimension concerns epistemic modality, e.g. the more traditional way of grasping the degree of reality of the uttered sentence. The second one relates to dynamic modality (see section 5.2.2).

5.2.1. Epistemic modality

20In most pedagogical grammars, and in construction grammar, epistemic modality is limited to grammatical verb forms for expressing potentiality (French conditionnel présent, Dutch voorwaardelijke wijs) or irreality (consider the case of conditional subordinate clauses in French and Dutch). In contrastive teaching of Dutch and French, the opposition between potentialis (POT of the future (9)) and irrealis (IRR of the present and, more specifically, of the past (10)) is usually taught from that perspective. In these two cases, conditional subclauses need indicatif imparfait (imperfective past tense or IPFV) and indicative plus-que-parfait (Past Perfect or PPRF), respectively with main verbs in conditional present (COND.PRS) and past conditional (PCOND).

21French grammar teaching for native speakers of Dutch insists on the correct use and the fixed form of French tenses, where French would not allow the use of the conditional in the subordinate conditional clause (Meunier 2017: 300–302). Similarly, modern pedagogical grammars of Dutch insist more on the discrepancy between temporal use of past perfect and modal distance in embedded constituents (Vandeweghe 2004).

22In our corpus, we see how students perceive their real or potential future, apart from the conventional forms taught [R-5]; irrealis here is used to express present ignorance about the future (see examples (11) and (12) below). The incorrect forms were only produced by L2 speakers.

23Projected reality seems to occur most frequently here, and no real difference is measured in terms of L1-L2 variation for realis and potentialis. Students master the expression of the conditional in both languages. More attention needs to be paid to the expression of the present irrealis (8%), as students seem to have a rather limited stock of conventional expressions [R-6]. In FR-L2, surprisingly, we find a larger number of irrealis forms. It could be that the students give a negative answer, such as ik weet het niet (I don’t know), a less conventional expression for ignorance in Dutch (cf. (11)).

24Our analysis has shown that the expression of uncertainty and ignorance should be more explicitly integrated into the teaching of epistemic modality [R-7]. A possible didactic approach would involve idiosyncratic (colloquial) expressions, as mentioned above, with the forms realis (negated present indicative (11a)), potentialis (negated conditional (11b and 12a)) or with elliptic forms as in (12b) (except for the incorrect *-forms used by some respondents). A good functional example of the expression of ideas, existing or nonexistent beliefs and intentions can be found in some Italian grammars (Proudfoot & Cardo 2005: chs. 24 and 32). The implication of this approach is that (especially in Romance language teaching) we may need to move from approaches based on morphological tenses to a conceptual-functional gradient approach to futurity (Herbst 2016: 25–27), in which we distinguish more broadly between what we can say about the future (epistemic meaning) and the (from a Dutch perspective)  non-inflectional forms used to indicate how epistemic positions, in the form of verb forms that are not always clearly marked, are combined with certain other grammatical markers  –  such as negation or the negated indefinite pronouns geen (12c), aucun(e)  (13b), or other lexical units (for instance the verb know in (12b) or the noun idea in (13a)) [R-8].

5.2.2. Dynamic modality

25Based on the combinations and complements of verbs and tenses for the present and the future (especially P00I, PF00 and P0CI, cf. 5.1.2) as well as the one to many ways of uttering epistemic modality (5.2.1), we mapped the conceptual content of the verbs used for thinking and saying futurity. The mind map [R-9] lists the different verbs (want, think, hope), sometimes also nouns (objective, dream) or adjectives (interesting, risky) that emerged from the corpus.

26These expressions render a dynamic modality that refers not so much to the potentiality of the event signaled by the proposition or to the necessity of the fact expressed, but to the internal control of the sentence subject (Palmer 2012: 7), i.e. the speaking respondent.

Figure 2. Concept map of verbal subtypes for dynamic modality in the corpus

27These verbs belong to six subclasses: willingness, ability (where can does not refer to epistemic possibility), perception, as well as emotive, cognitive, declarative expressions. Below the absolute frequencies are listed for each category. These absolute frequencies appear to be quite similar in L1 and L2 [R-10 | R-11]

28As the concept map shows, the formal side of dynamic modality largely depends on lexical verb subtypes that use a main matrix verb in the present and a subordinate that/to-clause in the future in a construction, such as I think (cognitive) or I hope (emotive) + that V.IND.FUT or I want/IND.PRS or I would like COND.PRES (willingness) to INF.PRS (used for posteriority). In other words, respondents use an integrated form that recruits syntactic, lexical and morphological elements, possibly as a simplified constructional template to compensate for lexical and tense deficits –  a strategy that could explain the overuse of the present indicative (P000) for futurity and the relatively high frequency of combined tenses PF00 in both languages (cf. 5.1.2).

29Dynamic modality appears to be more frequent in the productions of the French L1 speakers than in those of the Dutch L1 speakers. Also, both groups tend to use more dynamic modality when speaking their L2 (this is more strikingly the case for the Dutch-speaking learners of French). Our provisional conclusion is that basic lexical items in constructions for dynamic modality are more frequently used in L2 as a compensation strategy for morphological deficits.

5.2.3. Modal combinations

30Table 10 below shows the two opposites of objective epistemic modality and the subjectively colored dynamic modality, which is often combined with epistemic modality (since any dynamic attestation simultaneously emphasizes the reality or potentiality of this subjective stance). Future time is sometimes perceived as a deontic (internal) necessity. Dynamic modality coincides with epistemic (40%), deontic (4%) or both (1%), for a total of 45%.

31The question then arises how to conceive of the combination of different types of modalities in that/to-clauses, epistemic, dynamic and/or deontic (deontic verbs such as must and have to are not considered in the current discussion for reasons of space) in concrete constructions [R-12]. We briefly summarize two interesting constructional templates [R-13] that bring together both conceptual information about the future and the present position from which that future is conceptually accessed and imagined. Note how the relationship between present and future tense can be distributed in different ways across main and subordinate clauses by exploiting the matrix verb’s valency (especially in case I),

while in case II a combined satellite is used in a present subordinate causal clause and a second optional target clause:

6. Conclusions and recommendations

32In this contribution, we have investigated two sides of naturalistic L1-L2 constructions for the expression of imagined futurity and future self, based on a cognitive-grammatical epistemic model  in a spoken corpus of NL-FR learners: the conceptual interplay between future time and different modalities, epistemic and dynamic. Based on these initial descriptive results, and in contrast to traditional grammar teaching about tenses, we propose a set of form-function recommendations [R], first for the temporal dimension (A):

33To this first dimension, we add suggestions of future modality on the epistemic dimension (B):

34Importantly, dynamic modality (dimension C) transforms the binary tense-epistemic model into a three-dimensional future space. This is, in our view, a theoretical contribution compatible with Langacker’s spatial time model:

35At an advanced learning level, we suggest to

36Our future research will integrate constructions of futurity into the disciplinary future self at more advanced levels of language proficiency and communicative tasks, including a multimodal co-speech dimension for time and modality, for instance eye gaze, hand movements, facial expression and body orientation. A current reform of the Leuven MA programs in multicultural communication foresees the development of multilingual teaching methods and cross-linguistic communication projects (on future-oriented societal issues such as climate change or diversity). We are confident that the conceptual apparatus offered by cognitive linguistics can provide an overarching predictive framework for building the disciplinary selves of our students and our own disciplinary selves in applied cognitive linguistics.

37On a more personal note: thinking about the future of research, we are inevitably dealing with the inspiration that Sabine De Knop always shares with us, both in an intellectual and personal way. Therefore