Formulaic Expressions for Foreign Language Learning and Teaching

: Foreign language teaching experts unanimously insist on the necessity of acquiring formulaic expressions in order to communicate successfully in the target language. However, many of the treatises in favour of phraseme use, including semantically non-compositional idiomatic expressions, by foreign language learners seem to be marked by an insuﬃcient depth of reﬂection as to applied linguistic, methodological, and phraseodidactic1 criteria. e present contribution therefore aims at a diﬀerentiated treatment of prefabricated communicative constructions, starting out from an extended deﬁnition and classiﬁcation and by discussing the pros and cons of phraseme acquisition. ese considerations will lead to the delimitation of formulaic language fundamental for an operative foreign language competence (routine formulae, collocations and “constructions”) as opposed to those types of phrasemes which are not essential or even inappropriate for non-native speakers.


Introduction
[…] communicative competence is not a matter of knowing rules for the composition of sentences and being able to employ such rules to assemble expressions from scratch as and when occasion requires.It is much more a matter of knowing a stock of partially pre-assembled patterns, formulaic frameworks, and a kit of rules, so to speak, and being able to apply the rules to make whatever adjustments are necessary according to contextual demands.(Widdowson 1989: 135) Researchers in different domains of linguistics and second language acquisition unanimously admit the systematic presence, to a greater or lesser extent, of formulaic or prefabricated language in spoken and written communication (cf.Schmale 2021a: 9).Like Widdowson, quoted above, foreign language teaching experts generally insist on the necessity of acquiring formulaic expressions in order to communicate adequately and successfully in the target language.As early as 1909, Charles Bally pointed out the utmost importance of learners' acquisition of phraseological competence for the mastery of a foreign language: 2 L'étude des séries, et en général de tous les groupements phraséologiques, est très importante pour l'intelligence d'une langue étrangère.Inversement, l'emploi de séries incorrectes est un indice auquel on reconnaît qu'un étranger est peu avancé dans le maniement de la langue […].(Bally 1909a: 73) 3 Among the numerous protagonists in favour of creating phraseological competence, subsequently to Bally's initial postulation, Mel'čuk (1993: 84) is one of the most outspoken and probably influential ones, his convictions being shared by researchers in "phraseodidactics" across Europe: "l'apprentissage systématique des phrasèmes est indispensable dans l'enseignement d'une langue" (Mel'čuk 1993: 84).However, approximating Mel'čuk's quoted statements which will be discussed later in more detail, a great number of treatises advocating phraseme acquisition and use by foreign language learners (= FLL) seem to suffer from an insufficient differentiation and depth of linguistic, methodological and phraseodidactic criteria appertaining to the description and learning-teaching of prefabricated, formulaic, phraseological language or, rather: communicative structures.In fact, nonverbal and situational factors necessarily have to be considered with respect to the multimodal nature of speech in general and formulaic expressions in particular.
e present contribution therefore sets out to propose solutions to some of the major shortcomings of current phraseological and phraseodidactic research in order to deliver a description of discrete formulaic structures essential for the development of an operative intercultural communicative foreign language competence integrating prefabricated language as a pivotal component.
To start with, a brief overview of existing phraseodidactic studies will be delivered on the phraseological competence advocated by research in this field (section 1).is will be followed by a presentation and discussion of defining and classifying criteria for delimitating prefabricated communicative structures (section 2).Based on these necessary fundamental formulaic provisions, essential prefabricated communicative structures for the development of an operative communicative competence will then be delineated (section 3).In lieu of a summary, a reflection on methodological principles for the selection and teaching-learning of formulaic communicative structures is finally presented in section 4.
1. Phraseological competence as advocated by phraseodidactic research Following in the footsteps of Charles Bally (1909), many studies on phraseology and phraseodidactics outline the necessity of developing phraseological competence, however, without always distinguishing between productive and receptive proficiency (cf.infra).Mel'čuk, mentioned in our introduction, thus asserts: Un natif parle en phrasèmes.Si ce postulat crucial est accepté, et nous l'acceptons, il apparaît alors clairement que l'apprentissage systématique des phrasèmes est indispensable dans l'enseignement d'une langue, que ce soit la langue maternelle de l'apprenant ou une langue étrangère, et indépendamment de l'âge ou du niveau d'éducation de l'apprenant.(Mel'čuk 1993: 84) Some of the implications of Mel'čuk's assertions, which seem to be shared by numerous researchers in this domain (cf.supra), need to be examined from a more nuanced perspective as to their pertinence for foreign language learning: • "A native speaker talks in phrasemes."If he or she obviously uses prefabricated communicative means, all of his/her speech activities are by no means subject to the sole "idiom principle" (Sinclair 1991).A speaker equally has to call upon the "open choice principle" (ibd.) in order to communicate successfully.Even if one accepts that there is far more formulaic speech than is assumed nowadays, it is highly unlikely that, strictly speaking, every imaginable language activity is a reproduction of preexisting construction units.However, only -future -analyses of mass data will determine to which extent speakers effectively revert to formulaic communicative structures.• It goes without saying that the study of sufficiently vast corpora has to reveal which types of phrasemes are de facto employed by native speakers when communicating with whom, in which situations, in which way, to achieve which goal, etc.And it seems hardly justified to teach idioms to non-native learners which are not being recurrently used by natives themselves (cf.section 3.1 studies on the frequencies of idiom use).• Obviously, receptive and productive competence, not distinguished by Mel'čuk, cannot be treated on the same level.We are able to read and understand James Joyce's writings, but are we capable of expressing ourselves like the author of Dubliners?• Native speaker language acquisition in natural contexts and foreign language learning in institutional settings cannot be considered on the same level.Irrespective of the fact thatbilingual -native speaker competence is only very exceptionally attainable, 4 a non-native speaker is not expected to express him/ herself like a native for reasons which will be exposed in section 3.
Concurring with Mel'čuk's postulation, the Common European Framework of Reference for Languages (CEFRL; Council of Europe 2001: 112) equally stipulates "Good command of idiomatic expressions and colloquialisms", yet actively only from the C1-level onwards, offering nevertheless a hardly more nuanced view of phraseological competence.In fact, apart from obviously highly foreign language (= FL) relevant routine formulae such as greetings (good morning), "phrasal verbs" (put up with), "compound prepositions" (in front of) and collocations (make a speech), the section on "fixed expressions, expressions, consisting of several words, which are used and learnt as wholes" (ibd.: 111-112), also lists proverbs (e early bird catches the worm), "relict archaisms" (Be off with you!), "phrasal idioms" (He kicked the bucket or He drove hell for leather), etc. whose utility and necessity for an even advanced FLL are highly debatable as will be demonstrated.
(González Rey 2010: 1) Analogously to the observations put forward regarding Mel'čuk's assertion above, one cannot but notice the absence of a thoroughly differentiated treatment of what types of phrasemes a FLL should acquire at what age and stage of his/her learning process "from the very beginning" in pursuit of his/her communicative goals when talking about a particular topic in a specific social situation.
e following key section 2 therefore presents a detailed overview and discussion of criteria referring to the definition and classification of learner-relevant prefabricated communicative structures.

Definition and classification of prefabricated communicative structures
First and foremost, it is essential to circumscribe the concept of phraseme (2.1), then establish a classification of phraseme classes (2.2).

Towards a revised and enlarged understanding of formulaicity
In mainstream phraseology, definitions of "formulaic speech, as an umbrella term, 5 invariably recur to three concepts to distinguish offthe-rack recurrent verbal combinations from spontaneously created constructions: • polylexicality or, according to Wood (2015: 3), "multi-word combinations", consisting of at least two lexical autosemantic and/or synsemantic units; • stability: both lexical elements and their combination are more or less cognitively fixed and/or recurrently used in a specific form; • idiomaticity: the sum of the meaning of its constituents is not equivalent to the phaseological meaning, i. e. their meaning is semantically non-compositional (e. g. drop a brick = "unintentionally say or do something embarrassing, tactless, or indiscreet") (e Free Dictionary s. v. drop a brick), or else they have a syntactically irregular structure (e. g.Like father, like son but not *Like whiskey, like wine) or suffer from transformational restrictions (e. g.He dropped a brick but not *the brick he dropped). 6Given that most phrasemes correspond in fact to regular syntax and semantics, polylexicality and (relative) stability define their scope in a wider sense, whereas those which are also idiomatic but whose meaning is not transparent from a strictly semantic point of view, 7 belong to the class in a narrow sense.
As it is, existing definitions of "multi-word expressions" based on the aforementioned criteria no longer seem to reflect state-of-the-art research on formulaic speech.In the last decades, highly relevant aspects have in fact come to light, especially formulaic texts (Gülich 1997) and "constructions" (cf.infra).e purpose of the present section therefore resides in the discussion of new insights into the delimitation of formulaic speech.Our considerations are based on the pivotal hyponomic concept of "polyfactoriality", which will be delineated via the following hypernymic sub-concepts: the scope of formulaicity, the stability of formulaic expressions, their idiomaticity as well as their multimodal nature.ese considerations will lead to a revised definition of prefabricated communicative structures.
Defining polylexicality as a prerequisite for classifying an expression as a phraseme would logically lead to the exclusion of monolexical routine formulae 8 like hello, bye, thanks, etc. from the category of formulaic expressions, whereas their multi-word equivalents, having exactly the same communicative functions, good morning, goodbye or bye bye, thank you would be included.Even though several researchers, for instance Burger ( 4 2010: 28-29), Granger/Paquot (2008: 32) or Stein (1995: 27), include mono-lexical units in the class of pragmatemes, their definitions of the concept "phraseme" still call upon to the central criterion of polylexicality.However, unlike simplex nouns such as desk, cat, knife or roof, for which only cotexts can be determined, clear-cut contextual and situational components of use can be described for hello, buy or thanks.A distinctive definition of "phraseme" should therefore abandon the criterion of polylexicality whenever contextual or situational elements of use can be inexorably linked to a monolexical "routine formula".e decisive criterion would thus be "polyfactoriality", no longer "polylexicality", where situational factors can be closely linked to its use, for instance for thanks: the type of speech act executed (expressing gratitude), its stability (use of thanks being highly foreseeable and expected in this situation), its sequential position (following a beneficiary act), a subsequent activity (you're welcome) is likely to occur, specific kinesic activities (facing the benefactor, friendly facial expression, smiling, handshake, kissing in France).Furthermore, probably non-defining factors for thanks could be its prosody (vocative chant), stylistic level (neutral of informal), relation between participants (relative proximity or neutral) or its semantics (a priori transparent).e one-word formula thanks can consequently be considered as a phraseme via its polyfactorial nature.Notwithstanding their metaphorical nature, compound monolexical lexemes such as scapegoat, mainstream or spearhead are not included in the category given that semantic transparency is not a defining criterion for the phraseological nature of monolexical units, be they simplex or compound.As far as polylexical items are concerned, however, idiomaticity can be a distinctive measure for the differentiation of idioms from collocations or proverbs from commonplaces.
Certain types of polylexical phraseological expressions, especially proverbs, common places and idioms, and to a lesser extent collocations, can be equally affected by polyfactoriality as they can be subject to specific syntactic, prosodic, semantic, contextual or situational constraints.In fact, the use of the English proverb Strike while the iron is hot, polylexical, semantically non-compositional and used in a more or less stable form, is employed in a specific situation: Speaker A offers advice to speaker B, recommending action in order to obtain success when trying to solve a problem while conditions are most favourable.As for prosody, strike and hot are probably accentuated, and the proverb has to be pronounced in a convincing or at least non-hesitating manner.Furthermore, the speaker's general FL competence and social status have to be in accordance with his/her use of this high-level stylistic expression.is implies that a young person, especially a non-native speaker, would a priori not give advice to a much older interlocutor in this proverbial way (cf.infra).e polyfactorial criteria for Strike while the iron is hot would therefore comprise: its lexical form ("multi-word" in this case), metaphorical nature, prosody, problem to solve, age, and social status of speaker, his/ her relation with the addressee, etc.Not only routine formulae thus bear a pragmatic footprint; in fact, elements of use could also be described for most formulaic expressions.

Scope of formulaicity, stability of prefabricated communicative structures and their idiomaticity
As developed in the previous section, a phraseme may be composed of one single lexical unit as long as another factor is closely linked to its use.Whereas a great number of routine formulae are indeed monolexical, the majority of phrasemes described so far by phraseological research are polylexical, even if they also possess polyfactorial characteristics.Like Burger's ( 4 2010), most phraseological classifications do not go beyond the syntagmatic (good morning, put up with sth.) or sentence level (proverbs, commonplaces); textual phraseme structures are mentioned (e. g.Burger 4 2010: 37), but only exceptionally studied. 9Some authors like Luckmann (1988) go even further to include communicative genres such as wedding ceremonies or court trials, which he defines as "culturally and historically specific socially conventionalized and formalized solutions to communicative tasks" (Luckmann 1988: 281).ese genres have a fundamentally multimodal structure, comprising any type of linguistic, situational or even physical element constitutive for the communicative event.
At a lexicogrammatical level, traditional phraseology includes "phraseological models" or "patterns", for instance [xNP aer xNP], e. g. day aer day, night aer night, or [xNP/ADJ by xNP/ADJ], e. g. step by step, word by word, hence "syntactic frames" offering, on a paradigmatic level, slots for lexical completion by specific lexical classes.More recently, the scope of these lexicogrammatical patterns has been considerably widened by approaches from different research paradigms10 which one may, by simplifying, subsume under the heading "constructionist".According to Fillmore/Kay/O'Connor (1988) "constructions" are "form-meaning-pairs", composed of a syntactic frame lexically provided or providing slots to be filled more or less freely by items belonging to specific lexical classes.Two classes of "constructions" can be distinguished: "substantive idioms", corresponding by and large to classical phrasemes, thus grosso modo, lexically stable, and "formal idioms", morphosyntactic frames with slots partially or not at all filled lexically.Whereas "formal idioms" comprise the aforementioned phraseological model [x is x], they are obviously not limited to this category.Firstly, because co-and contextual as well as situational factors of use are an integral part of "constructions" while phraseological patterns are confined to syntax and lexis.But also, because the realm of "constructions" goes far beyond what has been researched so far under the heading "phraseological models" (Modellbildungen or Phraseoschablonen in German).Lewis (1993: vi) even believes that "Language consists of grammaticalized lexis, not lexicalised grammar."Presumably, the truth lies somewhere in the middle as oen: some lexemes are employed in a preferred syntactical form as opposed to syntactical models which allow for a great -but rarely unlimitednumber of lexical realizations.Having said that, a lot more lexemes or syntactic structures than one would expect could be associated more or less closely to certain types of "constructions".In fact, Sinclair's "idiom principle" -"A language user has available to him or her a large number of semi-preconstructed phrases that constitute single choices, even though they might appear to be analysable into segments."(Sinclair 1991: 110) -oen seems to outmatch "free choice" or at least confine it by way of grammatical norms or rules and paradigmatic lexical restrictions. 11 Hereaer, some instances of "constructions" so far described by corpus-based research in this paradigm: Since the beginning of phraseological studies, which started out from the idea of "frozenness" as a defining criterion, the notion of "fixedness" or "stability" of prefabricated expressions has been largely put into perspective.Burger ( 4 2010) considers: Nahezu alle Operationen, die man aufgrund der eigenen Sprachintuition für "nicht möglich" halten würde, lassen sich in realen Texten auffinden, gesprochenen und -was man noch weniger erwarten würde -geschriebenen Texten.
(Burger 4 2010: 23) Elspaß ( 1998) is even more outspoken: Mel'čuk (2012) therefore replaces the not ion of "stability" by "constraint" in order to retain degrees of relative fixity: A phrase is non-free (= phraseologized), iff13 at least one of its lexical components L is selected by the speaker in a linguistically constrained way -that is, as a function of the lexical entity of other component(s).
(Mel'čuk 2012: 33) 14 Mel'čuk's definition does not nevertheless explain in which way a language user may recognize an expression as being lexically constrained in situ (as he cannot carry out corpus-based semantic analyses to decide).Provided a speaker is familiar with an existing form of a phraseme, specific structural and/or lexical elements may provide for its recognition, which we chose to call the "smallest common phraseological" or "formulaic denominator".e most clear-cut lexical indicator for the existence of a phraseme are certainly unique "cranberry words" (spick and span, sandboy, tenterhooks) which survive in single specific phraseme cotexts only.Generally speaking, it seems that certain structural patterns, probably containing one typical lexical unit, may trigger off suitable coand contexts provided the interpretation of a respective locution.Word play procedures in fiction or advertising, particularly productive due to the coexistence of literal and figurative meanings of idioms, provide evidence for our "smallest-commonformulaic-denominator" hypothesis: as long as certain elements are maintained, recognition of a phraseme is feasible.
To delimit phrasemes in a narrower sense, research refers to the criterion of "idiomaticity", generally defined via semantic noncompositionality (cf.2.1.1.)irregularities.Burger (2010: 36-58) distinguishes full (kick the bucket), partial (drunk as a lord) and non-idioms or collocations (commit suicide) according to their semantic compositionality.Yet, for a non-native speaker even the wording of collocations is by no means transparent from an encoding point of view as their mother tongue might employ another verb: make a decision, prendre une décision in French, eine Entscheidung treffen in German.Hausmann (1997), as quoted above, therefore rightly concludes that everything is idiomatic for the learner of a foreign language, given the existence of combinatorial and selective preferences which have to be learned.In the same vein, Feilke (1998: 74) advocates the conviction that an 'idiomatic footprint' (our translation of German idiomatische Prägung) affects any language production.
As a consequence, not just semantically non-compositional idiomatic expressions have to be learned by non-native speakers, but intrinsically almost any formulaic communicative structure being subject to an "idiomatic footprint".What is more, language in general and, considering their polyfactorial nature, prefabricated expressions in particular, are also affected by a "pragmatic footprint" considering specific -co-and contextual, sequential, situational, social, stylistic, etc. -conditions of use which have to be acquired for adequate and successful implementation (cf.Schmale 2020: 11-15).
On the grounds of the preceding reflections on the nature of any type of prefabricated communicative structure, be it situated on the phrasal, utterance, constructional, textual, or genre level, Schmale (2020) forges the following revised definition of a formulaic sequence in a wider sense: A formulaic sequence, in a wider sense, is stamped by its polyfactorial nature and the presence of a smallest common formulaic denominator, and, in a narrower sense, by the existence of an idiomatic footprint.(Schmale 2020: 24) is definition requires several specifications: • Formulaic sequences (= FS) have a polyfactorial structure, i. e. a stable combination of at least two verbal and/or contextual, situational and/or corporal elements.• An FS is recognized by a competent member of the speech community via a smallest common formulaic denominator and known by its recurrent use.• In a narrower sense, a FS can be marked by an idiomatic or pragmatic footprint consisting of syntactic or semantic deviances or transformational restrictions and/or by specific connotations or conditions of use.

Classification of formulaic communicative constructions
Based on the preceding considerations as well as Burger's ( 4 2010) classification of phrasemes, the following categories of prefabricated communicative structures have to be retained.Burger ( 4 2010) distinguishes three main basic phraseological categories for German, which may also apply to other languages: • "Referential phrasemes" refer to objects, processes or states of affairs.ere are two types: "nominative phrasemes" with a syntagmatic structure, and 'propositional phrasemes" having utterance value.ree sub-types of "nominative phrasemes" exist: "idioms", which are semantically fully non-compositional (have an axe to grind, push the daisies); "partial idioms" having at least one semantically compositional element (as happy as a sandboy, drunk as a lord); semantically fully compositional "collocations", at least as far as decoding goes 15 (brush one's teeth, deliver a speech)."Propositional phrasemes" are subdivided into three classes: "proverbs" having non-compositional semantics (All that glitters is not gold, e early bird catches the worm), "commonplaces" (All's well that ends well, Don't put off until tomorrow what you can do today) and "fixed phrases", compositional or not, but characterized by the presence of an exophoric deictic element (at's how it goes, at takes the cake).• "Communicative phrasemes", "routine formulae" or "pragmatemes" serve the production of stereotyped speech acts in numerous situations, such as salutations (hello, bye bye), excuses (sorry, Excuse me), acknowledgements (thanks a lot -you're welcome), compliments (Happy birthday, congratulations) and many others, indispensable in all types of social encounters.• "Structural phrasemes" serve the creation of syntagmatic relations between parts of an utterance, such as either -or, A as well as B.
Considering that any valuable grammar (book) describes this type of phraseme under the heading "conjunctions" or "coordinators", this sub-category will not be given further attention in the present contribution, conceding nonetheless that corpus research as to their use in oral conversation has not yet been systematically carried out.
Apart from these basic categories, Burger ( 4 2010) establishes a class of specific phrasemes.ese are "phraseological models"; "binomials" (spick and span; hard and fast); "comparative phrasemes" (pleased as punch, deaf as a doorpost); "kinegrams" designating corporal action (knit one's brows, shrug one's shoulders); (famous) "quotations" (to be or not to be, A kingdom for a horse); "onymic phrasemes" serving as proper names (the White House, the Red Cross, National Health) and "phraseological terms" from various domains (the categorical imperative, temporary restraining order, hostile/unfriendly takeover).As the phrasemes of this special class are situated transversely to the basic category, a member can be idiomatic or not, of syntagmatic or utterance value, etc. us, the comparative phraseme deaf as a doorpost would be partially idiomatic; knit one's brows or spick and span plainly idiomatic.

Formulaic communicative constructions for non-native speaker competence
Starting out from the discussion of constitutive defining criteria for phraseological and formulaic sequences, it is now time to deliberate on those prefabricated structures which are essential the FLL.e elaboration of an operative communicative competence evidently has to start out from a learner's future linguistic and communicative needs.ese can, of course, not be determined by imitating what a native of similar age or social status might use in an equivalent situation.
e choice of linguistic structures in general and of formulaic language in particular therefore has to consider what is necessary so as to enable the learner to implement his communicative strategies.As developed in section 1, phraseodidactics attaches the utmost importance to the acquisition of phraseological competence without, however, providing a detailed description of targeted phraseme types.
Several substantial arguments run counter to the teaching and learning of figurative and/or metaphorical idioms for a productive use by an FLL: • Research on the German language is not always corpus-based according to the state-of-theart methodology and if so, mostly centred on corpora of written, mostly journalistic or fictional language productions (cf.Schmale 2021c: 203).It is by no means certain, even highly unlikely, that this highly elaborate text type adequately meets the needs of an FLL, regardless his/her age or proficiency level.• Furthermore, the study of mass data of English, French and German reveals that "[…] it seems reasonably certain that most phrasal lexemes (i.e. idioms; GS) are indeed infrequent".(Moon 1998: 100).Moon observes that many of the idioms known to competent English speakers are absent from the 211 million token "Bank of English".Grant (2005) arrives at the same conclusion: […] a corpus search of the final total of 103 "core idioms" was carried out in the British National Corpus (BNC).e search revealed that none of the 103 core idioms occurs frequently enough to merit inclusion in the 5,000 most frequent words of English.(Grant 2005: 429) Colson's research equally confirms these findings: A number of recent studies […] show clearly that the frequency of verbal idioms is very low.ere is no doubt about the importance of idioms as a whole in most texts, but if one is looking for a particular verbal idiom (e. g.Spill the beans) in a given corpus, its relative frequency will not be very high.
[…] Many verbal idioms of English, French and Dutch (and probably of all European languages) correspond to a frequency of less than 1 PMW, i. e. their occurrence in a corpus is inferior to one in a million words.(Colson 2003: 48) For French, Siepmann/Bürgel (2019), in a study of the Corpus de Référence du Français Contemporain (CRFC), 16 reach analogous conclusions: figurative expressions are negligible from a quantitative point of view as opposed to bigram-combinations such as un peu, parce que, par exemple, en plus which are frequent.
As for German, Schmale (2009) states that the conversational use of idiomatic expressions listed in the "phraseological optimum" (Hallsteinsdóttir/ Sajánková/Quasthoff 2006) is extremely limited; hardly any of the idioms retained are employed in a 42-hour German talkshow corpus.
In light of this insight into the actual discursive presence of figurative syntagmatic and especially sentence-equivalent proverbs or commonplaces, it seems hardly justifiable to integrate them into the indispensable lexical program for an operative foreign language competence.In fact, would it be logical to put language material at the disposal of non-natives which is not being called upon -at least in oral everyday communication, which should be the general guidelineby native speakers themselves?e answer tends to be a negative one considering two major obstacles: • e use of figurative expressions is subject to multiple syntactic, semantic, connotational, co(n)textual, situational and social constraints which can hardly be acquired in an institutional setting, especially as these have not been sufficiently described to this day by lexicographic research.Roulet (1989) rightly points out that adequate use of an idiom presupposes knowledge of its register, language level and situational variables.To quote an example a French example: Bárdosi/Ettinger/Stölting ( 3 2003: 88-89) mention, under the heading "conversation", the idiom tailler une bavette (~ gab, shoot the breeze), without mentioning any stylistic or social connotations of this rather colloquial expression.ey simply indicate bavarder (chat) as its meaning.A non-native speaker, be he/she at C1 level or not, would most certainly not succeed in communicating adequately if he/she used tailler une bavette (shoot the breeze-level) instead of bavarder (chat).Native speaking interlocutors would probably find it amusing.• In fact, another significant parameter of exolingual communication comes into play: "culturemes" (cf.Poyatos 1976; Oksaar 1988), which refer to a system of linguistic and behavioural preferences favoured by a language community.Whereas specific behaviour is expected from a non-native (and its absence probably sanctioned) in many domains of everyday life, e. g. table manners, dress code, tips, etc. as well as linguistic routines, for instance the use of expected routine formulae, politeness phenomena, non-verbal behaviour, other types of linguistic activities are not meant to be used by non-native speakers or rather reserved for natives, belonging to the culture in question as opposed to the foreigner who does not.A nonnative speaker, having an imperfect command of the language (marked foreign accent, grammatical and lexical errors), is not supposed to use highly figurative or metaphorical expressions, presumably rarely called upon by a native him/herself, for mainly two reasons.Firstly, because the gap between his/her manifested language competence and idiom use, stylistically situated on an elaborate level, is too vast; and secondly, because a foreign speaker, even one with a native-like proficiency, is not expected to use metaphorical language assimilating him/her to a native speaker.Dobrovol'skij/Lûbimova (1993) state: Wie unsere Beobachtungen sowie auch manche experimentellen Ergebnisse 17 zeigen, bewerten Muttersprachler die Verwendung bildhaer, expressiver Mittel durch Nichtmuttersprachler längst nicht immer positiv.Zum einen hängt das damit zusammen, daß der betreffende Sprachgebrauch o unkorrekt oder situativ inadäquat ist Supplementary empirical evidence would obviously be necessary to prove Dubrovol'skij/Lubimova's hypothesis; however, our personal 40-yearexperience as a non-native speaker in France supports their statement.Native speakers do not always seem to react to non-natives' use of stylistically marked idioms positively.

Indispensable formulaic sequences for the foreign language learner
In light of the high degree of prefabricated linguistic elements in written and spoken discourse, it is of course inconceivable to completely renounce any teaching of formulaic communicative structures -apart from idiomatic expressions for the reasons developed in the previous section.e formulaic structures which are indeed absolutely fundamental to any operative communicative competence as they cannot be replaced by non-phraseological means of communication are routine formulae (3.2.1.),collocations (3.2.2.) and constructions (3.2.3.), apart from rare partial idioms, for example blinder Passagier (a stowaway, literally a blind passenger), as well as onymic phrasemes or phraseological terms for professional purposes.

Routine formulae as a fundamental element of communicative competence
From the very start of the foreign language learning process, the acquisition of routine formulae, pragmatemes, communicative phrasemes or pragmatic idioms is essential for the development of an operative communicative competence as a great number of discursive activities necessitate the use of these expressions which do not only play a fundamental role in almost any type of communicative situation in oral discourse (greeting, thanking, apologizing, congratulating, etc.), but also in many environments of written language productions.e recourse to these formulae is inevitable, given that they can generally not be replaced by non-formulaic expressions as highly conventionalized and expected terms.Coulmas (1981: 119-120), who presented the first comprehensive study of "routine formulae", distinguishes five major types: 18  Table 1: Coulmas' classification of routine formulae ere is no doubt that routine formulae have to be integrated into realistic learner-relevant contexts, dialogical or textual, as a result of the study of naturally occurring discourse productions in order to retain any factor permitting successful usage.It seems, in fact, that dialogues from teaching manuals are still too oen based on the textbook authors' intuitions (cf.Schmale 2004), which can never yield equivalent results to the study of mass data.Even aer decades of discourse studies, one cannot trust one's accumulated empirical knowledge of language use; one should only "trust the text" (as Sinclair 2004 says in the title of his book).For obvious reasons, textbooks are not meant to reproduce authentic, naturalistic conversations, on the other hand these should serve as a basis for the development of realistic models integrating any relevant element appertaining to the use of the formula in question.

Collocations
e acquisition of collocations as recurrent and conventionalized lexical combinations of at least two lexical items is equally absolutely essential for the FLL as they have -from the encoding perspective -a noncompositional nature (cf.supra; 2.2.).In fact, the choice of one of the components may be arbitrary for the non-native whose mother tongue recurs to a different verb; for instance, get in touch, prendre contact, sich in Verbindung setzen.For both Burger ( 4 2010) and Mel'čuk (2013: 7, 9), 19 the latter being convinced that they amount to millions (sic) in a language, collocations consist of a base and a collocator.In most cases, the base is a noun phrase, the collocator a verb (make an appointment, meet the demands, come to an end, raise the alarm, etc.); more rarely, an adjective (brilliant, bright idea) or another noun (load of rubbish, suicide bombing).It is necessary to learn collocations as most verbs differ in English, French or German (cf.Schmale (2020: 26-27) for a table of occurrences), e. g. come to an end, zu Ende gehen, toucher à sa fin.Lüger (2019: 70) consequently pleads for an introduction of collocations from the very first stages of foreign language learning onwards, a perfectly logical postulation as Sinclair's "open choice principle" does not apply here.Even if phrasemes are generally open to modifications, collocations are not.Benigno/Grossmann/Kraif (2015:81) therefore attach the utmost importance to their teaching, which they consider as being insufficient or even neglected.ey also point out that frequency of use cannot be the sole criterion for learner-relevance.In point of fact, a collocation may be rarely found in a large corpus and nevertheless be highly useful for a learner, e. g. make an offer.As for any phraseme, frequency in a corpus alone does not allow to determine its importance for the FLL.To be more precise, frequency should be evaluated with regard to learner-relevant communicative contexts.So far studies of mass data have not been carried out under this angle.

"Constructions" as lexicogrammatical structures
In addition to routine formulae and collocations, "constructions" as lexicogrammatical combinations could be present in a language by a much greater number than so far described.While the studies mentioned above (cf.2.1.2) start out from an inductive "corpus-driven" perspective, we have adopted a more deductive "corpus-based" approach, choosing some grammatical phenomena of German not adequately mastered by foreign learners even aer many years of study.ese are, for instance, the correct use of the modal verbs sollen and müssen, particularly difficult for French learners as, in French, there is only one single verb for both modals (i.e. devoir), or else the differentiation of stative (with auxiliary sein) and dynamic passive (with werden), where again there is one single auxiliary for the formation of both (i.e. être).More recently, we studied the conversational use of the modal particle denn, which does not exit in English or French, in an attempt to describe its use in specific constructions (cf.Schmale 2021b).
Quantitative studies of mass data reveal that routine formulae, collocations and constructions are employed within specific syntactic and lexical environments which can be described in detail.e most frequently used types should then be adapted to the communicative needs of FL learners.We are indeed convinced that almost any phenomenon of a language can be described, provided one studies sufficiently representative large corpora, by way of lexicogrammatical constructions which must then be "didacticized" for FLL.In naturally occurring situations, language learning proceeds by the imitation and acquisition of communicative models, not theoretical rules.Fully conscious of the constraints of institutional FL learning, which can obviously not be compared to natural language acquisition in systematically authentic situations, why should one not try and proceed by way of lexicogrammatical constructions without recurring to abstract theoretical rules, whenever grammar-based approaches do not yield the expected success (cf.Schmale 2016)?
As a first approach to the description of constructions for German one might start out from the 34 'phrase models' (Satzbaupläne) proposed by the reference dictionary Duden 4 (Dudenredaktion 2009: 922-924) and study the presence of specific lexical configurations filling the slots within the depicted morphosyntactic frames, e. g. [subject + verb + noun phrase (accusative)]: Wir bauen ein Haus ('We are building a house').While this sentence model evidently allows an almost unlimited number of lexical completions, it is not excluded that, studying it in a learner-relevant corpus, specific lexical forms transpire, for instance the use of first and third person singular (ich, er/sie), "simple" verbs (sein, haben, nehmen, möchten) or noun phrases designating objects or states of affairs relevant to the learner's living environment (food, drinks, clothes, transport, sports, hobbies, etc.).
One might alternatively recur to valency models determining the complements accompanying a specific verb, thus setting out from a lexical element which provides slots for different complements, e. g. trinken: Ich trinke Milch (I am drinking milk).

Methodological principles for the selection and transmission of formulaic communicative structures
Eliminating proverbs, commonplaces and idioms from the core of foreign language lexical competence, does by no means imply that any text in a wider sense should be systematically expurgated of this type of phrasemes.As a rule, language teaching material should neither be chosen on behalf of its idiomatic or phraseological resources nor of its absence.Text content and style should always correspond to standards of naturally occurring spoken and written discourse, obviously in agreement with learners' language proficiency and their communicative needs.erefore, if a document comprises idioms, which is most likely the case for literary or journalistic texts studied at advanced stages, these should naturally be explained to become items of receptive competence.However, for the reasons detailed above, learners should not be expected to use them actively if a non-phraseological alternative is available?
Starting out from these considerations the following linguistic and (phraseo)didactic principles should guide any treatment of phrasemes in general and of idioms in particular in the foreign language class, thus an institutional setting which can of course not be compared to language contact in non-institutional settings (which have not been treated in this paper): • e focus of learning and teaching is put on routine formulae, collocations and lexicogrammatical constructions, at least until the learner reaches B1 or B2 level, particularly in spoken or everyday language.• e description of forms, conditions of use and functions is without exception based on large corpora of naturally occurring communicative events.• Phrasemes are invariably presented within contexts inspired by authentic situations without, however, reproducing their exact structure or wording.• Any phraseological expression is chosen according to learners' actual and targeted/ proficiency, by differentiating productive and receptive competence -and obviously learners' age as children or adults do not need to master the same types of prefabricated structures.• Phrasemes, as any language material in general, are chosen in keeping with the communicative situations learners may have to master in target language situations.• One might initiate learners to a recording sheet (cf.Lüger 1997: 118-119), considering form, conditions of use, occurrences, connotations, etc. in order to constitute a personal phraseme lexicon.
Even if foreign language teaching textbooks have made considerable progress concerning text model quality, it seems that text books still do not always reflect communicative reality.In fact, corpus research reveals that phraseme choice is oen deviant from attested authentic usage in everyday language.In fact, from a quantitative point of view, speakers rarely call upon those figurative idioms one encounters in collections of phraseological expressions.Only corpus linguistics, based on a large number of naturally occurring communicative manifestations, be they spoken or written, are apt to retrace authentic phraseme use.Researchers and teachers do have to "Trust the text!", following John Sinclair's maxim (2004).