Some remarks on pragmatics schemata and second language teaching

Se debe entender el siguiente articulo como un intento teorico de aclarar los limites conceptuales relativos a los terminos "competencia comunicativa" y "capacidad linguistica". Se espera asi que una mejor comprension de ambas nociones desde una perspectiva linguistica proporcione al ensenante el fundamento para una reorientacion metodologica seria. De tal forma, los principios aqui expuestos se pueden aplicar en la ensenanza de una segunda lengua


O. MODERN LINGUISTICS AND METHODOLOGY
The upheaval created by transformational grammar, which began to affect language pedagogy concerns by the mid-sixties, was followed in the seventies by a feeling of uncertainty among theorists as to the directions of development of L2 methodologies. The previous decades had seen the flourishing and vanishing of a number of methods such as the Intensive Language Teaching ofthe American Army during World War II, the well-known audio lingual (U.S.A.) and audiovisual (France/Great Britain) methods, the use of the language laboratory and others. Then the interest aroused by the distinction between 'competence' and 'performance', which marked the heyday of psycholinguistic research, resulted in new mcthodological oricntations of a mentalistic sort. Teaching based upon structural drills startcd to disappear and language learning began to be envisaged as the intcrnalization of rule systcms. The learning of syntactic relationships, it was thought, could not be successfully achieved through mere practice with surface structure drills since therc werc dccp structure factors which could deal more effectively with structural similarities, diffcrences and ambiguities. Furthermore, generative grammarians placcd a lot of emphasis on the creative nature of language. According to them, a speaker could produce an infinite number of sentences by a limited number of rules. The natural conscquence of this principie for language pedagogy was formulated by Diller (1978: 25) as follows:" a language learner does not ha veto store a large amount of ready-made scntences in h is hcad; he just needs the rules for creating and understanding these sentences".
Transformational grammar was al so conccrned with the common elements of all languages in contrast to the formcr structuralist emphasis on the idiosyncrasy of each language -as it was cvidenced by the prolifcration of contrastive studies such as those by Fries (1952) and Lado (1957). Audiolingualism in pedagogy, behaviourism in psychology and structuralism in linguistics were soon to be overriden by the ncw emerging rationalism.
In the 1970's Chom sky' s views underwcn tan u m bcr of revisions and modifications which affected language mcthodology once again. An outstanding example of this is the influenceofDellHymcs's(1979)conccptof'communicativecompetence'onanumber of proposals for the claboration of language tcaching syllabuses, which began to incorporate semantic categories of mcaning (notions), modality and communicative functions (for example, Wilkins, 1976), as well as syntactic rules. In the same way, the scholars in the Council of Europe Modero Languagcs Project, on the grounds of semantic and sociolinguistic work, addcd to this an inventory of situations in terms of learner's roles, settings and topics (van Ek, 1975). It is al so worth mentioning a number of projects based on sociolinguistic studics relatcd to Discourse Analysis (Allen and Widdowson, 1974), including Convcrsational Analysis.

l. COMPETENCE AND CAP ACITY
In all of thc post-structuralist studics mcntioncd abo ve the objective secms to ha ve been the acquisition by the language lcarncr of native-like competence, whether of a linguistic or a communicative varicty. The notion of linguistic competence, for example, is of central importancc to studies in Error Analysis and lnterlanguage (Selinker, 1972;Corder, 1981 ). According to thcse, an L2learner' s speciallanguage can be considered an idiosyn~ratic dialcct with a number of unstable rules which do not belong to either thc mothc~·tongue or thc target languagc. The methodological objective would be to cnablc thc learncr to bring his language behaviour into line with the conventions of the targct language in a proccss which, it has bcen suggcsted, would be natural: the leamcr, bcing in contact with the targct language, would formulate alternative hypotheses and gradually climinatc the wrong ones. The objective in this orientation is therefore the acquiring by the learner of linguistic competen ce, that is, of linguistic (or grammar) rules.
Criticism against this view by numerous authors (Allwright, 1979;Allen, 1977;Wilkins, 1976;Brumfit, 1979;Finnochiaro and Brumfit, 1983, among others) has been easy on the grounds that teaching a language is teaching how to use it, the target then becoming what is known as the attainment of communicative competence. This would include knowledge not only of grammar rules and vocabulary but also of: -Rules of speaking, which regula te, for example, when and how to begin or end a conversation, when to spcak into it and what to say, how to make use of address forms according to the situation. All thcse factors have been widely dealt with by studies in Conversational Analysis (cg. Coulthard, 1985).
Although the term 'communicative compctence' and its implications have been widely accepted, it has been suggcsted (\Viddowson 1983(\Viddowson , 1984) that it should not be treated as a language tcaching objectivc, that, rather, competcnce is to be achieved through the exercise of what Widdowson has termcd 'language capacity', the ability to make meaning at any lcarning stage with thc rcsources at hand: The human capacity for making mcaning out of linguistic resources is not, then, confined within competence. N or is it simply conv crtcd into competence in the language acquisition proccss .... Wc are led to bclicve that the creative force is channelled into a code and finds expression only in the production of sentences according to rule. But the fact that we are able to produce and intcrprct utterances which do violence to such rules makes itclcarthatcrcativecapacity has an indcpcnclcntcxistence (Widdowson 1984:246) It is assumed that the lcarncr will tcnd to accomodatc his linguistic behaviour to the normas he realizes that this will in crease his communicative capacity. And capacity is never lost in this process. Capacity docs not beco me compctence. It enables the learner to achieve competencc.
When we use language to communicatc (to make meaning) we make use of all available linguistic resourccs. But thesc rcsourccs are not confined to rule systems. There are also a number ol principies which guide our production and interpretation of utterances. Now I would like to suggest that thesc principies, which are in general the concern of pragmatics (see Lecch, 1983), are rclated to the concept of language capacity. This is consonant with the well-known intercst of pragmatic studies with the language user, justas it is evidenced by Charles Morris's (1938;see Morris, 1971) original division of semiotics into the study ficlds of 'syntax' (the study of the formal relations of signs to onc anothcr), 'scmantics' (the rclation between signs and objects) and 'pragmatics' (the relation of signs to thcir intcrprcters). This is particular! y true of Carnap's (1938) understanding of thc thrcc ficlds (as quoted in Lcvinson, 1983:2-3): If in an investigation cxplicit rcfercnce is made to the speaker, or to put it in more general terms, to the u ser of thc language, then we assign it [ thc investigation] to the field of pragmatics .... If we abstract from thc uscr of the language and analyze only the expressions and their designata, we are in the field of scmantics. And, finally, if we abstract from the designara also and an1yze only thc rclations between expressions, we are in (logical) syntax.
Our assumption here is that we can undcrstand capacity better by studying the nature of pragmatic principies and that thesc principies, as it will be shown later, are procedural in nature. B ut before we go into dccpcr detail, let us try to put into perspective how such principies opcrate in linguistic comprehcnsion and production.

STRATEGIES,PROCEDlJRES AND SCHEMATA
First, we shall make a difference bctween 'strategics' and 'procedures' in discourse processing. Both are part of our knowlcdge and must be regarded asan open set. Since they belong, as suggested, to the domain of language capacity (not to competence) they are not rule-govemed. They can be takcn as working hypotheses about the semantic and syntactic structure of discourse, in othcr words, a way of making meaning out of discourse. Strategies are scts of procedurcs engagcd upon in the discourse process and they are probably idiosyncratic. For example, considcr how a butler might interpret the sentence It' s cold in here, uttcred by the lord of thc manor. In a context in which it is obvious that it is cold üns sentencc may in principie be considered an unnecessary remark, a breach of one of thc maxims of Quantity ("make your contribution as informativeas possible") from Grice'sCoopcrativePrinciple(see below). Butifthelord is supposed to be relevant (Maxim ofRclation), the butler will have to attempt to infer the true meaning of the utterance (a requcst to close an open window, for instance). Maybe something like Labov ancl Fanshcl's so-called 'rule of requests' and 'rule for indirectrequests' will hold (sec Labov and Fanshel, 1977). These rules are in fact sets ofprocedures which guide the interpretation of direct and indirect requests, and they are based u pon the recognition by both speaker and hearer of a number of preconditions. In our example, the butler might interpret that he is required to close a window (that a request for action was meant by the sen ten ce) if there is a need for the request anda need for the action, ifhe has the ability and thcobligation to perform theaction, and ifthe lord has the right to demand the action. Gi ven these preconditions, the set of pragmatic (nongrammatical) procedures employed in undcrstanding the sen ten ce above might take the following form: 1-Assume S's assertwn (A) is relevant 2-Search for any relevant action X to be performed by H 3-Analyse the conscquences of performing action X 4-If performing action X cancels the conditions described in A, then hear A as a request for that action This set of procedures can be considered a personal substrategy of the language user, part of a larger strategy or set of strategies activated in discourse. It should be noted that other authors do not make this difference between strategies and procedures (for example, van Dijk and Kintsch, 1983). However, I believe the distinction m ay be useful. Examine the set of procedures stated above. It is fairly evident that each procedure works on the basis of our world knowledge, that is, knowledge about everyday situations, events and the way they are interrelated. Thus, the butler assumes that what the lord says is relevant because, among other things, people do not tend to supply redundant or too obvious information unless they mean something else. This is the origin of Grice's (1957) well-known distinction between natural and non-natural meaning or "meaning-nn", the lattcr being equivalent to intentional communication. According to Grice, saying 'A mcans.,N something by x' is roughly the same as saying 'A intended the utterance of x to produce sorne effect in an audience by means of the recognition of this intention'. In this dcfinition the importance of the in tended effect of an utterance is explicitly acknowledged. But such an effect is impossible unless both speaker and hearer share sorne common background knowledge. For example, in order to understand the intended meaning of It' s cold in here, the butler needs to know that coldness may disturb his lord as well as how to act in such a situation (what his role as a servant is).
In Artificial Intelligence and cogni ti ve psychology, the kind of know ledge required for procedural principies to opcrate successfully has been studied under the general name of "schcmatic knowledge". Thus, 'schcmas' or 'schemata' havc been proposed as organizing structures which guide much of our memory, abstraction and interpretation processes (see, for example, Alba and Hasher, 1983). Asan illustration of the way schemata work in comprehension, the following cxamplc (borrowed from Leech 1983: 92) may be useful: A: In the end we got through the back door. B : Did you have to break the lock ?
In order for B's answer to be rclevant (that is, to enforce the Maxim ofRelation), we must assume that the door had a lock which prevented A from entering. The knowledge that doors ha ve locks is not, strictly speaking, scmantic knowledge sin ce this cannot be considered a distinguishing feature of all doors (unlike the fact that a door is used to close an entrance). Since A and B share the same knowledge about doors (and more particularly about a certain door) understanding can take place.
Our knowledge of the world or 'background knowledge' is, we may assume, arranged and disposed for activation at any moment in the discourse process. Although a great deal has been writtcn on the subject, a thorough account of how schemata are organized has notas yet been made and will naturally be beyond the scope of this paper. However a few hints asto their nature not found in the literature on the subject may be useful. In any case for two good reviews on the subject the reader may be referred to Bransford (1979) and Matlin (1989).
In the interpretation ofthc uttcrancc We got through the back door we postulated the activation of a door-schcma. The activation is possible thanks to a lexical clue (door). Of course schemata do not neccssarily function indepcndently (we may invoke severa! at the same time; eg. adoor may beassociated with a house, acar, a room, asafe, and so on). But the rclevant qucstion is what information is to be includcd in a given schema. We do not always considcr all the details of a concept when it is activated but this does not mean the details are not thcre. Also, schemata are likely to be high1y idiosyncratic, varying from individual to individual. I believe a good approach to the problem is to separate the stable from the unstable features.
Rumelhartand Ortony (1977), in an illustration ofwhat might be a breakschema, specify the following variables: 1) an agent, 2) a (brittlc) objcct, and 3) a method of doing the breaking. Other relcvant factors to be takcn into account are a change of state ofthe object (from whole to not wholc), a cause andan effcct, an instrument, a possib1e intentionality, etc. Perhaps one way to define thc composition of a schema invoked by a lexical ítem can be derived from a distinction made by Fillmore (1971:282).
According to him, the vcrb accuse prcsupposcs abad act but asserts that somenone did something, while the verb criticise prcsupposes that someone did something and asserts the act was bad (John accused Mary ofstealing and John criticised Mary for stealing, as the reported versions of, for instance, Yo u stole it! and S he shouldn' t have stolen it).
Both the presuppositions and the asscrtions can be derived from the lexical meaning of the verbs, and they are indepcndent of other scntential constituents. But the issue is whether presupposition and assertion are part of our schemata. If we use the negation test -as proposed in Kiparsky and Kiparsky ( 1971 )-, we will be able to see that in J ohn didn' t accuse Mary of stealing and John didn' t criticise Mary for stealing the presuppositions remain constant under negation whilc thc asscrtions change. Therefore, this type of presupposition will invariably be associated with the lexical item and may be considered a stable conslitucnt of its corresponding schema. This is also true of the break-schema whcre an agcnt, an object and a method can be pressuposed. The verb hear presupposes an audible objcct (*/ heard atable) and the verb see that the object is visible (*/ saw a thunderclap). All these can be taken to be stable elements of a schema, while in the door example, the existencc of a lock cannot be said to be so in the same way.
Generative semanticists (Lakoff, 1970;McCawley, 1971) have postulated a deep structure semantic analysis in terms of 'semanlic primiti ves' which might pro vide sorne useful hints too. For example, it is proposed that the verb kili could be analysed as "DO something to CAUSE to BECO ME not alive", where DO, CAUSE and BECO ME would be semantic primitivcs bclonging toa large number of verbs. This type of analysis may reveal semantic ambiguities, as in 1 almost killed him, which can be interpreted in three ways: 1-I almost caused him to become not alive 2-I caused him to almost bccome not a1ivc 3-I caused him to become almost not alive The formalization of a cognitive schema for kili should probably include these semantic primitives as stable features, apart from specifying an agent, a method andan animated object (that is, the existence of an agent in vol ves a cause-effect relationship and the effect results from a process of becoming). Such aspects of the meaning of lexical items seem to have been largely ignored by schema theories.
So farit has beco me apparent that classical semantic analysis isrelevant toa schema theory. However semantic theories show a few weaknesses which m ay be sorted out by reference toa schema theory. Let us try to illustrate this point. The sentence The wind blew would be typical of a number of descriptions. A semantic case analysis -such as those elaborated by Fillmore (1968), anci Grimes (1975)-would tell us that the wind is a semantic object (it is in the 'objective' case), while it is a syntactic subject. A usual demonstration of the value of a case grammar will show how a same case can adopt different surface manifestations. Thc 'instrumental' case, for example, appears both as subject and instrument in Charles opened the door with a key and The key opened the door. But here we are confronted with a problem. We need a set of semantic criteria to be able to decide what case we are dealing with. Thus an 'agentive' m ay be defined as the "typically anima te instigatorof an action", an 'instrumental' as the "inanimate force or object causally involved" (lhe entity by means of which an action is carried out), and 'objective' as the entity we describe as being in a state or in volved in an action. But how do we really know that the key is instrumental when no surface structure marker indica tes so? How is it that we can tell for sure that thc key did not open the door by itself? What is it that makes us assume that thcrc is an animate agent involved? -for doors can be opened by natural forces like thc wind. One obvious answer is that we recognize a key is an inanimate entity which cannot perform the described action by itself. Our common knowledge of th~ world includes the information that people open doors with keys. And this simple answer might intuitivcly be closer to what really goes on in our minds than the notion of selcctional restrictions.
Still we can make make one further point in favour of a schema theory. What is it that enablcs us to understand the mcaning of thc following two events which are not related by any grammatical markcr?: The wind blew. Thc door suddcnly slammed.
In order to interpret the two sentences coherently we nced to assume that the wind is, on the one hand, in the objcctive case, and on the other hand, that it is to be understood as 'agentive' (or, in other tcrminologies, as 'force', since it is inanimate). But there is no textual clue whatsoever which facilitatcs such an interpretation. We can but have recourse to our world knowlcdgc which tclls us that the wind can move things. This belongs to the world of our expcriencc. So wc draw the infercnce (ifboth sentences are to be interpreted as just eme utterance) that it was the wind that slammed the door.
The question now is, how do we make such inferences as the one just described? The answer is again quite simple. Wc make use of our language capacity in order to access information from our background knowlcdgc. An interesting si de effect of the use of our capacity is the possibility of modifying or increasing the amount of information stored away in our minds in preparedness for use. It is also capacity that enables us to do violence to grammar rules in order, for example, to achieve a conversational goal. This latter aspect is particularly deait with by pragmatics. In a sense, pragmatic principies sometimes become the bridge between our schematic knowiedge and particular situations (when lhey cxist) according to the goals of the Ianguage u ser. Or in othcr words, thc principies which regulate language use are partand parcel of pragmatics.
It has been suggested (Widdowson 1984) that there are two basic types of schemata, 'ideational' (descriptions of objects and situations) and 'interpersonal' (stereotyped sequen ces of predictable actions). In addition, wc find two types ofpragmatic principies which reguiate the way we make use of our structurcd world knowledge (or schemata). These principies constitute a type of knowledge different from that of schemata. They are procedures which serve to cngage upon schcmatic knowledge in the discourse process. Sorne of them ha veto do with the way we organize the content of our messages for the sake of efficicncy and effectiveness in expression. These are textual in nature, such as those stated by Slobin (1979:188-194): Be proccssibie (Processibiiity Principie) (iii) Be quick and easy (Economy Principie) (iv) Be expressive (Expressivity Principie) In broad outiine, the first principie is conccmcd with transparency and avoidance of am biguity; the second one wi th the ordering of the parts of the message; the third one is related to the amount of time and cffort invoived in cncoding and decoding (thus dealing with 'pronominalization', 'substitution' and 'ellipsis'); and the Iast one involves effectiveness in cxpression. In consequence, among these principies all relevant aspects of the textual componcnt of a systemic grammar, such as 'thematization' and 'information' (which corrcspond to 'proccssibiiity') and cohcsion (which corresponds to 'economy') are adcquately dcalt with. It must be understood that the principies of 'clarity' and 'expressivity' are largely depcndent on the enforcement of maxims (aspects) of the othcr two. For instance, adeq uate use of pronominalization, substitution and ellipsis may have a bcaring on the clarity and redundancy of a message. Let us take these examples: (1) Don't kili your wife with work. Let clectricty do it 1 (2) John Brown was guilty of the crime, and John Brown would ha veto pay for it2 In (1)  te use of substitution and pronominaiization. If we were not aware that it is a mistake, the sentence might be takcn, enforcing a procedurai principie from the interpersonai rhetoric,as a pun oniy meaningfui on the basis of a piece of ourbackground knowledge about husbands and wifes.
Then in (2) we ha ve a breach of the Economy Principie for the sake ofredundancy in order to achieve effectiveness of expression.
Whiie sorne ofthe procedurai principies are textual, others are interpersonal. Leech ( 1983), on a survey of the interpersonai rhetoric, states the existence of at Ieast two main principies which can be integrated into the Gricean paradigm of implicature. These are the Cooperative Principie Uust as stated by Grice, 1975) and the Politeness Principie. The maxims of the Cooperativc Principie are: Quantity: Make your contribution as informative as possible. Do not be more informative than required.
Quality: Do not say what yo u beiieve to be faise. Do not say that for which you lack adequate evidence.
Reiation: Be reievant. Manner: Be perspicuous. A void obscurity and ambiguity. Be brief, orderly and poli te.
The maxims of thc Poiiteness Principie would include a number of variables such as tact, generosity, approbation, modesty, agreement, sympathy, and phatic communion among others. An apparent breach of one of the maxims of the Cooperative Principie may invoive the enforcing of a maxim from the Politeness Principie (or from any other principie, Iike Irony).
If I say John' s afinefriend and it is obviously a lie I may seem to be flouting one ofthe maxims of Quaiity, but at the same time the hearer ofthe message will probably interpret itas an ironicai remark. Another typicai example of the flouting of the maxims of the Cooperative Principie can be illustrated by the following exchange: A: Let's go to the movies B: I have an examination in the morning The apparent irrelevance of B' s remark is considered to be a breach of the Maxim ofRelation, but it gives w''Y to poiiteness (thus avoiding an overt refusai) by means of an excuse. Incidentally, thc excuse is based on a piece ofbackground know Iedge shared by both interiocutors: when someone has an examination the next day, he is likely to want to study and may not have enough spare time for other activities. In the present case, the Tact Maxim is enforced. From all the above examples, it might be suggested that the maxims of the Cooperative Principie will usually work in reiation to ideationai schemata, whiie those of the Poiiteness Principie, the Irony Principie and the Interest Principie, are more liabie to engage on interpersonai schemata. Specially, a breach of one of the maxims of the Cooperative Principie will normally make the language user enforce an appropriate interpersonal principie. For example, the Maxim of Modesty ("minimize self-praise; maximize self-dispraisc") controls the exprcssion 1-Iow stupid of me! (which violates one of thc maxims of Quality). It is bascd on the knowledge that attributing a mistake to oneself rather than blaming so m con e el se is part of the conventions which regulate polite social contact. And this is interpersonal knowledge.

CAPACITY AND METHODOLOGY
The account given abo veis not complete but it pro vides certain important clues as to the relationship between competence and capacity, the former being related to forms of declara ti ve knowlcdge -a pool of passive information, ideational and interpersonalwhile capacity is operational know-how (interpersonal and textual) regulated by pragmatic forces (like the conversational aims of bcing polite, ironical, cooperative, clear, expressive, and so on). It is important for the language teacher, as well as for the syllabus designer, to be awarc of all these facts. Capacity may be not only a theoretical concept but also a reality which a good languagc methodology might exploit in at least two different ways. First, as Widdowson himsclf has suggested, sin ce an L2learner' s conceptual pattcrns are different from those in thc target language speakers' minds, and sorne alignment is needed, it might be useful to try to use his own mother tongue capacity in order for him to build thc necessary L2 stratcgies into his transitional competence. Second, since the relationship between pragmatic knowledge and grammatical form varíes with the language it might be uscful to help the language learner become aware of such facts in a systematic way. That woulcl cnhance his performance potential in the second language. For example, if we only tcach grammar we m ay explain that the verb must is a modal auxiliary which expresses obligation. Then, if we teach functions, we are likely to add the information that in certain situations must is u sed to express a certain speech act, like an invitation (eg. You must come and have dinner with us, in the same way as We would be pleased ifyou carne and had dinner with us or You will come and have dinner with us, won' t you?). But thcrc is still the further possibility that we teach ourstudentsalittle bit more: *We must come and havedinnerwithyoucould-depending on the context-be taken as a hilarious, ironical or impolite remark, or simply a breach of the Tact Maxim of the Politeness Principie. This may not be evident to L2 students unless we help them realizc sin ce thcy m ay not sharc thc same set of cultural conventions (that is the same ideational ancl interpcrsonal schemata) with native speakers of the target language.
Also, whcn trying to trace the origin of sorne lcarncrs' mistakes, language teachers should be aware that misusing textual principies may result in oddities which do not necessarily reflecta dcficiency in grammatical compctence. Take the following cases which reflect native-like competence but, in their context, are by no means good examples of languagc use: These two sentences violate thc maxims of the Clarity Principle 3 : Woman wants cleaning three days a week Wanted-manto take care of cow that does not smoke or drink These others result from the misapplication of the Economy Principie: Wanted-edible oil technologist 4 If the baby won 't drink the mil k it should be boilecP The following is grammatically correct, but also a violation of the Processibility Principie: That the demonstrators were not willing to give in to threats from government officials and abandon their strike ata moment in which they were receiving important and encouraging back-up from most of the parties of the Opposition and in which significant public support was being attracted by them is a fact.
To con elude, it m ay be pointed out that an L2learner' s requirements should extend beyond the ability to express content and to produce situationally adequate utterances; learning how to use the target language in a clcar, cffective way entails much more than developing 'compctencc' in thc traditional sense. That is why it may be hoped that an adequate understanding of the rclationship bctwecn compctence and capacity will be useful in helping the lcarncr to mcet these rcquircments.