Plàcia a Déu ‘May it Please God’: Constructions with the Verb Plaure ‘Please’ in Medieval Catalan

The aim of the present paper is to discuss some uses of subjunctive constructions formed with the verb plaure ‘please’ in medieval Catalan. These constructions were used as permission requesting directives, and later, as indirect directives. In a group of the medieval Catalan occurrences the speaker applies for God’s or another divine persons’ consent. This paper demonstrates that the construction (no) plàcia a Déu ‘may it (not) please God’ became a conventionalized phrase with a special role in interpersonal interactions, arguing that it followed a specific path of semantic change and underwent subjectification. As to the methodology of diachronic speech act analysis, the aim of the present paper is to argue for a comparative stance and for a qualitative contextual analysis in order to identify illocutionary values of utterances surviving in written form. The findings show that, instead of analyzing only the speech act verbs themselves, it is important to take into account further factors as well, e.g. the social context of the interaction, encyclopedic information, the content and “weight” of the request, the cultural context of the period, and the structure of the interaction. Moreover, it is worth considering the whole semantic and functional field, investigating together constructions that are suitable to perform illocutionary acts belonging to the same type.


Introduction
The study presented in this paper addresses an issue of diachronic speech act analysis (see Arnovick 1999;Dasher 2004 [2002]; Taavitsainen and Jucker 2007;Kohnen 2000Kohnen , 2007Jucker andTaavitsainen 2000, 2008). It aims to reveal some uses of medieval Catalan constructions based on subjunctive forms of the verb plaure 'please'. These constructions were used to perform directives, in later occurrences in an indirect way (Nagy C. forthcoming). There is a number of occurrences, hitherto undiscussed, where by using the plaure-construction the speaker applies for God's or another divine persons' consent. The aim of the present paper is to demonstrate that the construction (no) plàcia/plagués a Déu 'may/might it (not) please God' became a conventionalized phrase with a special role in interpersonal interactions, followed a specific path of semantic change, and underwent subjectification. A further aim of this paper is to find effective ways to conduct historical pragmatic research on speech acts. Methodological issues discussed here primarily concern the identification of the illocutionary force of written utterances.
The structure of the paper is as follows. After presenting the data and methodology of the study in Sect. 2, Sect. 3 deals with Searle's (1975) generalization about possible ways of performing indirect directives. In Sect. 4, after introducing some historical data, I discuss sporadic mentions provided in the previous literature concerning the use of plaure 'please' in medieval Catalan language. In Sect. 5 I turn to various uses of some medieval Catalan structures formed with the verb plaure 'please' (namely, plaure (subjunctive) ? indirect object ? que 'may it please sb that', and plaure (subjunctive) ? indirect object ? infinitive 'may it please sb to do sth'), while Sect. 6 discusses theoretical considerations about the illocutionary force of these utterances. In Sect. 7 I argue that one of the uses of these constructions, namely, (no) plàcia/plagués a Déu 'may it (not) please God' underwent subjectification. Section 8 addresses methodological issues of diachronic speech act analysis, with special emphasis on the role of speech act verbs. Finally, Sect. 9 is dedicated to the discussion of findings and conclusions of the study.

Data and Methodology
For the present study I used a corpus of medieval Catalan texts from the thirteenth to sixteenth centuries (cf. Table 1, for more details see Historical sources at the end of this paper). My aim was to compile a historical corpus that contains a body of texts of different genres, covers a substantial period, and comprises a high enough number of occurrences to analyze the phenomenon under study.
Methodology of the present study was aimed at meeting the challenges of identifying and analyzing speech act types in a historical corpus (cf. Kohnen 2000Kohnen , 2007Jucker and Taavitsainen 2000). I primarily adopted the form-tofunction perspective (cf. Jacobs and Jucker 1995), based on a search for subjunctive forms of the verb plaure 'please'. The occurrences were then qualitatively analyzed in their broad contexts (for a more detailed discussion of methodological issues, see Sect. 8).

Theoretical Background: Possible Ways of Performing Indirect Directives
Research on directives has revealed various linguistic means of indirect and polite ways of communication. Searle (1975: 64) claimed that when we perform directives, we usually seek to find indirect means because of requirements of politeness. Concerning possible ways of performing directives in an indirect way, Searle (1975: 72) made the following generalization, on the basis of his observations about English language use: However, Searle (1975: 72) later added that '' [t]hese are generalizations and not rules''. Some medieval Catalan constructions with imperative or subjunctive forms of the verbs voler 'want', pensar 'think' and plaure 'please', 2 and parallel constructions in medieval Spanish, were used to perform directives in an indirect way that is not included in Searle's generalization and has not so far been discussed in the literature: the speaker does not ask, nor does s/he state that the hearer wants to perform a certain action, but s/he asks her/him to want to do the action wished for or to have the pleasure of doing it. In previous papers (Nagy C. 2011, forthcoming) I have discussed some of these ways of performing indirect directives, relying on a corpus of medieval Catalan and Spanish texts from the twelfth-sixteenth centuries. They are interesting from a theoretical point of view, because they call attention to the fact that the generalization above can be complemented.
In the next section I introduce some historical data regarding these specific uses and discuss previous findings on the use of plaure 'please' in medieval Catalan.

Directives in Medieval Romance Languages and Use of the Catalan Verb Plaure 'Please'
Consider the following two examples where the speaker performs a directive by appealing to the hearer's willingness to do a certain action. 34 1 The abbreviation ''S'' refers to the speaker, ''H'' to the hearer, and ''A'' to the act in question. 2 Consider that all these constructions have an imperative meaning. Imperative forms exist in Catalan and Spanish for second person singular and plural, for other persons the subjunctive present forms are used. For negative orders the subjunctive forms preceded by no are used for all persons. 3 The abbreviations used in the glosses are given under Abbreviations at the beginning of this paper. The glosses are not intended to capture all morphological properties of Catalan and Spanish words but indicate the necessary ones for the present purposes. 4 Some historical texts in my corpus have an English translation available, I partly used these to translate historical examples. Translations sometimes could seem strange, because they tried to preserve the archaic style of these medieval texts. Moreover, given that they are literary translations, sometimes I have had to modify the English version for the linguistic purposes of the present discussion. I have not adopted the translation entirely when it does not reflect the grammatical form of constructions examined in the present paper. In these cases I rather provide a literal translation for convenience. For bibliographic data of the translations, see Historical sources at the end of the paper.
(1) (a) e dixeren:- The Catalan text fragment in (1a) from Muntaner's Chronicle, includes two constructions that refer to the hearer's wish in relation with a certain state of affairs: one with the verb plaure 'please', the other with voler 'want'. The speakers direct a request to the king: they ask for permission to embark on the journey, using the verbs plaure 'please' and voler 'want'. The constructions with these verbs occur together in subordination, however, they could also be used separately to perform directives.
The Spanish text fragment in (1b), from the famous medieval epic poem written in the Castilian language, Cantar de mio Cid, includes a directive performed in a similar way. The speaker would like the hearer to carry out an action, which is described in an infinitive, while the imperative formally belongs to the verb pensar 'think'. However, the speaker probably aims to get the hearer to actually do the action and not only to think about it. Imperative forms of the verbs meaning 'think' and 'want' were used to perform indirect directives in a similar fashion in both medieval Catalan and Spanish, specifically, the verb querer 'want' in medieval Spanish had uses similar to (1a), while the medieval Catalan verb pensar 'think' had uses similar to (1b). The verbs 'want', 'think' and 'please' are worth considering together, because they all are linguistic means appropriate to perform directives in medieval Catalan. In the present paper I concentrate on the medieval Catalan constructions involving the verb plaure 'please' (cf. 1(a)).
Previous literature does not mention the Catalan construction involving the subjunctive forms of the verb plaure 'please'. Although this Catalan verb already was in use at the end of the sixteenth century, by the end of the nineteenth century it was displaced by the verb agradar 'please' (Alturo and Chodorowska-Pilch 2009: 24). However, it has survived in modern Catalan in some stereotyped expressions, for instance, in si a Déu plau 'if God pleases', and si us plau, sisplau 'please'. The grammaticalization of this latter expression is directly related to the issue discussed in this paper. Sisplau 'please' originates from a conditional construction (si ? indirect object ? plau 'if it pleases sb'), but later adopts a discourse function. This process has been investigated by Alturo and Chodorowska-Pilch (2009). The authors mention similar semantic change processes in other languages as well, such as the English please \ if you please, the French s'il vous plaît or s'il te plaît, and the Occitan si a vos platz. They relate the beginnings of this grammaticalization process 5 to contexts where the speaker performs a directive speech act. By using this conditional clause making reference to the hearer's willingness, the speaker politely admits that the hearer has the right not to take into account the speaker's wish. The original conditional construction changed into a stereotyped politeness formula, which is still used in current Catalan (sisplau 'please'). In contrast, Alturo and Chodorowska-Pilch (2009) do not deal with the cognate form, which is the topic of the present paper: when the speaker uses subjunctive forms of the verb plaure 'please', asking the addressee to have the pleasure of doing a certain action. In a previous study (Nagy C. forthcoming) I have already discussed some uses of this medieval Catalan construction. Therefore, in the next section I just summarize these briefly, whereas I will consider other uses in more detail.

Medieval Catalan Constructions with Subjunctive Forms of the Verb
Plaure 'Please' In a previous paper (Nagy C. forthcoming) I have discussed occurrences of constructions with plaure in a corpus of medieval Catalan texts from the thirteenth to the sixteenth centuries and classified them into two main groups: in 44 examples the speaker(s) aim(s) to win the approval of God, Jesus Christ, or the Virgin Mary, while in further 58 that of another person. The reason for dividing the occurrences according to these context types lies in the fact that, according to the findings of my historical study, illocutionary forces of the speech acts performed by these utterances can be different, and they underwent different paths of semantic change. The present study aims to emphasize the role of constructions in semantic change: the same lexeme can follow different paths of semantic change in different constructions (for the role of constructions in grammaticalization, see Traugott 2003).
Constructions 6 with the verb plaure 'please' making reference to the addressee's willingness occur often in the corpus studied when the speaker performs a directive speech act. From a grammatical point of view, the issue concerns two structures, although their meanings and functions seem to be the same: (i) the 3Sg subjunctive form of plaure (plàcia, plagués) 7 ? indirect object (Experiencer) ? que 'may it please sb that', (ii) the 3Sg subjunctive form of plaure (plàcia, plagués) ? indirect object (Experiencer) ? infinitive 'may it please sb to do sth'. The grammatical structure in (ii), that is, an infinitival clause can be used, when the person asked to be pleased (i.e. the experiencer) and the agent of the action requested coincide. The distribution of occurrences is shown in Table 2.

Directive Speech Acts Performed by Using the Plaure-Construction
In a previous study (Nagy C. forthcoming), I have already discussed the speech acts performed by using a construction involving plaure 'please' with a human Experiencer, therefore, I mention them here only for the sake of a brief summary. Usually we find these occurrences in directive contexts when the addressee is a person at the top of the social hierarchy (in the corpus of the present study: the king, the pope, the Infante, etc.). Therefore, we can suppose that it was a very respectful 6 Throughout the present paper I use the word construction in a pre-theoretical way. However, I aim to highlight that lexical items take new functions in constructions, i.e. in certain specific morphosyntactic and pragmatic contexts. Therefore, the present analysis is not at all incompatible with Construction Grammar approaches (e.g. in Traugott 2008;Hoffmann and Trousdale 2013). Constructions with plaure (i) and (ii) could be conceived of as micro-constructions pertaining to the same meso-construction, as defined in Traugott (2008: 236). 7 Plàcia is present tense subjunctive, while plagués past tense subjunctive form of the verb plaure.
form. I have distinguished between two main functions and claimed that they developed one after the other: (i) Type 1. Permission requesting directives: the speaker wishes to do a certain action, but s/he needs permission from the addressee and asks for it. This use can be related to the fact that in the state of the language under study the verb plaure 'please' was also used to give permission, approval or consent in behabitive utterances (cf. Austin 1962: 78-85). Therefore, the permission requesting directives are not indirect speech acts: the speaker asks the hearer to perform the speech act of approval, that is, to say em plau 'it pleases me'. (ii) Type 2. Indirect directives: occurrences found in contexts where the action wished for is the hearer's action cannot be interpreted as permission requesting directives. We have to assume that these occurrences with plaure 'please' convey a different meaning in these contexts: the speaker aims to get the hearer to do the wished-for action, although the literal meaning only refers to be pleased. In this sense, this second type is an indirect directive.
In the following let us consider examples of both types.

Permission Requesting Directives
Permission requesting directives are requests aimed at getting the addressee to permit that the speaker or a third party do a certain action. As an illustration, consider the following text fragment in (2). Muntaner relates in his Chronicle that King Charles decided to arrange a marriage between his son Philip and the Princess of Morea. A short fragment of this example has already been discussed in Nagy C. (forthcoming). However, the present example (2) contains more occurrences of the verb plaure, followed by a more thorough analysis 9 En was prefixed to the proper names of nobles in Catalan, it is equivalent to Don in Spanish From the fragment in (2) we get to know that the king asks for his son's consent to the marriage, using the expression si ell ho volia 'if he wished it'. Another matter is whether the son has an effective decision making right, anyway, he has to formally consent to the marriage. Consider that the Infante Philip performs this approval with the verb plaure (li plaïa molt 'it pleased him well', i.e. he agreed). This occurrence shows that the verb plaure was conventionally used to express approval in the language use of the period (see Nagy C. forthcoming). 10 Philip, King Charles's son agrees to marry the Princess, but on one stipulation. The occurrence of the verb plaure here (que plagués a ell que li donàs un do 'may it please him that he grant him a favor') is an indirect directive, belonging to Type 2 (see next Sect. 5.1.2). The condition Philip refers to is a request concerning a state of affairs, for which permission from his father is needed: namely, that his friend, the son of the count of Aria could take to wife the princess's sister, and both marriages would be celebrated together. This request belongs to Type 1: the Infante performs the permission requesting directive by using the construction with plaure 'please' (us plàcia que 'may it please you that'). Two speech act verbs appear in the fragment in a performative use: prec 'I ask/pray' and us deman 'I ask/request you'. The addressee's approval as a prerequisite for a state of affairs or action can occur independently as a clause (the politeness expression sisplau 'please' grammaticalized from this usage, see above in Sect. 4). Among these occurrences we can find cases where the emphasis is transferred from the literal meaning of plaure 'please' to the action itself, as in the case of the example in (3) 'And when the Catalans saw themselves thus settled in the Duchy of Athens and lords of the country, they all sent their messengers to Sicily, to the Lord King, to say that, if it pleased him to send them one of his sons, they would take the oath to him as their lord and would deliver up to him all the fortresses they possessed.' Consider that in the above context the promise of taking the oath concerns the case of performing the action requested, and not that of being pleased. We could also translate it as 'if he sent/agreed to send them one of his sons'.
Indirect Directives: Request for the Addressee to do a Certain Action Type 2 includes utterances in that the activity referred to is the hearer's own action, so, by uttering them, the speaker does not ask for permission but rather tries to get the hearer to perform a certain action. In the text fragment above in (2), one of the occurrences of plaure functions exactly like this: the Infante Philip asks his father with the following words, que plagués a ell que li donàs un do 'may it please him that he grant him a favor'. Also in the following example in (4), the action in question is the addressee's own action. In these utterances, it is reasonable to suppose that the speaker would like to get the hearer not only to be pleased, but also to perform the action in question (e.g. in 4, not to delay), in other words, they are indirect requests and we can assume that this indirectness is part of a politeness strategy. For more details on the use of these constructions, see Nagy C. (forthcoming). The following section deals with those occurrences in more detail where the experiencer of the verb plaure 'please' is God and that have not been discussed in Nagy C. (forthcoming).

Speech acts with the Plaure-Construction Addressed to God
In some occurrences the Experiencer of the verb plaure 'please' is a divine person: the speaker aims to win the approval of God, Jesus Christ, or the Virgin Mary. These utterances may appear at first sight to be supplications to God. However, they are ''public prayers'', which seem to fulfill a specific role in the verbal interaction between human participants, too. In order to analyze these occurrences we have to take into consideration the theocentric worldview of the Middle Ages, according to which God is the one who governs and controls the world. Only what He allows to happen will happen. This worldview manifests itself also in the language use and interpersonal verbal interactions of the period, not only in some linguistic forms, but also in the organization of interactions and argumentation. Reference to God's will is possible with constructions formed with the verbs plaure 'please' and voler 'want' as well. In my corpus many examples like this can be found, such as plach a Déu/Déus volch 'it pleased God/God wished it' or no plach a Déu/no ho volch nostre Seyor 'it did not please God/Our Lord did not wish it', and similar expressions that are conventionally used to express that a state of affairs or an event occurred or did not occur, respectively. The example in (5) shows an occurrence with the verb plaure 'please' and that one in (6)  be.SBJV.3SG manifest.PART '-Muntaner, get up and think of making a book of the great marvels that thou hast seen, which God has worked in the wars in which thou hast been. For it pleases God that by thee he should be manifested.' In the example in (8) we can see another directive, this time performed with a construction involving the verb pensar 'think', but this fragment also exemplifies how God's will motivates the actions of humans. The old men enumerates four reasons, one of them being the following: tell.INF 'The third reason is that it pleases God that thou shouldst recount these adventures and marvels, for there is not another alive to-day in the world who could relate them as truly.' God's will appears everywhere when relating the events, and also in the verbal interactions, as in (10) '-Lord, if God wishes, the army of Sicily arrives soon, and together with them we will make much gain on sea and land, if it pleases God.' (5)-(10) are linguistic manifestations of the faith that God's will and consent are prerequisites for anything to happen. These linguistic forms occur frequently in verbal interactions, so it is not surprising that some of them underwent specific semantic changes. In Sect. 5.1 above I have noted how the imperative plaure-construction was used to perform directives toward human addressees in the medieval period. But what happens in the case of occurrences that refer to God's pleasure? In this section I discuss this type of use of the plaureconstruction. The distribution of relevant occurrences in the historical texts is shown in Table 3. The religious worldview and frequent references to God's will are present also in the political discourse of the period. The records of medieval Catalan parliamentary speeches and also relevant parts of chronicles reflect correctly how these meetings took place and what the structure of the verbal interaction was like during them. Consider the example in (11), from James I's Chronicle. The context is the following: the king organizes a meeting of the Corts (the Catalan Parliament), assembling the archbishop, the bishops, and the nobles, and then he opens the parliament, speaking in the majestic plural, with the following words: In the continuation, the king refers to God in one way or another in six of his seven utterances. Then, Berenguer Girart delivers a speech, in a reply to the king's opening speech, on behalf of the citizens: 'The Lord, who is your Lord and ours, has put into your will those good things that you just said to us; and may it please Him that our reply be such that you may fulfill your wish to God's honor and your own.'

All the above examples in
The above occurrences in (11) God's will played in the structure of the argumentation and verbal interaction in the political discourse of the period. At first sight the above utterances with subjunctive forms of the verb plaure seem to be examples of directives toward God: the speaker expects God's help in achieving a wished-for state of affairs. However, it is also true that s/he expresses this wish verbally; consequently, s/he would like to make other people be aware of it as well. In other words, the speaker performs a speech act toward God and toward the other participants of the speech event at the same time.
In addition, in most of these contexts the speakers do not direct their words directly at God. Therefore, the interpretation of these utterances is not straightforward. Are they directed at God or at the other participants of the speech event? How can we identify the illocutionary force of these utterances? These questions are addressed in the following section.
On the Illocutionary Force of Utterances with God as Experiencer of Plaure From a present-day perspective it is hard to recognize the illocutionary intentions behind these utterances with God as experiencer object of the verb plaure 'please'. Anyway, these uses of the plaure-construction can be related to two speech act types: directives and expressives. If speakers express feelings about themselves or the world, the speech act is called expressive. If the psychological state expressed is a wish, the speech act in question is only one step away from directives, because the sincerity condition of directives is exactly the speaker's wish that the hearer do a certain act. The essential condition of directives is to get the hearer to act according to the propositional content, in other words, the speaker does not only have a wish but also wants that the hearer play a role in the realization of the state of affairs wished for. The question is whether the essential condition is fulfilled or not. Therefore, if the speaker only expresses a wish but does not attempt to get the hearer to do the act wished for, the speech act is an expressive. In contrast, if the speaker expresses a wish and, by performing this act, attempts to get the hearer to do the act wished for, it is a directive.
In what follows, let us use the analysis of the fragment in (12) as an illustration of the problem of identifying the illocutionary force of these utterances. How should we interpret the speech act performed by the speaker in (12)? First of all, it is important to note that it is impossible to exclude subjectivity when judging the illocutionary force of utterances. However, on the basis of our encyclopedic knowledge about the religious worldview of the period, we can assume that the first utterance of the fragment in (12) is sincere, and the speaker believes in God paving the way for the people by influencing their intentions. As far as the second part is concerned, we cannot disregard two facts. On the one hand, Berenguer Girart addresses the utterance to the king. On the other hand, although probably it is God from whom he expects the realization of a state of affairs, he expresses this wish verbally, that is, he makes those present aware of it. Therefore, the utterance should be interpreted as part of the verbal interaction between human participants. In a first approach, this utterance can be interpreted in two different ways: (i) as a directive, in which sense the speaker asks for God's consent for the things to go right, or (ii) as an expressive, in which sense the speaker merely expresses a wish. According to the interpretation in (i), the utterance belongs to the type of permission requesting directives. I have already characterized this use above, in Sect. 5.1.1, as directed to human addressees (when the speaker asks for a superior's consent in relation to an act or event). This use seems to be possible also in those cases when God is the experiencer of plaure, as the one in the highest place of the social hierarchy. However, the interpretation in (ii), as an expressive is also plausible. Next, I present arguments in favor of the analysis of these occurrences as expressives.
One of the arguments can be the grammatical form: the indirect object of these occurrences is in the 3rd person singular, that is, they are not addressed directly to God. The other argument is an indication of a contemporary speaker's linguistic intuition. In the example in (13) (14) is also worth considering because of its grammatical properties. It is uttered by the Saracen admiral, ''who was a wise seaman and had been in many feats of arms and had proof of what the Catalans are'', to his people before a battle: In the case of the utterance in (14), the speaker asks for God's consent concerning a situation contrasted with the present one. However, the subsequent utterance makes it clear that he has expressed only a wish, and he actually does not expect from God the realization of the situation wished for but accepts the present state of affairs. Therefore, the interpretation as an expressive speech act is plausible here. Tense and mood of the verb are other reasons for this interpretation in this context. The form plagués is the past subjunctive form of the verb plaure 'please', which also appears in conditional and optative sentences. From the 102 occurrences found in the corpus, 28 examples of plaure are conjugated in past subjunctive (cf. Table 2). However, in 24 cases the past tense is motivated by the past narrative context (cf. the past form plagués in 2). The remaining 4 examples appear in utterances with God as the experiencer of plaure 'please'. Their contexts are similar to the one in (14), and we can assume that the past tense form of plaure in them indicates the interpretation as expressives. A further argument for the hypothesis that we should not analyze these utterances exclusively as directives toward God is that they tend to appear at particular points of the discourse and play a particular role in its organization. For instance, they are used very often at the end of a stretch of discourse. We could characterize this function as a kind of ''concluding evaluation'', maybe directed at God and the discourse participants at the same time. In the example in (15), the king, after having discussed strategies of war with the gentlemen, finishes the interaction with these words: The king's utterance can be regarded as a closure of the interaction, by which he expresses a wish, and in this sense it is an expressive. But how do we know whether he expects somebody to play a role in the realization of this wish or not? And who does he expect to do something? Is it God? Or the other participants in the speech situation? Does he want to suggest to the listeners that they have to act the way they agreed to act before? I argue that these functions of the utterance are hard to distinguish from each other in the actual language use, and the illocutionary force of these utterances cannot be identified in a straightforward manner. We can find the construction with voler 'want' in similar contexts, conveying similar meanings, as in the fragment of a letter in (16) We can see a more negative evaluation in the following interaction in (17). In this context, the king sends a messenger to the Pope. The messenger has to tell the Pope everything the Lord King ordered him to say: he asks for assistance and money for a crusade, but the Pope refuses to give these. The messenger, indignant, utters the following words immediately before leaving: With the utterance in (17), the speaker expresses a wish and asks for God's pleasure concerning its realization. However, if we take into account the whole discourse, we should interpret the utterance in its context as one directed at the Pope, as the second person pronouns (vostra 'your', us 'to you') also suggest. However, it is not possible to judge with certainty whether the noble only expresses his psychological state, his anger (expressive), he performs a threat, or he wants to provoke -in an indirect way -a certain reaction as well (directive). The function of closing the interaction is also present in the parliamentary discourses, as in the closure of Peter III's parliamentary speech on 9 March 1370 in Tarragona. The king finishes his speech with a quote from the Bible, which says that if the owner of the house had known at what time of night the thief was coming, he would have kept watch. The king encourages the participants to be similarly vigilant in the affairs of the country: The utterance in (18) illustrates that the construction was probably employed as a rhetorical device as well: we should have no doubt that the wish expressed toward God in (18)  Amén. (Parl 72,12) amen 'and may it please God, Our Lord, that he give us as much grace that we can reign in a way that would serve his glory and your benefit, and so that we merit his grace here and his glory over there. Amen.' This type of closure is very common in the parliamentary orations: the king explains that he would like to reign according to God's wishes and for the people's benefit at the same time. Based on its frequency it can be assumed that it has to be interpreted as a conventional way of finishing a speech, a routine formula, characteristic of the genre. However, it is impossible to judge the speakers' intentions with certainty. By this utterance the speaker asks for God's pleasure in connection with a certain state of affairs. At the same time, s/he directs the utterance at the participants of the speech event, and it can be interpreted as a closure of the speech addressed to them. It is worth adding that the duality documented in the case of utterances with the verb plaure addressed to a human addressee (cf. Type 1 and 2 discussed in Sect. 5.1) can also be observed in utterances with God as the experiencer of 'please'. The action wished for can be (i) the speaker's act or some independent matter or (ii) God's own act. While in the former case the speaker seems to try to achieve God's consent, in the latter case s/he tries to get God to do something, using the verb plaure 'please' as a linguistic means of indirectness (Table 4).
As a summary, we can analyze these linguistic forms as directed toward God and toward human participants at the same time. Such an analysis can be provided in a theoretical framework that presupposes the presence of lateral speech acts in linguistic interactions. In their paper on speech acts, Clark and Carlson (1982)  introduce the concept of informative speech acts. They argue that in the context of verbal interaction with multiple participants we have to distinguish between addressees and other participants. The authors claim that the speaker performs two types of illocutionary act with each utterance: a ''traditional'' kind of speech act, directed at the addressees, and, in addition, an informative speech act, directed at all the participants in the conversation. With the informative act the speaker informs the addressee and the other participants of the illocutionary act that s/he is simultaneously performing toward the addressee(s). Moreover, Clark and Carlson (1982: 333) claim that ''[a]ll addressee-directed illocutionary acts are performed by means of informatives''. Clark and Carlson (1982: 336) distinguish between two distinct types of indirect illocutionary acts, called linear and lateral indirect illocutionary acts, respectively. In linear indirect illocutionary acts the direct and indirect addressees are the same, while in lateral indirect illocutionary acts, which can occur in conversations involving more than two people, they are different. It is not easy to reveal the communicative intentions of speakers of the medieval period (or of present day speakers, for that matter). However, the occurrences presented above, associable with two speech act types, could be analyzed in this theoretical framework according to the following possibilities: (i) analysis of the utterances as directives: we suppose that the speaker asks for God's or the participants' activity in the realization of a state of affairs (a) this is a request performed by means of an informative act, the addressee is God (b) linear indirectness: a directive performed by means of an expressive, the addressee can be God or the other human participant(s) (c) lateral indirectness: the directive is intended to be directed at the human participant(s) in the conversation, but the addressee is God (ii) analysis of the utterances as expressives: we suppose that the speaker purely expresses a wish (a) the addressee can be God, the other human participants are informed by an informative speech act of the performance of the expressive (b) the addressees are the human participants of the speech situation: since the speaker expresses his/her wish directed at God verbally, s/he probably intends the speech act to be understood by the others.
In summary, the main question is who the addressee is (God or the people listening to the utterance) and who is simply a participant of the speech event, and whether there is-either lateral or linear-indirectness (cf. Clark and Carlson 1982: 336). Finally, it is worth noticing that although in the majority of contexts, the speaker does not address God directly, that is, the indirect object is in 3rd person singular, there are seven occurrences with a 2nd person singular indirect object, directly addressing God. These forms appear in only one text (Muntaner's Chronicle), so perhaps they can be due to the chronicler's personal narrative style, who may have aimed to reach a lifelike effect in this way. Nevertheless, these uses show that we cannot analyze these utterances in every case as expressives, instead, we have to take into account the interpretation as directives addressed directly to God. Consider the following example: How should we analyze the body language described in (21)? Does the speaker assign the role of addressee with these gestures? Or does he only want to add emphasis to the content of the utterance and emphasize the sincerity of his request? Anyway, these uses suggest that we should take into account the analysis as a directive speech act addressing God, too. Be they directives or expressives, all of the utterances presented above tend to appear at certain points of the discourse and seem to play a special role in its structure. The contexts of occurrences with 2nd and 3rd person singular pronouns are similar: they are all used when the speaker intends to express his/her own evaluation in relation to an issue. They all seem to convey the speaker's attitude, in this sense the request for God's help or anger seems to play a special role in the discourse. To conclude, these findings suggest that the construction followed a particular path of semantic change: it became a conventionalized phrase and underwent subjectification. In the next section I turn to elaborate on this idea.

Subjectification
In Sect. 6 I have discussed two groups of utterances with the plaure-construction. First, this construction is used in contexts where it seems to express the speaker's desire, as in optative sentences (cf. examples in 13-14). Second, the plaureconstruction appears at the closure of an interaction (cf. examples in 15-19), conveying the speaker's own subjective evaluation on a certain issue. These functions are close to the uses presented in this section, when the speaker expresses subjective feelings in relation to a content present in the discourse context. The process whereby a linguistic expression acquires meanings that convey the speaker's subjective attitude or viewpoint is called subjectification. 11 I use this concept in the Traugottian sense (cf. Traugott 1989Traugott , 2010, which connects subjectivity to the meaning of a morpheme, word, phrase or construction. 12 Another significant view of subjectification is provided in Langacker (2006). The following two collections of papers represent these two different points of view on subjectification: Athanasiadou et al. (2006) represents the Langackerian view, while Cuyckens et al. (2010a) the Traugottian view. After her early research on subjectivity (Traugott 1982), in her later work Traugott (1989: 34-35) explains the process in more details, mentioning three main related semantic change tendencies. For the focus of the present paper the third one is relevant: ''Tendency III: Meanings tend to become increasingly based in the speaker's subjective belief state/attitude toward the proposition'' (Traugott 1989: 35). There are some contexts which suggest that the meaning of the plaure-construction has changed according to this tendency, and it has undergone subjectification. It has begun to convey the speaker's subjective attitude: affirmative utterances with the construction express the speaker's wish and positive attitude, while the negative version of the construction is used to express indignation and refusal concerning a proposition known from the previous context or mentioned subsequently in the discourse. We find one of these examples when Desclot tells the legend of the good count of Barcelona: he is going to defend the empress of Germany, who is falsely accused of adultery, in legal combat. The knights are indignant at his intention to go with only one companion and offer five hundred or a thousand knights to go with him. But the count says the following, expressing his refusal: A similar shade of meaning can also be conveyed with the verb voler 'want' in a similar construction referring to God's will.
These two examples above cannot be analyzed as directives: both cases concern the speaker's own actions, on which the speakers can decide freely. Consequently, the speakers in (22) and (23)  find.PART 'And then the admiral did a thing which should be counted to him more for madness than for sense; he said, might God not want that he attacked them in their sleep, rather did he wish the trumpets and nakers to be sounded in the galleys to wake them up, and he would let them prepare. He did not wish that any man could say to him that he would not have defeated them if he had not found them asleep.' In the following context, the speaker also expresses indignation concerning a proposition previously introduced into the context. The king asks whether those people wish much evil to the king of Aragon, but they answer in the following way: In the case of the utterance in (24), the interpretation that the speakers would direct a request at God can be excluded. This utterance does not refer to a future action but to a wish of the speakers themselves, and they are aware of whether it is true of them. Finally, consider the text fragment in (25). In contrast with the negative utterances in (22)-(24), it contains the construction in an affirmative context, but also seems to convey a subjective meaning: In (25), the plaure-construction occurs in a subordinate clause, in an expansion of the adverb before (en tal bona ventura que 'with such good fortune'), therefore, it is implausible that the speech act in question would be a directive directed at God. This fragment is also interesting because the chronicler expresses his own opinion, addressing the speech act to a later interpreter of his text. The function of the 'And the noble En Roger threw himself on the ground and kissed the feet of the Lord King of Aragon and then his hands and took the baton with such good fortune that may it please God that all the officials to whom the Lord King will entrust his offices in future will administer them as well as did the said noble.' construction seems to be to express Muntaner's subjective point of view and to add a certain emphasis. All these examples suggest that this construction referring to God's wish or pleasure took a specific path of semantic change and became a conventional way of expressing subjective attitudes, in other words, it underwent subjectification.
In summary we can say that the plaure-construction tends to appear at certain points of discourse and seems to play a particular role in its structure and in the argumentation. As regards the illocutionary force, we can analyze it either as a directive or as an expressive, but this is not straightforward in every context. The construction, mainly in its negated form, underwent subjectification, that is, it acquired a meaning that conveyed the speaker's subjective attitude or viewpoint at certain points of the interaction.

Methodological and Theoretical Issues
Diachronic speech act analysis raises several methodological questions. Some of them are discussed in Taavitsainen and Jucker (2007), who mention two methods of investigating speech acts from a historical point of view. The authors base their research on searches with speech act verb lists when investigating speech acts of verbal aggression in earlier stages of the English language. They examine the distribution of speech act verbs previously compiled with the help of dictionaries on a large corpus. This method does not guarantee the finding of all examples of verbal aggression in the corpus. According to Taavitsainen and Jucker (2007: 108), speech act verbs ''do not give us any direct access to the speech acts that they name, but they provide an interesting ethnographic view of how a speech community perceives specific speech acts and which ones are important enough to be labelled with a speech act verb which the speakers use to talk about the speech act and-in some cases-even perform the speech act in question''. They classify the verbs into semantic fields and investigate neighboring speech acts together in a pragmatic space. Another method is applied by Kohnen (2007), who aims to find all realizations of a speech act type, including the so called hidden manifestations (see Kohnen 2007), which is possible only by manual search.
In the present study, speech act verbs play a different role. My research also goes in part from form to function (cf. Jucker 1994;Jacobs and Jucker 1995), although I do not search for speech act verbs but for occurrences of constructions with the verbs 'want', 'think' and 'please'. Speech act verbs serve only as a complementary source of data when I try to identify the functions of these constructions. Therefore, I do not work with a predefined set of speech act verbs. This method presupposes manual search. One of the reasons why a manual search is necessary is that we cannot foreknow the inventory of speech act verbs in the medieval period. The other reason is that speech act verbs can occur in the very broad context of the construction, which in some cases covers more than one paragraph of text.
How can we reveal the illocutionary force of utterances, especially in historical investigations? One could assume that verbs describing speech acts in the utterances in question or in their contexts can be our primary source of information when judging the illocutionary force. But the issue is not that simple. Already Searle (1969: 70) has claimed that [b]oth because there are several different dimensions of illocutionary force, and because the same utterance act may be performed with a variety of different intentions, it is important to realize that one and the same utterance may constitute the performance of several different illocutionary acts. There may be several different non-synonymous illocutionary verbs that correctly characterize the utterance.
Although speech act verbs-occurring in descriptive or performative uses-can be reference points, in some cases they can be misleading. Not even in the case of performative uses can we assume that the speech act verb denominates exactly the speech act performed. As regards the descriptive use, this is also not a reliable source: it reveals only the intuition of the person who reports the performance of the speech act. According to Ervin-Tripp (1976: 28), [t]he verbs used in indirect speech are a source of information about native speakers' categories in the ethnography of speaking. The English verbs used in reporting are not necessarily the best analytic categories for classifying speech events, though they are a plausible first hypothesis.
In historical texts it also occurs that the narrator denominates the interaction by using a speech act verb in the description, which reveals how s/he interprets the utterance in question. Naturally, speech act verbs cannot occur only in narrative parts but also in text fragments describing verbal interaction. In addition, in most of the cases the quoted words are introduced by verbs that refer only to the locutionary act ('said', 'answered', 'told', etc.).
In the case of descriptive uses, the question arises whether the utterance described as an indirect quote corresponds word by word to the original utterance. We have to take into account that we usually do not remember the exact utterances, but the content of the interaction as a whole. In the description of interactions the narrator's interpretation plays an important role, in the formation of which s/he usually takes into account several pieces of information (e.g. the context of the utterance, formal markers, etc.). A usable theoretical framework is provided in van Dijk's (1977) paper on the macro-structures of discourses. The author relates the notion of illocutionary force not only to a single utterance, but he believes that it is applicable to larger units of discourse. In a global sense we can perform a speech act by uttering more than one sentence, which he calls macro-speech-act. He argues that discourse processing is complicated and we should assume a macro level in it, since we do not remember the exact words uttered, but organize information into structures of a manageable size. Functionally related speech acts are structured into larger units that we interpret as a single entity. Sentences and speech acts are ordered into macro-structures, in which the construction under study is only one of the linguistic means used to perform the directive.
If we aim to identify which constructions were used to perform directives in a certain language period, it is worth comparing the linguistic devices found in descriptive versus performative uses. For example, we can describe the performance of a speech act with the words She asked me to carry her bag, but it would be strange (in most situations) if we asked somebody in this way: Carry my bag! In some contexts we can perform a request by uttering a single sentence (Carry my bag, please), but in most cases we utter more sentences due to politeness considerations (cf. Blum-Kulka et al. 1989). Therefore, when we try to reconstruct the performance of a speech act relying on reports containing speech act verbs in descriptive use, we complete the linguistic form given in the description with information from our encyclopedic knowledge. For instance, we interpret a description like asked that 'he performed the request in a way appropriate/expected in that context', and we reconstruct a plausible way of performing the act in question relying on our pragmatic competence. According to that, in the case of a description like He asked me to do the shopping, the original utterance can be various, e.g.: Would you mind if I asked you to do the shopping? or Could you do the shopping, please? Moreover, descriptions also can be various. We can describe appropriately these speech acts also with the following words: He asked me whether I could do the shopping or He kindly asked me to do the shopping, etc. Anyway, we can presuppose a kind of pragmatic correspondence between the original utterance and its description: if the original was a directive, than the forms used in the description should be appropriate to perform a directive. We can also suppose that if some extra words appear in comparison with the propositional content of the request (in our case the expression could), it would have been used in the original context as well. In this way, although an exact formal reconstruction of the original utterance is impossible, we can have conclusions concerning the linguistic means usable to perform directives on the basis of the form of descriptive parts as well.

Speech Act Verbs in the Context of the Plaure-Construction
In this section I discuss how research into speech act verbs and similar linguistic forms 13 helps us to identify the illocutionary force of utterances with the plaureconstruction. They occur both in descriptive and performative uses. We can find the following verbs:
As I have mentioned above, there is no one-to-one correspondence between speech act verbs and illocutionary values. These speech act verbs only reveal the illocutionary act type of directives, the neutral verb 'ask' is the most frequent. However, we do not know whether the utterance is a request, a proposal, a demand or something else. In order to judge the degree of politeness of these forms in their particular contexts and their function, we have to take into account the broader context of occurrences: not only the co-text, but-among other things-the propositional content of the utterance, encyclopedic information concerning the ''weight'' of the request, the genre of texts, and the relationship between speaker and addressee. As for the latter one, addressees in early contexts almost exclusively belong to higher levels of social hierarchy: king, the Pope, Infante, admiral, prince, etc. In later contexts, other addressees are possible as well, and the use of the construction seems to be less polite: the speech act seems to be more pressing, and the content of the request can also be different. However, in order to use these pieces of information, we have to take into account the use of other forms of performing a directive and compare all of them (for instance, the use of the construction with the verb pensar 'think'-cf. the example in (8)seems to be a less polite strategy). Directive utterances can be interpreted in their relation to the others, taking into account the distribution of the different ways of performing directives across context types. Also, the speech act verbs in the contexts of various constructions used for performing directives have to be compared.
It is not surprising that the neutral verb pregar 'ask' is the most frequent, revealing the illocutionary act type (directive). The verb suplicar 'entreat', which expresses deference and refers to asymmetrical social relations, also occurs in high numbers. On the basis of this distribution we might assume that the construction was rather polite. However, there are also verbs that contradict this presupposition. For example, there is an occurrence of the verb requerir 'require': however, if we examine the context, we see that it is the Infante who is ''required'' to take the kingdom after the death of his father. Therefore, if we take into account the participants' interests, this use does not seem impolite anymore. Kohnen (2007: 151, 157) also emphasizes that when judging which speech act is performed by an utterance we have to consider also the predicates and the broader context. This example illustrates that several factors (the content and weight of the request, the interests of speaker and hearer, etc.) play an important role when judging the illocutionary force of utterances. The above analysis has shown that this is true even when in the context of the construction there is a speech act verb. On the basis of these considerations it is at least questionable whether we can gain a realistic picture of illocutionary force on the basis of only a quantitative analysis of the distribution of speech act verbs (cf. Taavitsainen and Jucker 2007).
We can also observe that not every verb can be used performatively to the same extent: the more neutral verbs, used to describe various illocutionary acts, occur more frequently in a performative use, while more specific verbs tend to occur in a coordinated way in specific genres, as part of conventional formulae. For instance, the verb suplicar 'entreat' has 5 occurrences, two of them in a coordinated phrase, supposedly used to add emphasis. This verb of a humble tone, however, occurs exclusively in the parliamentary speeches (namely, in speeches delivered between 1410 and 1421), thus, it can be considered an institutional formula characteristic of that period. Its use seems to mitigate the rough, crude tone of the verbs exortar 'exhort' and requerir 'require' at the beginning of the coordination and to make them more polite. However, the presence of these latter verbs may also suggest that the use of the plaure-construction perhaps could not be regarded as really polite in this period. Its contexts of use could have been extended, but to confirm this suggestion we have to check also other circumstances, e.g. the content of the request. In the parliamentary speeches containing these occurrences requests are always performed by a person of authority, who has the power to require (e.g. the king). Therefore, we can suppose that the person of authority aims to mitigate the face threatening directive by using this polite form. It should be noted that we have to differentiate between the politeness of performing a request (of a speech act) and the politeness of a linguistic form. Sometimes the speaker seems to aim to reduce the face threat by using polite formulae when they perform inherently face threatening speech acts. Certainly, even by using them, a request is still a request, and a demand is still a demand, and language users clearly perceive this.
Except for the parliamentary speeches, only Muntaner's Chronicle includes an occurrence of the verb requerir 'require' in the context of the plaure-construction. It is interesting that the context of this discourse (presented in Sect. 5.1.2 as 4 and repeated here for convenience as 26) is similar: the king is asking for material assistance for his campaign (the aim of parliamentary speeches often is the same). For this reason, he sends a messenger to the Pope, who has to replicate exactly those words that the king demanded that he say: to greet the Pope and the cardinals, to beg him to assemble the consistory, because he wished to say something on behalf of the king. Then he has to greet them again and say what the king commanded him to say. And at this point of the interaction we can see a very convincing argumentation, containing the expression lo dit senyor rei vos requer de part de Déu 'the said Lord King requires you in the name of God', finally the messenger ends the talk with the following words: Perhaps the aim of using this construction is to decrease the face threat of this imperative content. Anyway, there are only a few contexts as crude as this, the construction instead tends to appear in situations that require a high degree of politeness. Taavitsainen and Jucker (2007: 114-116) argue that we have to consider the original context of the utterance, and also the genre of the text in which the utterance is evoked. The authors emphasize that there are two levels of considering the context: taking into account (i) the context formed by previous discourse, and (ii) the socio-cultural characteristics of the communicative event. In order to judge the illocutionary force of these utterances, we have to take into account who the addressee can be. Addressees in early contexts almost exclusively belong to higher levels of the social hierarchy: the king, the Pope, the Infante, the admiral, the prince, etc. When in later contexts other addressees are possible as well, the context also reveals a difference in tone: the content of the request seems to be less polite and the speech act seems to be more pressing (see Nagy C. forthcoming).
A further factor can be that the inventory of speech act verbs can be different not only in different languages but also in different historical periods of the same language. When interpreting utterances in historical documents we should consider that speech act verbs in use in earlier periods may be out of use today, or they may have acquired new meanings in the course of time.
Last but not least, it is important to mention that the investigation presented in this paper is only a part of a larger research project on constructions used to perform directives in medieval Catalan and Spanish in a comparative approach. This is important because it is almost impossible to identify the functions and degree of politeness of a single construction without situating it in a broader ''pragmatic space'', in other words, without saying how and when it is used in relation with other linguistic means appropriate to perform similar illocutionary acts. Consequently, linguistic means used to perform directives should be investigated all together in a comparative approach. Apart from the construction presented in this paper, medieval Catalan used the pure imperative to perform directives and other constructions as well, e.g. with the verbs pensar 'think' (cf. ex. 6), voler 'want' (cf. 1, 14, 21 and 22) and deure 'have to'.

Conclusion
In this paper I have discussed some uses of constructions formed with subjunctive forms of the verb plaure 'please' in medieval Catalan. This study has emphasized the role of constructions in semantic change: the same lexeme can follow different paths of semantic change in different constructions. The verb plaure 'please' was used in some utterances as a linguistic means of indirectness when performing directives, while in other contexts it acquired specific meanings conveying the speaker's subjective attitude. The former use is interesting from a theoretical point of view, because its description complements Searle's generalization about possible ways of performing indirect directives (see Nagy C. forthcoming). However, the main focus of the present paper has been on occurrences where the verb plaure 'please' has a divine person as Experiencer, in other words, where the speech act is aimed to win God's approval. My findings have revealed that this construction became a conventionalized phrase with a special role in interpersonal interactions. It tended to appear at certain points of discourse and played a particular role in its structure and in the argumentation. As regards the illocutionary force, we can analyze these utterances either as directives or as expressives. However, this is not straightforward in every context, given that these speech acts can be interpreted as directed at God and the discourse participants at the same time and they fulfilled various interrelated functions in the interaction. They were used to express the speaker's desire in optative sentences, to convey a kind of ''concluding evaluation'' at the closure of interactions, or to reflect the speaker's own subjective evaluation on a certain issue. The construction plàcia a Déu 'may it please God' followed a particular path of semantic change and, mainly in its negated form no plàcia a Déu 'may it not please God', underwent subjectification.
Furthermore, I have claimed that qualitative analysis is indispensable in diachronic speech act analysis. Speech act verbs can be searched for in large corpora automatically and analyzed quantitatively, but they in themselves do not entirely reveal the illocutionary force and function of utterances. We should take into account a lot of additional information and keep in mind that different sets of speech act verbs can be used in different periods of a language's history, and the speech act verbs used nowadays may have been used differently at earlier stages.
As a next step of my investigations, I will include other constructions used in similar contexts in medieval Catalan and Spanish, in the hope that the role and function of the construction presented in this study may become even clearer.