FREQUENCY AND TYPES OF INDIRECT REQUESTS IN HUNGARIAN

The paper presents the frequency and types of illocutionary metonymy in Hungarian directives in various social contexts, taking into account the evoked elements of the request scenarios. The recognized and construed social relations have an impact not only on addressive forms but also on the appearance of other elements such as indirectness and its scalarity, which clearly shows that human language activity reflects the physical and social worlds of the intersubjective context (cf. Verschueren 1999). Indirect directives are based on illocutionary metonymic scenarios (Panther–Thornburg 1998) and by evoking a part of the scenario referring to the core action, they give access to the illocutionary scenario domain. The scalar nature of indirectness (Panther–Thornburg 1997, 1999, 2007) depends on the number of evoked elements and their conceptual distance from the core of the request. As the gathered material shows, social context has an impact not only on indirectness of the request but also its structure and type. The paper, based on a discourse completion test conducted among Hungarian data providers, presents not only the frequency of particular types of illocutionary indirectness in specific social contexts but also analyses the most frequent strategies attested by Hungarian data, taking into consideration their type and place in the illocutionary request scenario and compares the social contexts of indirectness with the social context of direct directives. The analysis of material gathered by a discourse completion test shows a correlation between social context and directness in requests in Hungarian, also highlighting the practices of being indirect by using hints, referring to common background knowledge, and employing conventional indirect requests.


Introduction
Based on Hungarian data, the paper presents the frequency of illocutionary metonymy in directives in various social contexts, evoked by various elements of request scenarios and analyses the types of appearing request scenarios. The adopted perspective takes into account the socio-cultural situatedness of language activity (Croft 2009), since grammatical constructions and operations are manifestations of domain-general cognitive abilities (Tátrai 2013: 199) and knowledge about our social world.
As human language activity is understood in social interaction, and at the same time as social interaction, the recognized and construed social relations have an impact on the usage of addressive forms, often understood as main markers of linguistic politeness. Nevertheless, as the presented research shows, social context has an impact on the appearance of indirectness and its scalarity. Indirectness in directives is based on illocutionary metonymic scenarios (Panther-Thornburg 1998). By referring to the part -but not to the core -of the scenario, access is granted to the scenario domain, and the indirect request can be contextualized in the frames of the referred scenario (for more on the role of domains in metonymy, see Croft 1993).
Depending on the part of the scenario being evoked, directives may be based on the BEFORE, CORE, or AFTER metonymy. In the gathered material, both in Polish and Hungarian, the majority of indirect requests is based on the BEFORE metonymy. When being direct the data providers referred to the CORE of the scenario.
The scalarity of indirectness (Panther-Thornburg 1998, 1999, 2007Thornburg-Panther 1997), is governed by the accessibility of the scenario domain -the more accessible it is, the clearer and more evident the request is perceived, leaving a small possibility of other interpretations of the speaker's intentions. The scenario domain's accessibility in turn depends on the number of evoked elements and the conceptual distance of elements from the core. As the presented research shows, referring to the most distant parts of the domain could be observed in distant social relations, and when the request was perceived as a big and difficult one. However, the accessibility of the domain does not only depend on how we place the scenario evoked in the request in terms of the request scenario, but also on how mentally distant a given element is. This mental distance can be shaped by language-specific language conventions, conventional and frequent use of a given type of indirect request, but also but the shared knowledge and experiences of participants. This question will be analysed in more detailed way in the analytic part of the paper (Section 4).
The paper, by presenting the types of illocutionary metonymy appearing in various social contexts, shows that the understood and construed social context correlates not only with the addressive forms being adopted and the use of grammatical person but also on other aspects of the request such as its indirectness, showing also -depending on the social context -patterns in the use of various types of indirectness (for the interaction of metonymy and grammar see: Brdar-Brdar-Szabó 2017: 126-149).
As human language activity consists in social interaction, the recognized and construed social relations have an impact on the usage of addressive forms, oftentimes understood as main markers of linguistic politeness. Nevertheless, as the presented research shows, social context has an impact on the appearance of indirectness, its type and number of supportive strategies.

Research method and preliminary remarks
The research is based on data elicited from Hungarian informants (aged 19 to 35, total number of L1 data providers: 44) in their L1 by a discourse completion test. The test contained 9 different situations, with socio-culturally diverse situations, adding up to a total of 28 situations. The collected data (1232 answers in total) was analysed regarding the types, number, frequency and social context of instances of illocutionary metonymy in directives.
The informants were aged 19 to 35. 90.7% of Hungarian women had finished secondary education, and 9.3% of Hungarians had completed higher education. The social relations differed in terms of hierarchy (boss, teacher), acquaintanceship (stranger, acquaintance, friend) and family bonds (sibling, mother, mother-in-law, grandmother). The analysed directives also differed in terms of the type of request. The discourse completion test (DCT) was filled in by participants on hard copies, within a specified time frame (45 minutes), and was designed to examine conventional linguistic knowledge in recurring contexts. As in other questionnaire-type tests, the more conventional was the given situation the more similar answers were given, which shows that DCT's examine conventional employment of language in conventional social contexts. The aim of the study was to examine types of indirectness appearing in Hungarian requests, and their hypothetical co-occurrence in conventional social contexts. The repetitiveness of the context was provided by DCT, and the need to employ the most mentally accessible devices in a given social context was reinforced by the time limit.
The research was driven by the research questions presented below: • What types of indirectness appear in Hungarian request?
• Is there a co-occurrence of given type of indirectness and social context?
• In what social context direct request appear?
• And if indirectness/directness shows correlation with social context can it be considered as social marker in our linguistic activity?
The paper presents answers gathered among Hungarian data providers, nevertheless the analogical DCT was taken among Polish data providers, so the appearance of indirectness and its cooccurrence with social context can be examined in other languages, as using metonymical structures in language and thought are commonly used in human activity.
As the research by Csató and Pléh (1988) showed that the type of request had an impact on the appearance of indirectness in the analysis the answers were classified for the types of directive (asking for an object, or making an effort), furthermore they were classified in terms of the directness/indirectness, and conventionality of used indirectness. In the next step happened the analysis of the type of the employed non-conventional indirectness and its consequences in terms of construing. The occurrence and frequency of given types of the requests was analysed in various social to verify the hypothetical co-occurrence of the type on the request and social context.

Indirectness as metonymic structure in requests
The main interest of presented paper are metonymical structures in Hungarian request and their co-occurrence with social context. Indirectness in our language activity has a crucial role in expressing the metonymic nature of our thinking. Metonymy (similarly to metaphor) is one of the basic conceptual and linguistic figures of thought and language (Panther-Thornburg 2017: 275;Panther-Thornburg 2011: 239;Panther-Radden 1999;Barcelona 2000), which is present not only in language activity but also in interpreting meanings in general (Panther-Thornburg 2007). In the processes of metonymy, one entity in a conceptual domain is reached through another entity (Langacker 1993: 30;see also: Kövecses 2005: 149;and Kövecses-Radden 1998: 21). Metonymic structures build relations within conceptual domains so that the target content can be reached through another element, the source content. The relation's strength is subject to variation, with the distance between the two elements determining to what extent the source element is salient (Panther-Thornburg 2011: 242). Language activity is perceived here as social cognitive activity, and as the perception of social context has an impact on our linguistic choices the arising question is whether the social context has an influence on indirectness, and if so, what is its type and frequency. Therefore the main topic of the study is not requests as a directive speech act, but the appearance of indirectness in them as a marker of perception of our social word (for the link between politeness and indirectness, see Wierzbicka 1991;Blum-Kulka 1987, 1989 The classification of speech acts is not always easy since illocutionary force is often defined by the speaker's intention (Sadock 2006: 53), which can often be associated with conventional devices for given community. Speech acts, including directives, can be implemented with sentences belonging to different sentence types. Moreover, the sentence types belonging to each type belong to conventional constructions that are related to a conventional speech action value (Croft 1994;Goldberg 1998Goldberg , 2006König-Siemund 2007;Sadock-Zwicky 1985). The sentence type expressing a given speech act value involves typical constructive solutions in a typical social context. These solutions are language-specific, but for the given type show common features in terms of indirectness, supportive strategies, epistemic uncertainty, and the fact that their elaboration shows typical connections with the social context.
In the gathered material the purpose of the utterance is to make a recipient act in such a way that is expected by the speaker, and in such terms -regarding illocutionary force -that should be understood as matching the speaker's intention (see also : Szili 2004) -therefore utterances gathered in the discourse completion test should be regarded as directives. For the prupose of the study regarding the requests, I will make a few brief remarks only. With regard to requests one should take into account their elaboration -the extent to which the request is elaborated (addressive forms, preparatory strategies, supportive strategies), and in what way it is elaborated (indirectness, lexical choices, expressed epistemic uncertainness, conventionality). The presented paper is focused on in indirectness in request and its co-occurrence with social context. Requests were used on purpose, since if one wants to be successful with his/her request s/he should take into consideration social context by employing the most appropriate linguistic expressions, including the construction of the request. The second factor that can have an impact on the form of a directive is the weight of the request (Csató-Pléh 1989;Pléh 2012). The request is perceived as bigger if the investment of time or effort in fulfilling it is considered to be big, or smaller if the investment of time or energy needed to fulfil the request is smaller. While considering the weight of the request, the speaker takes into account his/her socio-cultural knowledge about what one can ask in the given type of relation and whether this kind of request is conventional in a given situation.
Indirect requests do not express the intention of the speaker in an obvious way but rather through metonymic structures which refer to the intended action, relying on elements from the script associated with the request. Characteristically for metonymy, one of the elements activates (in varying degrees of strength) the script of the action, depending on how far conceptually it is in a conceptual frame (see Panther-Thornburg 2011: 256). Conventionally, indirect requests are also based on metonymic relations; in Hungarian, they involve the tud ('can') auxiliary verb and conditional mood. When making requests, the speaker should take into account the character of his or her social relation with the addressee and the type of request. Depending on the type, fulfilling it demands various amounts of energy, time or money. In the case of an information request, the amount of invested energy is relatively small; this type requires verbal action and the sharing of information from the speech partner. In the research presented here, the sharing of informationon the whereabouts of a fitting room and a street -did not have any strategic value. In the case of asking for value, the addressee had to give or lend something and in the case of asking for a favour, he or she had to make a time and energy investment.
Panther and Thornburg do not only discuss predicative, propositional, or referential metonymy (Panther-Thornburg 2011: 246-247). In a series of other papers (Panther-Thornburg 1998, 1999, 2003, 2007, 2009) they state that metonymy appears also at the speech act level, which they call illocutionary metonymy. An indirect speech act presents one of the parts of the scenario for a given speech act, and depending on which part is presented, the BEFORE, CORE, or AFTER type indirect speech act is adopted.
Depending on which part of the scenario is being represented, that before the requested action (BEFORE), that after the action (AFTER) or a part of the action itself (CORE), a BEFORE, CORE or AFTER indirectness appears. When asking about ability (e.g. Can you pass the salt, please?), an indirect metonymic structure is presented in such a way that the BEFORE part of the scenario is highlighted, where we assume an ability to execute an action.
If the speech act contains more elements of a given scenario, the target is more easily identified, and it is easier to understand the intention of the speaker. That is why indirectness should be considered as a scalar phenomenon, as it varies in extent, it has a scalar nature (Panther-Thornburg 1998: 768). It is plausible to assume that the extent to which supportive strategies are involved in evoking elements of an action scenario will have an impact on how the indirectness of the utterance is perceived. This is because meaning, including discursive meaning, is presented through frames (Fillmore [1982] 2006), ICMs (Lakoff 1987) and scenarios (Panther-Thornburg 1998:756). The scenarios are built up from elements that are in metonymic relation with the entire scenario, and the elements that create a connection between the whole scenario and the evoked elements build up the context of the utterance. Also, it is important to mention that during socialization, through experience, we learn a variety of scenarios, and thus the interpretation of indirect acts is connected with identifying a proper scenario by the speaker and the addressee. These scenarios contain sets of actions combined with typical (language-specific) language activity.
As Panther and Thornburg argue in several papers, the scalarity of indirectness is also governed by the place of the element in the illocutionary scenario, a model of which is presented below: Figure 1. Illocutionary frame of the request (Panther-Thornburg 2017: 282) The scalarity in their model depends on whether the evoked element is closer to or more distant from the CORE of the scenario, and the closer it is to the core, the more easily it is accessed, in this way reducing the extent of indirectness. Also a higher number of evoked elements of the request scenario will make it more accessible, easier to identify, and thus less indirect.
The metonymic structures appearing in utterances evoke a part of the scenario, namely the part preceding the action (BEFORE type indirectness), the consequences of the action (AFTER), or the action itself (CORE). This is why the metonymic structure of asking about ability (Could you pass the salt?) highlights the fact that in the scenario of fulfilment there should be an ability to do the requested action in the first place.
In the results presented below, the main strategies concentrated on the presence of the needed item (Nincs egy fölösleges jegyed? Eng. Do you have a spare ticket?), asking about ability (using auxiliary verbs both in Polish and in Hungarian), mentioning that there is no good reason not to give/pass the item, or no good reason not to carry out the requested action, so indirectness was based on the conditions of fulfilling the scenario, in other words, the BEFORE type indirectness played a dominant role.
The CORE type directives are realized as direct ones, with the speaker asking the addressee to do something for him/her without using metonymic structures. E.g.: (1) Adjál egy ezrest! 'Give me one thousand!' (2) Kapcsod ki a zenét! 'Switch off the music!' The research shows that this strategy was used in the case of small social distance, mainly among friends, provided that the nature of request was not perceived as a big one or potentially offensive one. The CORE type strategy is based on expressing obligation (must, should, mustn't, shouldn't), and it did not appear in the collected material (answers provided by Hungarians data providers).
Predicative metonymy, based on the expression of obligation (must), ability (can, could) or permission (to be allowed) is productive in English or German but its application in Hungarian is much more limited (Panther-Thornburg 2017: 281).
In Hungarian answers, the AFTER type metonymy did not appear at all. Panther and Thornburg (2017: 282-283) mention only one strategy in this type, in which the addressee is going to give the requested item to the speaker (realized by will, won't, would, wouldn't auxiliary).
It should be mentioned, however, that understanding indirect directives containing metonymic structures requires not only accessing the domain of the illocutionary scenario but also that the speaker and the addressee, relying on shared socio-cultural and contextual knowledge, should access the same scenarios, so they can identify each others' intentions (for more on identifying intentions, see Tomasello 2016: 33-54; Gazzaniga [2008] 2011), which are crucial for identifying the type of speech act (Tátrai 2017) and accessing the right scenario. This mental accessibility is not only governed by where we can place a request on the above presented request scenario schema, but also by mutually shared knowledge and experiences of the participants. If, for example, one is driving with his/her friend and does not like the music, they can say: (3) Oh, not Dankó radio again! When the participants share mutual knowledge, the meaning of this utterance will be clear for the other participant, and he/she can switch off the radio or change the music without being directly asked to do it. The mutual knowledge, that is, that the participant cannot stand Dankó radio, enables him/her to interpret this linguistic behaviour as an indirect request and leads him/her to action. In the used example other elements express negative attitude, but with mutually shared, well-entrenched knowledge the participants will be able to adequately interpret an utterance such as (4) as well.
(4) Oh, Dankó radio! Being indirect is not only the representation of our metonymic way of thinking but it should also be considered as a politeness strategy. When using an indirect speech act, the speaker can still be understood while also giving the addressee more space. In particular, the latter can react in such a way as if he or she had understood the utterance differently (in some cases directly) and in this way the addressee with his or her reaction can override the intended meaning, which can save him or her from saying "no".
The outcome of the research shows that the appearance of indirect directives correlates with social context -the number of indirect directives increased when social distance between speech partners increased, and when fulfilling the request demanded more (energy or value) investment. Politeness here is not understood in the way presented in Brown and Levinson's (1987) work, but rather as being adequate in a given social context (Watts 2003;Watts-Locher 2005). The received answers show that metonymic structures appear in specific patterns depending on the social context of the directive.
The same request is differently formulated in different social contexts and social context has an influence on the type of indirectness, so its appearance should also correlate with the use of T/V, where T stands for second person singular forms of verbs, and V for third person singular forms of verbs and associated addressive forms; for more on this distinction see Wardhaugh ([1995] 2005: 233-256). As Veres-Guśpiel (2017) shows, a correlation between social distance, the use of T/V forms, and the appearance of indirectness can be experienced not only in Hungarian but also in Polish (Veres-Guśpiel 2017: 75-78). Their occurrence depends on social context, as shown below with percentage data. The figures show the occurrence and type of indirect directives used when informants asked for a bus ticket (Figures no. 2 and 3) and when they were supposed to ask the listener to stop smoking (Figure no. 4 and 5). The above mentioned is illustrated with Figures 2 and 3, and Figures 4 and 5.
When the speaker was asking for a spare ticket, direct requests appeared only when the addressee was a friend, in the rest of cases indirect BEFORE requests were used. As the next figure shows, the supportive strategies rarely appeared in the case of direct requests, and when used with indirect requests their rate increased when the social distance grew between participants. In the second presented situation (when data providers were to ask someone to stop smoking), the direct request appeared in relations in which T addressive forms were used and V forms as well, nevertheless in the case of V addressive forms a smaller number of data providers decided to make any remark. The supportive strategies once again appeared more frequently in the case of indirect request (see Figure 5).  Initial data show that dependencies can be experienced between social context, nature of request and its construction. A further remark that should be made at this point is that, as previous research has shown (Blum-Kulka-Olshtain 1984), the structure of a given speech act is language-specific, and so the type of illocutionary indirectness and also its appearance may vary cross-linguistically. As indirectness is based on evoking one or more elements of the request scenario, the way of our linguistic socialization will have an impact on how we interpret a given element in context -if our addressee can access similar scenarios, s/he can interpret it in accordance with our intentions. Nevertheless, interpretation can be influenced by one's own mental world and other contextual factors depending on which element of experience-based knowledge is activated.
Before presenting detailed results, a few general remarks are in order. According to the research data, the number of non-conventionally indirect directives increased with social distance and the weight of the request. Being direct was restricted only to close social relations (e.g. friend, sibling, mother). In Hungarian, however, the most conventional way for asking for something is to adopt the tud auxiliary verb, and thus this kind of request can be regarded as a conventional indirect request. As the object of a request gets more unconventional or the request itself can be perceived as a big one, the rate of this strategy decreases and we see more diversity in answers regarding types of request. Conventional indirect directives appeared in typical social relations and social situations, nonconventional indirectness and hinting appeared in the case of non-typical requests, in both close and distant social relations, or requests perceived as face-threatening. Finally, hinting was based on evoking the very first elements of the scenario or a condition of fulfilling it and often relied on shared contextual knowledge.

Results
The gathered material has been analysed and presented accordingly to the results of Csató-Pléh (1988), taking into consideration the weight of the request and the social context of the request. As for the types of request, one group was formed by requests of an item, value (bus ticket, loan and key), the second category was formed by requests that had a potentially face-threatening character (asking someone to stop smoking or to switch off the radio) and the third group were requests of favour (handing over a letter and proofreading). The first group of the request needed a sort of financial investment, the second one (FTA request) could be potentially hurtful for the addressee, and the third one was connected with energy and time investment. In each category, requests differed in terms of the how much efforts should be invested to fulfil the request, or how much they can be considered as face-threatening. In each situation, the same request was made in various social contexts evoking close and distant relationships based on solidarity and hierarchy as well.

Directness in Hungarian
Before presenting the appearance of indirect directives in Hungarian answers, it is worth taking a look at what social context directness appears in.
In the Hungarian material, direct requests did not have a high share and appeared in specific situations. In the case of asking for an item, they appeared when speakers were requesting it from a relative (mother: 35%, sibling: 44.3%), grandmother: 25.6%, but only 6.8% in the case of a motherin-law. What is more, the numbers presented above show that the number of direct requests decreased as social distance increased.
When asking for a favour, direct requests appear when the speaker is asking a friend to change the music in the car (39%), asking friends to stop smoking in a non-smoking train compartment (41%) and asking a friend to proofread a translation (16%).
This shows that direct requests appear with small social distance (friend), and close family relations (sibling, mother). Not only Hungarian data but also Polish data show that when choosing whether to be direct or indirect the social distance plays a crucial role (for a detailed cross-linguistic comparison, see Veres-Guśpiel 2020) and has a similar impact on frequency of in/directness. When analysing indirectness, one should not only take into consideration the proportion of direct and indirect requests, but also types of indirectness scenarios appearing in various social contexts. Close relations are often characterized by mutually shared knowledge, and that is why being indirect can exploit this shared knowledge at the same time also highlighting the discourse partners' relation and sense of community, especially in such situations when the request can be potentially perceived as hurtful or invasive.
Although the paper focuses on indirectness, it should be said that in the analysed material direct requests -in a small rate -also appeared. In requests for an item or value (loan, bus ticket, key), the social context and the nature of request had a big impact on being/not being direct. In requests for a loan, direct requests appeared in a small rate (4,5%) and only in the case of close relations (friend -4,5%, acquaintance -4,5%). In requests for a bus ticket, direct request appeared when addressed to a friend (9%). In the case of asking for a value or item, direct requests appeared with the biggest frequency in family relations. Age gap had a clear impact on their frequency -when asking their grandmother, 25% of data providers used direct request, whilst when asking their mother, 43,2% employed direct requests. The second factor that had an influence on their frequency was the type of family bond and associated distance -when asking their mother-in-law, only 6,8% of responders used a direct form, while when asking a grandmother 25% and with a sibling 34%. Other studies (Veres-Guśpiel 2017) show that in Hungarian, age is a social factor that also has an impact on how we interpret the virtual use of WE in requests, so age gap between participants is potentially significantly influencing our understanding of the social world.
In requests for a favour, direct requests appeared between friends (39% asking someone to change the music, 41% asking someone to stop smoking), and when a request was connected with obeying the rules (asking someone to stop smoking in a non-smoking area). In this case it appeared also in requests directed at a stranger (13.6%), teacher (11.4%) and a boss (6.8%). Although the request was justified by general rules, the data show that not only social distance had an impact on being direct/indirect, but also relations of dominance. In the case of favours that demanded an investment of time and energy, direct requests appeared in a small rate when the speaker was asking someone to give back a letter (young speaker and an elderly neighbour 2.3%-2.3%), and more frequently when they were asking someone to proofread a translation -15.9% of data providers were direct when asking a friend, and 11.4 when asking a teacher addressed with T forms, and 4.5% when asking a teacher addressed with V forms.
Based on these data it should be said that direct requests appear mainly in close social relations, and often when a speaker has justification to make the request (direct requests appeared most frequently when the speaker was asking someone to stop smoking in a non-smoking area), which shows that directness often appears when a speaker can perceive a directive as an instruction.

Social context and types of indirectness
In the gathered material, indirect requests played a dominant role, and in the case of Hungarian the BEFORE type of request was typical. In the presented work, I do not analyse the prevalence of the supportive strategies. However, it can be observed that in various social contexts the supportive strategies appear in different percentages, indicating certain tendency patterns.

Asking for an item, value
In two situations, the data providers were to ask for a small loan (from a friend and a teacher acquaintance), for a bus ticket at a bus stop (from a friend, a stranger and a boss), and for a key to the apartment (from their mother, sibling, grandmother and mother-in-law).
When asking for the item (bus ticket, key), the most frequent strategy was an indirect request. The most conventional device involves the tud 'can' auxiliary verb and a conditional suffix, expressing uncertainty as to whether one's partner will be willing to fulfill the request. This conventionally indirect strategy introduces the condition of fulfilling the request. As this is the most typical way of requesting in Hungarian, the accessibility of the request meaning is high, but also as it is one of the most conventional ways, it can be perceived as the most neutral and thus adequate when asking a stranger about whom the utterer has a limited knowledge. When asking for a ticket at a bus stop, the informants asked if the addressee had a spare ticket (stranger: 27.3%, boss: 36.4%, friend: 43.2%).
'Excuse me, don't you have a spare ticket?
(8) Ne haragudjon. Nincs véletlenül egy felesleges jegye? 'Excuse me. Don't you have a spare ticket by any chance?' This strategy differs in some ways from the above-mentioned one -first of all the utterer asks about having a ticket, so the physical condition of fulfilling the request. Considering the request scenario this strategy should be placed just before the very starting point on the presented schema (Panther-Thornburg 2017: 282). It is worth adding that in the case of asking for a ticket from a boss such preparatory and supportive strategies appear as giving a reason, saying that one does not have a ticket and has no possibility to buy one, or promising to give back its price. Being indirect and asking the addressee about having a ticket, the speaker gives the opportunity for the hearer to deny fulfilling the request just by saying that he/she does not have one. When s/he is willing to fulfil the request, the discourse partner creates a sense of solidarity not only by handing over the ticket or giving a loan but also by the request's adequate interpretation.
When informants asked for a key to the apartment, in close family relations (mother, sibling), direct requests appeared (mother: 43.2%, sibling: 34%), and when the distance grew, attention was directed onto the speaker with the expression elkérhetem ('May I ask'), which is by its form drawing attention to the first person. This expression also involves the -hat/-het morpheme used when asking for permission, and is used normally by a person in asymmetrical social relations (younger to older, children to parents, student to teacher, etc.).
(9) Csókolom! Elkérhetem az X néni kulcsát egy kis időre? Nagyon fontos lenne... Az enyémet odaadtam... 'Good morning! May I ask for a key for a little time? I would be really important… I have lent mine…' (10) Ne haragudjon, kölcsönkérhetném a kulcsát addig, amíg vissza nem kapom a sajátomat? 'Excuse meg, may I ask for a key till I get back mine?' This strategy also is an indirect BEFORE strategy pointing not only to the person of the speaker and should be placed just at the beginning of the request illocutionary scenario since the speaker asks if s/he can pose a question. As it can be seen in other situations, this type of indirectness appears mainly in social contexts characterized by larger distance. The tud ('can') verb plus conditional was the most frequent choice in requests for a loan (friend: 45.5%, acquaintance: 41%), or when the social distance was regarded as fairly large (mother-in-law, asking for a key: 34%). What should be said, though, is that this construction appeared in similar rates (between 23% and 34%) in the case of close family relationships (mother and siblings). The second most popular strategy was a direct request and in the case of larger social distance (motherin-law). Then came the elkérhetem 'may I ask' expression. This shows that the conditional form of tud ('can') can be used in various social contexts and can be perceived as a neutral device. To summarize, direct requests typically appeared in the case of small social distance, a high amount of shared background knowledge, and close family relations. Social relations had an impact on expressing uncertainty regarding fulfilling the request, at the same time expressing various expectations if the request could/should be fulfilled. In the case of typically smaller social distance, the expression of uncertainty decreased, and in the event of larger distance, it grew.

FTA (Face-Threatening Act) type of request
In the first of two requests considered next, data providers were to ask their friend, acquaintance, boss and teacher to change the music or switch off the radio in a car, which is connected with expressing displeasure and questioning another person's taste. By violating the addressee's selfesteem, choice or taste, these can be considered as FTAs -face-threatening acts (Brown-Levinson 1987). In the second situation, the participants were to ask a friend, stranger, boss, and teacher to stop smoking in a non-smoking area, and in this situation, they had to point to the fact that the other person was violating the rules. Both requests can be regarded as FTA requests, as they show our displeasure with the behaviour of the addressee. In this case, direct requests appeared in interactions with friends (music: 39%, smoking: 41%). When asking a teacher or a stranger to stop smoking, the number of direct requests was over 10%, but here it has to be added that the speaker had justification to ask for it, as it was a non-smoking area. These types of directives were perceived as awkward and it is shown also that several informants decided to stay silent, especially when hierarchical relations came into consideration (boss, teacher). The second strategy that appeared was the use of tud ('can') in conditional mood and often in first person plural (Why couldn't we…), which appeared when the speaker was asking someone to change the music. It was characteristic of the situations where both the speaker and the addressee were in the same closed space, but the action itself should be executed by one person only (preferably the owner of the car), so first plural person was used in a virtual manner (Veres-Guśpiel 2017), and it was a way to express solidarity with the other person and mentally incorporate him/her in the action. This mental incorporation not only expresses solidarity but also implies that both the speaker and the addressee have the same interests. (11) Nem hallgathatnánk másik zenét? 'Couldn't we listen to some other music?' (12) Bocsi, nem hallgathatnánk valami mást? 'Sorry, couldn't we listen to something else?
When addressing their boss, a case of increased social distance, informants once more resorted to the megkérhetem, hogy expression ('May I ask') (11.4%). This strategy appears when the social distance grows and complying with the request is considered difficult. Hence, this context is similar to the case when the data providers were to ask for keys. This tendency can be observed also in the next situation. The data providers were to make a remark expressing that the other person was violating the rules. The most popular strategy was to state that the compartment was a non-smoking compartment, i.e. they pointed at the very first element of the speech act scenario. As the social distance grew, this strategy was more frequent (friend -22.7%, teacher (T) 27.3%, boss -29.6%, stranger 45.5%). The megkérhetem, hogy ('May I ask') expression also appeared in requests directed at a boss (9%).

Asking for a favour
In the next two situations, participants had to ask someone to check a few pages of translation, which required the investment of time and effort (from a friend, a teacher addressed with T address forms, and a teacher addressed with V forms), and to ask for a letter that has fallen on the neighbour's balcony (here the variants concerned the age of the neighbour -young, elderly and middle-aged). In the first situation, one of the most popular strategies was to say that the letter has fallen down, and it was used in similar rates in each case (26-30%).
(13) Bocsi, az erkélyetekre leesett egy levelem. 'Sorry, my letter has fallen down on your balcony' (14) Szia! Ne haragudj, leesett egy fontos levél az erkélyetekre... 'Hi, my bad, an important letter has fallen down on your balcony…' The second most popular strategy was to make a request by using the megkérhetem, hogy ('May I ask that…') expression and its frequency grew with the age difference (and in fact, it surpassed the frequency of stating that the letter had fallen down when the neighbour was older than the informant). A similar tendency can be seen in the case of asking someone to proofread a translation (2nd person singular verb in conditional: friend: 70.5%, teacher (T address forms): 38.6%, teacher (V address forms): 34%). The megkérhetlek ('May I ask you…') form had a ratio of 2.3% in requests directed at a friend, 22.3% in utterances targeted at a teacher (T address forms), and 27.3% when the addressee was a teacher with whom the speaker uses V address forms.

Summary
In the collected material, a majority of indirect directives were observed, and the indirectness included in them is of the BEFORE type, there was no example for the AFTER type indirectness. The gathered material shows that indirect directives were used in few different patterns and clearly co-occur with specific social contexts: one was asking about having/not having a physical item needed to fulfil the request (eg.: ticket). The second strategy, used mainly in the case of larger social distance (mother-in-law, teacher, boss), was represented by indirect directives inquiring if one can ask for something. These request involved the -hat/-het morpheme (megkérhetem/ megkérhetlek 'may I ask you'), often appearing in Hungarian requests for permission. This strategy was the most popular not only in the case of larger social distance but also when the relationship between the participants was strongly hierarchical.
In both cases these strategies are BEFORE indirect directives, and requests in this type should be placed just at the beginning of the illocutionary scenario.
When complying with the request is considered difficult, especially in the case of larger social distance, informants used shifting the attention to first person singular and often they made a statement of discomfort (eg. Zavar a füst 'The smoke bothers me'), building the foundation for the the requested action, based on the assumption that our communicative actions are intentionally motivated.
In some contexts, using the first person plural was also seen as a device for making a request that potentially could be considered as violating one's taste (asking to change the music). In such cases, using the first person plural pronoun 'we' expressed solidarity, but clearly it was employed in the virtual use when the addressee was the participant who was expected to take an action. This form mentally engaged the addressee as well, whilst expressing common interests.
Hungarian equivalents of the can auxiliary are used in conditional and indicative forms and the variety of situations and social contexts in which this device appeared shows its universal character in Hungarian requests.
Activating the very first element of the speech-act scenario, as mentioned before, appeared in various forms, but with two different outcomes. Exploiting the very first step of the illocutionary scenario (an expression of displeasure, pointing out the violation of rules, asking about the possibility of meeting the conditions) shows that this type of indirect directives can be perceived as strong instructions (in the case of large social distance, and when the speaker is pointing out the addressee's violation of a rule). In interactions with friends and family members, they indicate higher intimacy and the existence of a substantial amount of shared background knowledge. Exploiting background knowledge underlines shared experiences and suggests that there is no reason to express the request in an explicit way, since the speaker and the addressee understand each other with no need to put it bald on-record.
Direct directives appear when the relations between participants could be interpreted as close ones, based on shared background knowledge about each other, or between participants classified as close family members. With the increase of social distance, the number of direct directives decreased.
In the end, it is worth mentioning that the more conventional the request, the smaller the variety of answers. In conventional situations the tendencies are stronger, since it is in conventional situations that we tend to make conventional linguistic choices, as we pay less attention to them, perceiving no need for additional linguistic effort. Consequently, non-conventional language choices in very conventional situations can be seen as conveying extra pragmatic meaning.