“ Sorry for your consideration ” : The (in) adequacy of English speech act labels in describing ‘ apologies ’ and ‘ thanks ’ in Japanese

: Despite speech act theory being very influential in pragmatics, the notion of what constitutes a speech act in languages other than English has not received the attention it deserves in the literature. After a brief outline of traditional speech act theory, this paper problematizes the use of English speech act labels by comparing English and Japanese conceptualizations of ‘ apologies ’ and ‘ thanks ’ . The notion of indebtedness and the norm of reciprocity are then discussed, arguing that they can help revealing similarities between ‘ apologies ’ and ‘ thanks ’ in Japanese that are not observed in English. The second part of the paper is empirical in nature and adopts a corpus-assisted approach. The Japanese expression su ( m ) imasen [sorry], usually signaled as apologetic, is used as key word in two web corpora of written Japanese for retrieving metapragmatic comments and naturally occurring exchanges where su ( m ) imasen is framed as an expression of gratitude – a function English apologies do not serve. Finally, the paper proposes the notion of pragmatic space to investigate ‘ apologies ’ and ‘ thanks ’ as neighboring speech acts that overlap to different degrees and present different prototypical features in Japanese and English. The analysis reveals that the acritical use of English speech act labels is not suitable for describing ‘ apologies ’ and ‘ thanks ’ in Japanese.

2 Theoretical background 2.1 Speech act theory problematized Traditional speech act theory (SAT), as outlined by Austin ([1962Austin ([ ]1975 and Searle (1969), contributed greatly to the field of pragmatics by observing that some forms speaker may utter a number of (often conventionalized) words (locutionary act) in order to convey regret for something (illocutionary force), e.g. being late, with the effect to persuade the addressee that they are sorry and, consequently, obtain forgiveness and reset the relationship (perlocutionary effect). On these grounds, I argue that a prototypical 'apology', despite all of its varieties, can be reduced to four essential elements (see Jucker and Taavitsainen's delimitation of 'insults' [2000: 73]). First, an offender, i.e., the speaker or apologizer. Second, a target (or more than one) of the utterance -'apologies' are relevant on the interpersonal level and, thus, require the presence of a target to be communicated to. Third, an "action or event or consequence thereof" (Coulmas 1981: 71; what in Section 2.3 will be referred to as the "object of regret") potentially opened to a negative evaluation on the target(s)' part-in our previous example, being late. Fourth, the target experiences the utterance as remedial and face-saving because it addresses their desire to be free from imposition (Brown and Levinson's negative face [1987]), while also manifesting the speaker's desire to be appreciated by others (Brown and Levinson's positive face [1987]). Such recognition leads to forgiveness. 2 Notably, while 'offender', 'target of the apology' and 'object of regret' are obligatory, 'forgiveness' is optional: I consider as 'apologies' also ineffective 'apologies' that do not have the intended perlocutionary effect on the target. This is also due to the nature of the data employed in the present study, where responses to 'apologies' are often missing and, therefore, it is often the case that I do not have access to how the target experiences the speech act.

'Apologies' and 'thanks' in English and Japanese
In the attempt to find an alternative solution to the use of potentially misleading English verbs for the identification of speech acts cross-linguistically, Wierzbicka (1987Wierzbicka ( , 1991Wierzbicka ( , 1996 proposes the adoption of the Natural Semantic Metalanguage (NSM). NSM employs reductive paraphrases, claimed to be universal, to define speech acts without using terms that may assume (Western) culture-specific set of values. For example, the cultural script (i.e., "universal human concepts, lexicalized in all languages of the world" [Wierzbicka 1996: 527]) of Japanese 'apologies' is characterized as follows: when something bad happens to someone because I did something I have to say something like this to this person: "I feel something bad" 2 See also Deutschmann (2003: 46) for a slightly different conceptualization of the prototypical apology.
Diegoli I have to do something because of this. (Wierzbicka 1996: 532) Notably, in a subsequent work the conceptualization of the English performative verb apologize is slightly different: X apologized to Y (for doing A).
a. someone X said something to someone else Y at that time b. this someone said something like this: "I feel something bad now because I think like this: 'I did something (A) before, you can feel something bad because of it'" c. this someone said it like someone can say something like this to someone else when this someone thinks like this: "I know that this someone can feel something bad towards me I don't want this" (Goddard and Wierzbicka 2013: 171) These wordings suggest that admitting responsibility is not necessarily implied in Japanese nor English 'apologies'. This explains the use of apologetic expressions for expressing sympathy, as in the English conventionalized formulae I'm sorry to hear thata function that, as will be seen, Japanese 'apologies' do not perform. This is in line with Coulmas' assumption that "regret, not necessarily responsibility for an unwelcome change in the course of events, is the point of making apologies" (1981: 76). However, in the conceptualization of English 'apologies' outlined above we do not have any modal verbs of obligation, while the description of Japanese 'apologies' employs the deontic modal marker have to twice (I have to say something like this; I have to do something). It is precisely this obligation associated with social norms that links (im)polite behavior with a social dimension. For the sake of completeness, it should be noted that in previous research Wierzbicka does include into the constituents of apologize a deontic modal verb ('I think I should say it to you') (1996: 215). Nonetheless, it is my contention that it can be inferred with a certain degree of probability that in the Japanese lingua-culture there is relatively more pressure on the speaker to apologize, which explains why apology-like expressions seem to be used much more frequently in Japanese than in English (Coulmas 1981: 82). The quantitative difference between 'apologies' in Japanese and English is also related to what Coulmas refers to as the "object of regret" (1981: 75). In order to better understand what is meant by "object of regret", it might be helpful to mention again (see Section 2.2) Coulmas' rather generic definition of Sorry for your consideration 'apologies' as acts "directed towards some action or event or a consequence thereof […] considered negative and unwanted for the recipient of the apology" (1981: 75). The object of regret is, thus, any "action, event or a consequence thereof" more or less directly linked to the speaker for which a feeling of regret is expected. This expectation of regret on the hearer's part may eventually lead to an 'apology'. Previous studies argued that in Japanese something is regarded as an object of regret even when there is no direct causal relation between the speaker's actions and the offence (Wierzbicka 1996: 530-531; see also Doi 1981). This sociopragmatic difference can result in 'apologies' that English speakers may see as unnecessary.

Indebtedness
Despite the fact that the social expectation for the speaker to apologize is indeed very relevant in a number of contexts, it is my contention that a qualitative difference between speech acts in Japanese and English provides a better explanation for the wider range of application of Japanese 'apologies'. In fact, Japanese 'apologies' seem to overlap to a certain degree with the English notion of 'thanks'. Wierzbicka describes the English speech act of thanking as follows: When someone did something good for me I have to say something like this to the person: "I feel something good because of this " (1996: 534) In the Anglo lingua-culture, when someone does something that produces a positive outcome for the speaker, this is expected to respond in a positive way. However, the Japanese conceptualization of 'thanks' is very different: When I know that someone did something good to me It is good to say something like this to this person: "I feel something bad because of this" (1996: 534-535; emphasis mine) We can thus infer that the focus of 'thanks' in Japanese is not so much on the positive outcome produced by the object of gratitude, but rather on the regret the speaker feels for having impinged on the other person. This is mirrored in Miyake's discussion of two salient notions of the Japanese moral order, namely kari 借り [debit] and futan 負担 [burden]. Japanese speakers use apology-like expressions to convey gratitude when they want the addressee to infer that they acknowledge their debit (kari) and appreciate the (symbolic) burden (futan) s/he is forced to carry (Miyake 1993: 22). What is a benefit for the speaker is thus conceptualized as a burden for the other person and the distinction between object of gratitude and object of regret, though often clear cut in English, becomes rather blurred. Along these lines, Coulmas (1981: 79-81) proposes the theoretical notion of indebtedness, which can be paraphrased as the acknowledgment that the speaker has done something bad (i.e., the object of regret) to the hearer, or that the hearer has done something good (i.e., the object of gratitude) for the speaker. Either way, the speaker needs to give something to the other person as a symbolic compensation. Indebtedness can thus be construed as the link between object of regret and object of gratitude, and consequently between 'apologies' and 'thanks', in the Japanese lingua-culture. In other words, if we reckon that by apologizing the speaker acknowledges an offence and their indebtedness to the interlocutor, the same is true for thanking, where the indebtedness of the recipient of the benefit can also be made explicit. This explains why Japanese speakers use su(m)imasen where English speakers would say thanks (I owe you one).

Norm of reciprocity
A different but complementary explanation for the blurring of 'apologies' and 'thanks' in Japanese is offered by Ohashi (2008), who also claims that traditional SAT does not adequately account for Japanese speech acts, and that cultural frames of reference must be taken into account when interpreting speech acts in non-Western languages. More specifically, he advocates that the norm of reciprocity, namely "the social obligation to reciprocate benefits to one another" (Ohashi 2008(Ohashi : 2150, is more suited than traditional notions of 'gratitude' and 'regret' for explaining speech acts in Japanese. This is in line with Culpeper and Tantucci's claim that a significant part of interpersonal communication is about "maintaining a balance of payments" (2021: 150). Although reciprocity and indebtedness are clearly closely related to each other, in my view the norm of reciprocity is more general than the notion of indebtedness: while the former constrains virtually any kind of human interaction and includes polite as well as impolite behaviour, the latter was developed to describe 'apologies' and 'thanks' specifically. Notably, the norm of reciprocity also makes explicit the conceptualization of (im)polite speech acts as money, and argues that the social goal underlying 'thanks' and 'apologies' is to maintain a debit-credit equilibrium between the interactants. When we cause offence to someone or someone does something for our benefit, this results in a negative balance sheet. This imbalance in the credit-debit equilibrium is generally thought of as something to avoid and, when not properly addressed, potentially opened to a negative evaluation of our interlocutor as non-cooperative (see Culpeper and Tantucci 2021: 148-150). This set of understandings guide interactants' choices and accounts for 'apologies' and 'thanks' uttered as a symbolic compensation to counterbalance the debt and ensure the maintenance of social bonds. On these grounds, we can safely assess that, generally speaking, achieving a debit-credit equilibrium is regarded as a good thing both in English and in Japanese and the non-adherence to it may result in sanctions. However, as I have already pointed out when presenting Wierzbicka's paraphrases for 'apologies' and 'thanks', in Japanese the adherence to the norm of reciprocity, described by Ohashi as "social obligation" (2008: 2150, see above), is mandatory, as opposed to English (see also Culpeper and Haugh 2014: 178).
So far, I have pointed out how the notions of indebtedness and the norm of reciprocity apply to both 'apologies' and 'thanks', highlighting a link between the two that seems to be particularly relevant in the Japanese context. This assumption is reinforced by Ohashi's finding that, as observed in his corpus of telephone conversations, the 'apology' formula su(m)imasen is commonly employed in thanking, a result showing that 'thanks' and 'apologies' "are closely related in the indebtedness the speaker feels towards the interlocutor" (Ohashi 2008(Ohashi : 2161. In order to test the link between 'apologies' and 'thanks' in Japanese, the remaining part of the paper elaborates on the apology-like expression su(m)imasen and applies the theoretical notions of indebtedness and reciprocity to the analysis of its occurrences, as observed in two different web corpora.

Su(m)imasen
The expression chosen as key word for the corpus-assisted pragmatic investigation of apology-like behaviors in Japanese is sumimasen, often written with the omission of the first 'm' as suimasen. Su(m)imasen is what Blum-Kulka and Olsthain (1984) define as illocutionary force indicating device (henceforth IFID), i.e. an explicit linguistic marker of pragmatic force which "fulfils the function of signaling regret (on the S[peaker]'s part) for X (the violation)" (Blum-Kulka and Olsthain 1984: 206; this terminology first appeared in Searle 1969: 64). The link between 'apologies' and (often formulaic) regret is based on the assumption that 'apologies' are uttered when the speaker's conduct (or something related to it) somehow damaged the addressee (see Section 2.2). Within this interactional setting, it is thus conventionally expected thatother things being equalthe speaker regrets that conduct and will avoid it in the future (Oishi 2013: 5). This is not to say that an IFID by itself conveys the attitude of regret, but that, in conjunction with other syntactic and contextual information, it has a typically associated meaning that includes such attitude.
In terms of morphology, su(m)imasen is the negative form of the verb sumu [to finish]; its literal meaning is [this is not the end] or [it is not finished]. It is conventionally labelled as polite or formal not because it is marked in terms of (im) politeness, but because it contains the formal verbal suffix -masu (-masen being its negative form). In Japanese, formal style is usually considered to index either interpersonal distance between the participants (see Brown and Levinson 1978) and/or the speaker's heightened awareness of the addressee (Cook 2018: 636). Thus, it is commonly chosen when there is a perceived distance between unacquainted people, but it is not automatically perceived as more polite than other speech styles.
The multi-dimensional pragmatic nature of su(m)imasen has received focused attention in the literature and Ide (1998) identifies up to seven pragmatic functions of it: sincere apology, quasi-thanks and apology, request marker, attention-getting device or conversation opener, leave-taking device, affirmative and confirmational response and, finally, reciprocal exchange of acknowledgment. Particularly relevant for this paper are the first two, namely (a) sincere apology and (b) quasithanks and apology, for which Ide proposes the following two examples: Okurete sumimasen be.late sumimasen 'Sorry for being late' (b) 1 N Senyonhyakuen arimasu? Rokujūen 1,400 yen are 60 yen '1,400 yen. (Do you) have 60 yen?' 2 P Arimasu are '(I do) have ((starts to count the coins))' 3 R A ano sumimasen mōshiwake gozaimasen ano uh sumimasen sorry-HON umm 'Uh, urn. sumimasen, I'm very sorry, umm' 3 (Adapted from Ide 1998: 514) The 'apology' illustrated in (a) is quite ritual in nature. As Jucker and Taavitsainen (2000) point out, "the ritual should not be seen in direct opposition to the personal" (2000: 74) and highly conventionalized social routines may indeed correlate with some authentic regret. However, since the evaluation of the 'apology' as sincere is very problematic and goes beyond the scope of the 3 Ide contextualizes this second exchange as follows: "A male patient (P: 30-40) is making a payment, while the receptionist (R: 40s) and the nurse (N: 40s) are at the counter. The fee is 1,460 yen (= $15.00). The patient understands that the receptionist is short of change, and after handing 1,400 yen, tries to get the exact change from his wallet" (1998: 154).
Sorry for your consideration present study, I will employ the traditional term 'apology' to refer to utterances such as (a).
With reference to (b), the label "quasi-thank and apology" emphasizes that the receptionist feels regret for having the patient looking for the correct change, while also feeling grateful for it. In my data I found many similar cases where su(m) imasen was used in circumstances where an English speaker would have said thank you, as in: ki o tsukatte itadaite sumimasen deshita concern OBJ use BEN.HON sumimasen-PST *Sorry/thank you for your concern Here, the translation of sumimasen with sorry sounds odd and inappropriate to an English speaker, because, in this specific context, sumimasen does not express the English notion of 'apology', but rather the English notion of 'thanks', functioning in a supportive manner. I thus argue that the label 'thanks' may be more accurate to describe the function sumimasen has in examples (b) and (c), as confirmed by the fact that they can be translated in English with thank you without sounding awkward.

Data and method
The first data source selected for the analysis is the Japanese Q&A website Yahoo! Chiebukuro 'https://chiebukuro.yahoo.co.jp/'. Yahoo! Chiebukuro was chosen for two main reasons. First, the sense of conduct in Yahoo is the one shared by the larger Japanese community: polite-seeming, standard written Japanese is used, and the dominant speech level is the formal one (Nishimura 2008(Nishimura , 2010. This might allow for more generalizable considerations. It was also reasonable to assume that su(m)imasen (which is in formal style, see Section 2.4) might be quite frequent. Second, and most importantly, Yahoo! Chiebukuro has sections dedicated to judgments of (often linguistic) behaviour, such as ningen kankei no nayami 'troubles with interpersonal relationships' and manā 'manners'. These sections are very valuable sources of metapragmatic comments (see Section 4.1) that may reveal an ethnographic view of what is considered as a given speech act in a culture (Jucker et al. 2009: 273). Since I am interested only in those instances that contain su(m)imasen, for the data collection I employed the free software BootCat, a tool for the automated extraction of specialized or topic-specific corpora by web-mining (see Baroni and Bernardini 2004 for more information on the BootCat toolkit). The data set so collected, which contains 98,709 tokens, was then uploaded onto the Sketch Engine software (Kilgarriff et al. 2004). Finally, I used the Sketch Engine concordance tool to identify and then download in plain text format a total of 375 occurrences of su(m)imasen within their co-text. After duplicates were manually removed, 322 occurrences of su(m)imasen were placed into an Excel spreadsheet and manually coded. However, this data collection procedure is quite subjective, and the corpus is very small in size. With the aim of at least partially mitigating these limitations, the patternings that emerged from this first data set were verified in a larger and more balanced web corpus, namely the JaTenTen11 provided by Sketch Engine.

Metapragmatic comments in the Yahoo! Chiebukuro corpus
Using su(m)imasen as key word, I read line-by-line a total of 322 occurrences in the Yahoo! Chiebukuro data set. As a result, 10 metapragmatic comments made by Yahoo! Chiebukuro users on the use of su(m)imasen for expressing gratitude were found. Drawing from Haugh, I take metapragmatic comments to be instances of "the use of language on the part of ordinary users or observers, which reflects awareness on their part about the various ways in which we can use language to interact and communicate with others" (2014: 239). Due to space limitations, in what follows I report only two out of the 10 metapragmatic comments linking sumimasen to both 'apologies' and 'thanks': (1) Kansha no imi o arawasu 'sumimasen' wa, 'nani no okaeshi mo dekizu sumimasen' no imi kara ka, 'kokoro ga sumikiranai' no imi kara hanare, kansha o arawasu yō ni natte kara no hyōgen desu.
The 'sumimasen' that expresses gratitude [kansha] takes the distance from 'sumimasen for not being able to give you anything in return', or the idea that 'my heart is not at peace', and it has developed after expressions of apology [shazai].
(2) Hito ni nanika shite morattari, osewa ni nattari shita toki wa 'arigatō' to iu beki na noni, 'sumimasen' to iu hito ga ōi desu. 'Sumimasen' wa owabi no kotoba de atte, kansha o arawasu kotoba wa 'arigatō' de aru hazu desu. 'Arigatō' wa kansha igai no kimochi wa zenzen nai, 100% jiritsu shita kotoba de aru no ni taishi, 'sumimasen' wa kansha shinagara mo nani ka hara ni ichi motsu motte iru yōna, ato de monku o ie sōna 'kurai imēji' ga aru no desu ga … Between square brackets there are a number of key words as observed in the Japanese text. Such key words are quite frequent in the data and are relevant for a number of reasons, which will be addressed at the end of the section. In (1) the speaker acknowledges that the use of sumimasen to express gratitude comes from feeling bad for not being able to reciprocate a "gift" (whether this is a material gift or any kind of action or event that brought benefit to the apologizer). Sumimasen is thus framed as a pragmatic device that shows awareness of the temporary imbalance in terms of debit-credit equilibrium and attempts to reduce indebtedness on the speaker's part. Example (2) attests the ambivalent nature of sumimasen and, in this specific excerpt, is associated with a negative connotation. Such negative connotation becomes clear when we look at the rest of thread (not reported here for space constraints), where one of the comments discussing the use of sumimasen to convey gratitude reads kibun no ii mono dewa nai 'it's not something that makes you feel good'. Interestingly, sumimasen is then compared with arigatō, the unmarked and, according to the speaker, straightforward expression of gratitude in Japanese. Here the use of sumimasen to convey gratitude seems to trigger an assumption of relevance and may lead to the inference of additional meanings (see Ide 1998: 515), but it also supports the assumption that Japanese speakers conceptualize 'apologies' as closely associated with 'thanks', though not as fully complementarya point I will go back to in Section 4.2. It is also worth noting that in the above two examples the speakers use specific terms to denote what in English is the speech act of apologizing, namely shazai 謝 罪 (Example [1]) and owabi お詫び (Example [2]), both translated as 'apology' in the Taishukan's Genius English -Japanese Dictionary (3rd edition). However, the counterpart of the English speech act label 'thanks' is kansha 感謝, which literally means 'gratitude'. Indeed, Japanese does not seem to have a specific label for the speech act that in English is referred to as 'thanks', which is more generally indicated using the feeling verbalized by the speech act. This suggests that the English speech act label 'thanks' and its Japanese counterpart 'kansha' (which here is also in single quotation marks to point out that it is not the word 'used' in actual interaction, but it is employed as a second-order theoretical speech act categorization) may occupy different pragmatic spaces-something I will discuss in more detail in Section 5.

Verification from the JaTenTen11 corpus
Metapragmatic comments are very helpful resources because they provide us with access to the layperson's understandings of what constitutes (im)polite and (in) appropriate behaviour (Eelen's [2001] "(im)politeness1" and Watts' [2003] "firstorder (im)politeness"). However, they may reflect idiosyncratic uses of the language and/or stereotypes about (im)politeness concepts rather than what's normally happening. In light of these considerations and with the aim to test the above lay interpretations of su(m)imasen as a marker of both 'apology' and gratitude, its use was cross-verified using the concordance tool in the much larger web corpus JaTenTen11. These numbers are very small and should be treated cautiously. However, some interesting features are apparent. In Examples (64) and (92) the link between sumimasen and (ritual) 'thanks' is made explicit by the use of the straightforward expression of 'thanks' arigatō gozaimasu (where gozaimasu is the form in -masu of the honorific verb gozaru and, when added to the plain arigatō, enhances the formality level of the expression). The co-occurrence of conventionalized expressions of 'apologies' and 'thanks', further discussed below, supports the assumption that the two speech acts are related. Example (71) was categorised as 'thanks' because of the presence of the benefactive verb kureru, which appears twice at the left of the key word. Kureru (lit. 'receive') is commonly employed in a giving event where the speaker is the receiver and, when used as a verb auxiliary in the construction V-te kureru, it indicates that the subject of the predicate (in our example the recipient of the speech act) provides some favourable effects on the affectee of the event (in our example the producer of the utterance) (Hasegawa 2018: 511; on the co-occurrence of kureru and 'thanks' see also Ohashi 2008Ohashi : 2157. The acknowledgment on the speaker's part that the other person did something good for them motivates the interpretation of the utterance as 'thanks'. Notably, in example (64) we have another benefactive verb, the form itadaku (the honorific counterpart of morau 'receive'). Geyer translates itadaku with the paraphrases "to receive something or a favor from a person of higher standing" (2018: 151). This description fits into the frame of 'thanks' and points out that the use of itadaku is relevant also in terms of social deixis.
In order to test the assumption that the benefactive verb itadaku and its nonhonorific counterpart morau, when used in the L1 position of su(m)imasen, constrain its use to the expression of gratitude, I searched again the JaTenTen11 for the construction -te itadaku/-te morau + su(m)imasen. Figure 1 illustrates a concordance of the first 15 lines, as observed in 100 random samples.
Because of space limitation, here I report only transcript and translation of the relevant clause as observed in Lines 1 and 2:

2.
Wazawaza okoshi itadaki sumimasen take.the.trouble come.BEN BEN.HON sumimasen *Sorry/thank you for coming all this way As expected, the use of a benefactive verb immediately at the left of su(m)imasen frames the utterance as an interactional practice with a benefactor (the receiver) and a beneficiary (the producer)two social actors prototypically involved in the speech act of thanking. This interpretation is supported by the fact that the straightforward expression of gratitude arigatō gozaimasu occurs five times in four different lines (Lines 3, 4, 8, 15) out of fifteen. As already noted, the co-occurrence of the apology-like expression su(m)imasen and straightforward expressions of gratitude reinforces the assumption that the two speech acts operate on a spectrum and that they are less distinct than they are commonly taken to be. Drawing from the distinction between similarity and proximity (Ziem and Schwerin 2018: 215-216; this terminology first appear in Wertheimer 1923: 308-309), 'apologies' and 'thanks' in Japanese may be strongly associated with each other not only because Sorry for your consideration they share a number of properties (i.e., they are similar), but also on the basis of interactants' previous knowledge that they often occur together (i.e., they are spatially and temporarily contiguous). However, on the other hand, the fact that su(m)imasen and arigatō (gozaimasu) are employed in the same clause highlights that they are not fully interchangeableor their use would just result in a redundant repetition with no additional pragmatic effect but a mild intensification of the illocutionary force. Elaborating on this idea that Japanese 'apologies' and 'thanks' are indeed related but do not map exactly onto each other, Miyake (1993: 22-23) touches upon the inappropriateness of apology-like expressions to convey gratitude in specific settings. She points out that su(m)imasen as an expression of gratitude wouldn't be used in situations where there is no futan 'burden' imposed on the recipient of the speech act. For instance, a sale assistance would use arigatō gozaimasu, and not su(m)imasen, for thanking a customer that purchased a product, since this activity can hardly be considered a burden for them.

A prototype approach to speech acts
In the previous section I presented a number of considerations supporting the idea that 'apologies' and 'thanks' in Japanese are more similar than their English counterparts, but do not map exactly onto each other. In what follows I outline an approach that, to the best of my knowledge, has not yet received much attention in the literature, but which may nevertheless be useful to better clarify the fuzziness of speech acts. This approach frames 'apologies' and 'thanks' as entities that occupy different but contiguous and somehow overlapping pragmatic spaces. The notion of pragmatic space was first proposed by Taavitsainen (2000, 2008) and is reminiscent of Hofmann's semantic field (1993: 298-299) and Simon-Vandenbergen and Aijmer's semantic map (2007: 247). In his work, Hofmann discusses the fuzziness of word meanings as follows: Most concepts do not apply only to a well-defined set of things but to a more or less fuzzy group of things with no clear boundary line between what is in the group and what is not. (1993: 297) He then adopts a lexical approach and applies this argument to words referring to concrete objects, such as chair and dog. I propose that the same is true for more abstract notions such as speech acts, though they may require an approach that goes beyond lexical features alone. What constitutes a given speech act is an issue that may also be conceptualized in terms of cognitive frames, where the term frame refers to structures capturing the typical features of the speech act (see, among others, Bednarek 2005). However, here I prefer the term space because it favors the visualization of speech acts.
Drawing from this view, I argue that we can conceptualize speech acts as multidimensional spaces containing a number of features that vary in their degree of centrality or prototypeness: the more central a given feature is, the more it is likely to be mandatory, while, on the other hand, more peripheral features are optional. 4 For instance, with reference to 'apologies', the acknowledgement that "something bad happens to someone" (Wierzbicka 1996: 532; see Section 2.2) is a central feature of the 'apology' pragmatic space in Japanese as well as in English. This is observed also in Coulmas' (1981) definition of 'apologies', where the speech act is explicitly linked to something perceived as negative by the target of the utterance (i.e., the object of regret [Coulmas 1981: 71]; see Section 2.2). On the other hand, 'admission of responsibility' is peripheral and optional because, as we have seen in Section 2.3.1, both in English and Japanese (though arguably to different degrees) one may apologize for something that is only loosely linked with their behavior. However, there are also important differences between the two linguacultures. A case in point is the sympathetic use of apology-like expression in English, a function that is not located in the Japanese pragmatic space of 'apologies' because it was not observed in my data set.
With reference to 'thanks', we may say that the acknowledgement that "someone did something good to me" (Wierzbicka 1996: 534-535), or what Coulmas refers to as the object of gratitude (1991: 74; see Section 2.2), is a mandatory element in the pragmatic space. However, while a feeling of gratitude is a central feature of 'thanks' in English, this may not be the case for 'thanks' in Japanese, where a feeling of regret and indebtedness also occupies a relevant area. We can thus assume that in Japanese the object of regret, prototypically associated in English with 'apologies' alone, is observable in 'thanks' as well and, consequently, that the two pragmatic spaces overlap to a certain degree. This assumption that in Japanese the pragmatic space of 'apologies' takes over some of the functions that in English are associated with 'thanks' seems to be confirmed by two additional findings I have touched upon in Section 2.3.1. Firstly, Japanese 'apologies' can be reactive speech acts preceded or accompanied by something beneficial for the speakera function that their English counterparts do not present because usually related with 'thanks'. Secondly, the Japanese language does not seem to provide a specific speech act label for what in English is referred to as 'thanks', which is indirectly referred to as kansha 'gratitude'. Two etymological observations can further back up the argument that Japanese 'apologies' and 'thanks' overlap to some degree. First, the Japanese word arigatō, conventionalized for the expression of 'thanks', derives from arigatai 有難い, which literally means 'hard to be/accept'. This reinforces the link between the speech act of 'thanks' in Japanese and the notion of indebtedness. Second, both kansha 感謝 'gratitude' and shazai 謝罪 'apology' present the 'apologetic' character 謝, hinting at some relationship between the two concepts at the morphological level too.
A tentative representation of the different spaces occupied by the two speech acts and their most relevant features is illustrated in Figures 2 and 3. Notably, while the Japanese pragmatic space (Figure 2) is at least partly based on empirical observations on the use of su(m)imasen in two web corpora, the English one ( Figure 3) is based exclusively on previous research and, as such, may be less robust. Nevertheless, their comparison allows us to make a number of assumptions. For instance, in the Japanese pragmatic space ( Figure 2) the object of regret encompasses both notions of 'apologies' and 'thanks' but, while it is mandatory for the former, it is optional for the latter, a difference that is visually conveyed by its marginal position alongside the boundaries of 'thanks'. This blurring of object of  regret and object of gratitude is illustrated also by their proximity in the pragmatic space. As we have seen in previous sections, in fact, in Japanese it can be difficult to distinguish them, a finding in line with the assumption that "there is a common domain where thanks and apologies are both appropriate" (Coulmas 1981: 79).
With reference to the English language, the object of regret is located exclusively in the pragmatic space of 'apologies', as reflected in their linguistic realizations. Indeed, apology-like expressions cannot be used as remedial moves for something the speaker benefitted from (i.e., to convey gratitude) and there is no standardized English construction that can convey both 'apologies' and 'thanks'. The fact that different English expressions are employed to convey 'apologies' and 'thanks' suggests that their functions are more remote from each other in the pragmatic space than their Japanese counterparts. However, and most importantly, the pragmatic spaces of 'apologies' and 'thanks' overlap in English as wellthough to a less extent than in Japanese. In fact, I argue that the norm of reciprocity and the related notion of indebtedness, so relevant for expressive speech acts, are observed cross-linguistically. One reason underlying the assumption that 'apologies' and 'thanks' have fuzzy boundaries in English as well as in Japanese is that the second pair in standardized sequences of 'thanks' and 'apology' expressions can be the same, as in the following examples: Although they are by no means the same thing, the fact that standardized responses to both 'thanks' and 'apologies' imply denying the indebtedness of the addressee strongly suggests that the two speech acts are indeed related. Finally, 'admission of responsibility' is illustrated as a peripheral, hence optional, feature in both Figures 2 and 3, because a great number of routine apologetic formulae do not imply responsibility on the speaker's part but can nevertheless be categorized as 'apologies'. This applies also to 'expression of sympathy', a function of 'apologies' strongly bounded to the context, which, as already mentioned, was observed in English but not in Japanese. Notably, 'admission of responsibility' and 'expression of sympathy' are positioned at the top and at the bottom of the English pragmatic space respectively to convey their mutual exclusiveness: expressions of sympathy, such as I'm sorry to hear that, encode that there is no responsibility for the object of regret on the side of the speaker.
The conceptualization of speech acts as entities that occupy a pragmatic space has a number of advantages. First, it shows how speech acts overlap and how they are distinguished from each other, thus favoring cross-linguistic comparison. Second, it allows for a prototype approach to speech acts, according to which, within the pragmatic space, we have a number of features that operate on a scale ranging from central (mandatory) and peripheral (optional). More central features are likely to be stable across different lingua-cultures and to have a more established association with that specific speech actbut it remains to future studies to identify, if there are any, such prototypical speech act features cross-linguistically. Last but not least, this idea that speech acts can be analyzed in relation to one other in the pragmatic space Taavitsainen 2000, 2008) allows for the visualization of speech acts, which, I argue, is potentially very helpful for problematizing English speech act labels for the description of languages other than English.
The notion of pragmatic space as proposed here presents nevertheless a number of limitations. For instance, it might be argued that bidimensional diagrams, such as those illustrated above, do not adequately allow for the multidimensionality of features that are relevant for the description of speech acts. For example, the temporality and the time frame of 'apologies' (not illustrated here because not as relevant as other factors) are also important in the realization of 'apologies' and 'thanks'. In fact, generally speaking, speech act features stand in a specific temporal relationship, with the object of regret temporarily preceding expressions of responsibility or sympathy, simply because it is a necessary condition for the acknowledgment of indebtedness on the speaker's part and, thus, for the 'apology' to be recognized as such. It should also be emphasized that there are other speech acts not mentioned here but closely related to 'apologies' and 'thanks'. This is exemplified by the finding that, in my Japanese data, apology-like expressions may introduce a request or function as attention-getting devices. It follows that the pragmatic space of, for instance, 'request' most likely overlaps to some degree with the pragmatic space of 'apology'. Bidimensional diagrams such as those shown above may thus be overlayed to include additional dimensions involved in the realization of speech acts.
As discussed at the beginning of this paper, English notions of 'apologies' and 'thanks' may be different from their counterparts in other lingua-cultures. This is an important aspect heavily overlooked in traditional SAT, where the focus is on English alone. Following previous research pointing out that the use of Japanese apology-like expressions to convey gratitude has no direct counterpart in English, the paper showed that English and Japanese speakers may conceptualize 'apologies' and 'thanks' differently. This sociopragmatic difference is mirrored in a number of pragmalinguistic differences. First, the Japanese language does not seem to have a direct counterpart for the English speech act label 'thanks', which is indirectly referred to as 'kansha' 'gratitude'. Second, in Japanese 'apologies' and 'thanks' are closely related speech acts, as confirmed by Yahoo! Chiebukuro users' metapragmatic comments on the use of su(m)imasen, framed as a lexical device that can convey both regret and gratitude. These metapragmatic considerations were tested through a collocation analysis of su(m)imasen in 100 random samples retrieved from the web corpus JaTenTen11. The analysis revealed three instances where su(m)imasen was used to symbolically compensate a favor. Although the use of su(m)imasen in contexts where English speakers would say thank you seems not very prominent, a subsequent corpus query revealed that, when immediately preceded by benefactive verbs, the thanking function of su(m)imasen takes over and was actually observed in virtually every occurrence of the JaTenTen11 corpus. It was also briefly noted that su(m)imamsen often co-occurred with the conventionalized expression of 'thanks' arigatō (gozaimasu), a pattern supporting the conceptualization of Japanese 'apologies' and 'thanks' as related activities because both similar (i.e., they share a number of features) and contiguous (i.e., in common practice they often occur together). In light of these considerations, if we want to adopt English speech act verbs to describe roughly equivalent phenomena in Japanese, we need to be aware that they may impose boundaries and constraints on speech acts that do not necessarily reflect their actual use in the Japanese lingua-culture. It follows that traditional SAT alone may not be suitable for the description of speech acts cross-linguistically, unless notions of indebtedness and reciprocity are incorporated, as they have proven to provide a very helpful toolkit for the analysis of speech acts in Japanese.
Finally, Section 5 focused on the visualization of speech acts as entities that occupy pragmatic space, an under-researched approach that has proven to be a viable tool to (attempt to) reveal the conceptual background of 'apologies' and 'thanks'. This pragmatic space concept is a useful one for a number of reasons. First, it allows the researcher to approach highly generalized and abstract categories with a more practical methodology. Second, it convincingly shows that speech acts are interlinked entities and, as such, should be investigated in relation to one another. This link between speech acts is heavily overlooked in traditional SAT, according to which "to study the speech acts of promising or apologizing we need only study sentences whose literal and correct utterance would constitute making a promise or issuing an apology" (Searle 1969: 21). As we have seen, this early conceptualization of speech acts does not reflect the variety of apology-like expressions observable in real-life communication, where prototypical 'apologies' with no additional pragmatic function but the issue of the 'apology' itself are the exception rather than the norm. Third, and related to this latter point, in the pragmatic space speech acts' features can be encoded in terms of the (proto)typical semantic features that instantiate that speech act. With reference to this, I attempted to propose some basic, prototypical aspects of 'apologies' and 'thanks' in English and Japanese, though it is beyond of the scope of the present study to assess them in any scientifically adequate way.
In terms of methodology, the co-deployment of different approaches allowed for a thorough investigation of 'apologies' and 'thanks' in Japanese. Theoretical notions of what do we mean by 'thanks' and 'apology' are indeed important for avoiding acritical use of English terminology, but they need to be complemented by a more empirical, corpus-assisted approach which, ideally, takes into account both how concrete linguistic realizations of speech acts are employed in real-life situations, and language users' evaluations of such linguistic realizations in terms of (im)politeness. Finally, quantitative and qualitative restrictions of the Yahoo! Chiebukuro corpus were hopefully at least partly mitigated by the co-deployment of a second data set, allowing for more generalizable considerations.
A number of important aspects relating to 'apologies' and 'thanks' remain as topics for future research. For instance, the extent to which the use of su(m)imasen to convey gratitude is representative of the whole JaTenTen11 corpus and of other genres is a topic that deserves focused attention. Moreover, further connections between 'apologies' and 'thanks' may emerge if we consider the response strategies appropriate for these two speech acts cross-linguistically. An analysis of English speech acts that takes into account the larger interactional exchange may support the assumption made in Section 5 that, in specific circumstances, the second pair of 'apologies' may be suitable also for 'thanks', because it conventionally involves denying the indebtedness of the speaker. This potentially applies also to Japanese, where tondemonai とんでもない 'it's ridiculous' is perceived as the appropriate response in a number of situations involving indebtedness (see Inoue and Suzuki 2019). A third limitation of the study is that it neglects nonverbal, visual elements, such as kaomoji (Japanese emoticons) and non-standard script choices, which have proven to have interpersonal functions (see, among others, Miyake 2007Miyake , 2020Backhaus 2013). For instance, and with reference to 'apologies' and 'thanks' specifically, Ohashi (2010) identifies the body movement of bowing, typically associated with 'apologies', as an essential feature of thanking episodes in spoken Japanese. Future studies may investigate whether this applies also to the bowing kaomoji 'm ( _ _ ) m,', a potentially salient feature of written CMD in Japanese.