Metarepresentational phenomena in Japanese and English: Implications for comparative linguistics

Seiji Uchida
Nara University

Abstract

Contrastive studies of languages usually focus on differences in lexical items, syntactic structures, semantic expressions, collocations, and so on. In the present paper we take a cognitive pragmatic approach, assuming that metarepresentation in the sense of Sperber (2000) ed. 2000 Metarepresentations: A Multidisciplinary Perspective. Oxford: Oxford University Press.Google Scholar and Wilson (2000)Wilson, Deirdre 2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar offers a crucial perspective in such studies. We discuss how the speech act component of higher-level explicatures is linguistically realized in Japanese and English, focusing on sentence adverbials, ‘because’ clauses, speech act particles, reported speech, private predicates, and desiderative predicates. We conclude that in the Japanese language, information concerning the speech act component tends to be linguistically realized, while such information is not necessarily realized in English. We suggest that this cognitive pragmatic approach can be applied to other languages where higher-level explicatures are basically explicit as in Japanese or implicit as in English.

Keywords:
Publication history

Date received: 21 November 2021

Date accepted: 17 January 2023

Published online: 16 May 2023

Table of contents

1.Introduction

Grice (1989)Grice, Paul H. 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press.Google Scholar distinguishes ‘what is said’ from ‘what is not said,’ referring to the latter as ‘implicature.’ However, the distinction is not so simple: ‘what is said’ and ‘what is not said’ are not actually complementary to each other. We must also take into account what is not said when we interpret ‘what is said.’ In other words, what is explicitly communicated is in fact much richer than ‘what is said.’ Sperber and Wilson (1986/1995)Sperber, Dan, and Dierdre Wilson 1986/1995 Relevance: Communication and Cognition. Oxford: Blackwell.Google Scholar coins the word ‘explicature’ on the analogy of ‘implicature’ to cover this innovative view. Explicatures are obtained through the processes of disambiguation, saturation, free enrichment, and ad hoc concept formation (Carston 2002Carston, Robyn 2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. DOI logoGoogle Scholar).

Wilson and Sperber (1993)Wilson, Dierdre, and Dan Sperber 1993 “Linguistic Form and Relevance.” Lingua 93: 1–25. DOI logoGoogle Scholar introduced ‘higher-level explicature,’ which is realized by an implicit verb phrase set higher than what is explicitly stated, i.e. the ‘basic explicature,’ mainly reflecting the speech acts performed and the speaker’s propositional attitudes. (2a–c) are possible higher-level explicatures of Mary’s utterance in (1) when it is uttered with appropriate paralinguistic features such as facial expression, gestures, and intonation.

(1)

Peter:Can you help?
Mary (sadly):I can’t.

(2)

  1. Mary says she can’t help Peter to find a job.

  2. Mary believes that she can’t help Peter to find a job.

  3. Mary regrets she can’t help Peter to find a job. (Wilson and Sperber 1993Wilson, Dierdre, and Dan Sperber 1993 “Linguistic Form and Relevance.” Lingua 93: 1–25. DOI logoGoogle Scholar)

‘Say’ in (2a) indicates a speech act, while ‘believe’ in (2b) and ‘regret’ in (2c) imply Mary’s attitudes towards the proposition ‘she can’t help Peter to find a job.’

Metarepresentational aspects of communication is one of the major topics in relevance theory. It can be said that the discussion started in Sperber (1994)Sperber, Dan 1994 “Understanding Verbal Understanding.” In What is Intelligence? ed. by Jean Khalfa, 179–198. Cambridge: Cambridge University Press.Google Scholar, followed by Sperber (2000) 2000 “Metarepresentations in an Evolutionary Perspective.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 117–137. Oxford: Oxford University Press.Google Scholar, Wilson (2000)Wilson, Deirdre 2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar, Noh (2000) 2000 Metarepresentation. Amsterdam: John Benjamins. DOI logoGoogle Scholar, Sperber and Wilson (2002) 2002 “Pragmatics, Modularity and Mindreading.” Mind and Language 17: 3–23. DOI logoGoogle Scholar, Uchida and Noh (2018)Uchida, Seiji, and Eun-Ju Noh 2018 “Metarepresentational Phenomena in Japanese and Korean.” Memoirs of Nara University 46: 1–23.Google Scholar, and Noh (2021) 2021 “On Linguistic Communication Based on Resemblance in Form.” Journal of Pragmatics 186: 20–32. DOI logoGoogle Scholar, among others. Wilson (2000)Wilson, Deirdre 2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar notes that a metarepresentation is “a representation of a representation: a higher-order representation with a lower-order representation embedded within it.” If the original representation is the speaker’s own thought, it is a case of first-order metarepresentation, and if it is someone else’s representation, it is a second-order metarepresentation (Noh 2021 2021 “On Linguistic Communication Based on Resemblance in Form.” Journal of Pragmatics 186: 20–32. DOI logoGoogle Scholar).

Linguistic phenomena of higher-level explicatures and metarepresentations are, as it were, two sides of the same coin: the former is concerned with what the speaker communicates and the latter with what the addressee interprets. That is to say, the addressee tries to entertain what the speaker intends to communicate from the basic and higher-level explicatures. For example, if Peter processes Mary’s utterance in (1), he may interpret it as conveying one of the candidates in (2). It might be the case that not only the basic explicature “Mary can’t help Peter to find a job” but also “Mary says” in (2a), “Mary believes” in (2b), and “Mary regrets” in (2c) are metarepresented in Peter’s mind, reflecting higher-level explicatures of Mary’s utterance. They can be paraphrased as follows:

(3)

  1. Peter:[Mary says [Mary can’t help Peter to find a job]]

  2. Peter:[Mary believes [Mary can’t help Peter to find a job]]

  3. Peter:[Mary regrets [Mary can’t help Peter to find a job]]11.As noted by one of the reviewers, the paraphrases in (3) are examples of ‘second-order’ metarepresentaion in the sense of Wilson (2000)Wilson, Deirdre 2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar and Noh (2000) 2000 Metarepresentation. Amsterdam: John Benjamins. DOI logoGoogle Scholar. Metarepresentations, whether first-order or second-order, are linguistically related in some way in languages like Japanese, as we discuss below. We mention these two types of metareprentation in Section 3, where information from metarepresentation is crucial.

(3a) shows that Peter metarepresents what Mary says, and (3b) and (3c) are Peter’s metarepresentations of what she believes and what she regrets respectively. According to Sperber (2000 ed. 2000 Metarepresentations: A Multidisciplinary Perspective. Oxford: Oxford University Press.Google Scholar, 3), (3a) to (3c) are examples of ‘mental representation of public representation,’ while if Peter said ‘Mary said she can’t help,’ it would exemplify an example of “public representations of public representations.”

The present paper primarily focuses on the descriptions of metarepresentations in utterances and discusses how such information is linguistically realized in Japanese and in English. The discussion leads to a conclusion that metarepresentational perspectives are an encouraging approach to comparative studies in languages.

In the following section we show how higher-level explicatures or metarepresentation processes are reflected in Japanese and in English. Section 3 discusses cases involving inflection in Japanese. Section 4 is the conclusion.

2.Realizations of metarepresentations

We saw in (1) above that paralinguistic features may provide us with information on various higher-level explicatures and that the metarepresentations can be shown as in (3). There are, of course, linguistic devices that perform the same functions as those features. In this section we observe such linguistic facts both in Japanese and in English.

2.1Metarepresentational adverbials

There are some sentence adverbs which are related to the realization of basic or higher-level explicatures.22.See Wilson and Sperber (1993)Wilson, Dierdre, and Dan Sperber 1993 “Linguistic Form and Relevance.” Lingua 93: 1–25. DOI logoGoogle Scholar, Ifantidou (2001Ifantidou, Elly 2001 Evidentials and Relevance. Amsterdam: John Benjamins. DOI logoGoogle Scholar, 97–118), and Carston (2002Carston, Robyn 2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. DOI logoGoogle Scholar, 120–125) for relevance theoretic approaches to sentence adverbs. Let us call such adverbs ‘metarepresentational adverbials’ if metarepresentations are implicitly or explicitly incorporated in them.

Consider the following utterances with -ly adverbs:

(4)

  1. Mary:Generally, French wine is superior to Australian wine.

  2. Mary:Technically, French wine is superior to Australian wine.

(5)

  1. Mary:Frankly, French wine is superior to Australian wine.

  2. Mary:Honestly, French wine is superior to Australian wine.

The basic and higher-level explicatures of (4) and (5) are seen in (6) and (7) respectively:

(6)

  1. [Mary says from a general point of view [French wine is superior to Australian wine]]

  2. [Mary says from a technical point of view [French wine is superior to Australian wine]]

(7)

  1. [Mary says frankly [French wine is superior to Australian wine]]

  2. [Mary says honestly [French wine is superior to Australian wine]]

It can be seen from (6) and (7) that those sentence adverbs, generally, technically, frankly, and honestly, behave in the same way in that they modify the higher-level explicature verb say, but notice that (6) differs from (7) in quality. That is to say, generally and technically are concerned with the speech act aspect of higher-level explicatures and frankly and honestly with the propositional attitude, which can be confirmed from the fact that it is possible to paraphrase the utterances in (5) into (8) but not from (4) into (9).

(8)

  1. Mary:To be frank, French wine is superior to Australian wine.

  2. Mary:To be honest, French wine is superior to Australian wine.

(9)

  1. Mary: *To be general, French wine is superior to Australian wine.

  2. Mary: *To be technical, French wine is superior to Australian wine.

If we add the phrase ‘with you’ to the phrases in (8), it becomes even more evident that the utterances reveal propositional attitudes towards the hearer.

(10)

  1. Mary:To be frank with you, French wine is superior to Australian wine.

  2. Mary:To be honest with you, French wine is superior to Australian wine.

If we translate (4) and (5) into Japanese focusing on the -ly adverbs, we get (11) and (12).

(11)

  1. Mary: Ippantekini iu-to, French wine is superior to Australian wine.generally say-Quot

  2. Mary: Senmontekini iu-to, French wine is superior to Australian wine.technically say-Quot

(12)

  1. Mary: Socchokuni iu-to, French wine is superior to Australian wine.frankly say-Quot

  2. Mary: Shoujikini iu-to, French wine is superior to Australian wine.honestly say-Quot

Note that speaking can be inserted in English, as in ‘Generally speaking’ and ‘Frankly speaking’ while, on the other hand, iu-to, which corresponds to the speech act ‘saying that …’, is required in Japanese. (13a) to (13d) are therefore not possible.

(13)

  1. Mary: * Ippantekini, French wine is superior to Australian wine.

  2. Mary: * Senmontekini, French wine is superior to Australian wine.

  3. Mary: * Socchokuni, French wine is superior to Australian wine.

  4. Mary: * Shoujikini, French wine is superior to Australian wine.

In some cases, ‘-ly adverbs + speaking’ can be expressed by other constructions in English, as in (14):

(14)

  1. In other words, French wine is superior to Australian wine.

  2. In short, French wine is superior to Australian wine.

  3. To conclude, French wine is superior to Australian wine.

These sentence initial adverbials can be put into Japanese as follows:

(15)

  1. Iikaeru-to, French wine is superior to Australian wine.

    say -Quot (in another way)

  2. Kanketsuni iu-to, French wine is superior to Australian wine.

    briefly say-Quot

  3. Ketsuron-o iu-to, French wine is superior to Australian wine.

    conclusion-Obj say-Quot

‘In other words,’ ‘in short,’ and ‘to conclude’ do not reflect the speech act component on the surface, but they clearly contain ‘saying’ in the higher-level explicatures.33. Blakemore (1996) 1996 “Are Apposition Markers Discourse Markers?” Journal of Linguistics 32: 325–347. DOI logoGoogle Scholar takes the same standpoint in that ‘in other words’ and ‘in short’ put constraints on higher-level explicatures as reformation markers. We can trace back the process in iikaeru (=say in another way) and iu (=say) in (15), while in the English counterparts of (14) there are no linguistic markers to suggest speech act information.

The following are parallel constructions, where the speech act component is crucial for interpretation (Uchida and Noh 2018Uchida, Seiji, and Eun-Ju Noh 2018 “Metarepresentational Phenomena in Japanese and Korean.” Memoirs of Nara University 46: 1–23.Google Scholar):

(16)

  1. Mary to Peter:John was killed last night, in case you haven’t heard.

  2. Mary to Peter:Since you know so much, why is Paul leaving?44. Sweetser (1990)Sweetser, Eve 1990 From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar classifies ‘because’ clauses into three types; content, epistemic, and speech-act types. The use of the ‘since’ clause of (16b) corresponds to the speech act domain. cf. footnote 5.

  3. Mary to Peter:(The door bell is ringing.) If that’s John, I’m not here. (Noh 1998 1998 “Echo Questions: Metarepresentation and Pragmatic Enrichment.” Linguistics and Philosophy 21 (6): 603–628. DOI logoGoogle Scholar)

  4. Mary to Peter:Jill was pretty rude to me. I am neglecting my job! (Wilson 2000Wilson, Deirdre 2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar)

On the surface there seem to be no logical connections between the two clauses in (16a) to (16d). We have to infer the information in order to bridge them. In (16a) Mary tells Peter the news presuming that he doesn’t already know. (16b) suggests that Mary asks Peter assuming he knows the reason why Paul is leaving. In (16c) Mary asks Peter to tell the visitor that she is away if that person is John. (16d) is a kind of reported speech, suggesting that Jill is criticizing Mary. The verbs ‘tell,’ ‘ask,’ and ‘suggest’ are not realized in (16) but clearly communicated as the speech act component of higher-level explicatures.

These processes of interpretation are intuitively clear to native speakers of English, but Japanese people may find it quite difficult to understand the meaning of these utterances. If we translate the utterances in (16) into Japanese, we get those in (17), where the italicized parts are required in Japanese.

(17)

  1. In case you haven’t heard ii-masu ga (=I tell you), John has left.

  2. Since you know so much kiki-masu ga (=I ask you), why is Paul leaving?

  3. (The door bell is ringing.) If that’s John, I’m not here to itte kudasai (=tell him).

  4. Jill was pretty rude to me. I am neglecting my job to iunda (=she says)!

The inserted Japanese phrases reflect the higher-level explicatures of the related clauses in (16) and the utterances sound unnatural without them.

We observed above that higher-level explicatures are reflected in some metarepresentational adverbials and that information concerning propositional attitudes is linguistically realized both in English and Japanese, while information implying speech acts should be linguistically realized in Japanese but not necessarily in English.

2.2‘…, because …’ construction

As is often pointed out, there are two types of ‘because’ clauses following main clauses: those with a comma and those without, as in (18):

(18)

  1. Mary:Peter couldn’t walk straight because he was drunk.

  2. Mary:Peter was drunk, because he couldn’t walk straight.55.According to Sweetser (1990Sweetser, Eve 1990 From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar, 76–78), (18a) and (18b) are cases of the content domain and the epistemic domain respectively, and Rutherford (1970)Rutherford, William E. 1970 “Some Observations Concerning Subordinate Clauses in English.” Language 46 (1): 97–115. DOI logoGoogle Scholar calls the ‘because’ clause of (18a) the restrictive use and that of (18b) the non-restrictive. However, neither Sweetser nor Rutherford mention the speech act performed by saying ‘Peter was drunk.’

In (18a), his being drunk caused Peter not to walk straight, while in (18b) knowing that he couldn’t walk straight caused Mary to conclude that Peter was drunk.

With no comma as in (18a) ‘because’ clauses are subordinated to the main verb clauses and internally processed in the utterance. In ‘comma+because’ utterances like (18b), the speaker’s assertion or conclusion is presented first and the reason for saying or thinking so follows. Wilson (2000)Wilson, Deirdre 2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar mentions a similar example when discussing cases where higher-level metarepresentations need not be linguistically marked.

(19)

It’s raining, because the grass is wet.

She states that the causal relation is found between a state of affairs and an utterance or thought in (19), where the hearer “must construct a higher-order representation of the type ‘she says’, or ‘she thinks’, and attribute it as part of the speaker’s meaning.” (Wilson 2000Wilson, Deirdre 2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar, 431) (19) corresponds to (18b) in our analysis and both ‘because’ clauses mean that ‘the reason for saying or thinking so is as follows.’

There are typical translation patterns in Japanese for those two types of ‘because’ clause: (…) (da)kara/(na)node … (da) for the former and (…) to-iu-nowa … (kara)-da for the latter. With this in mind, the utterances in (18) would be translated as in (20).

(20)

  1. Mary:
    Peter
    Peter
    wa
    Top
    yopparatte
    drunk
    ita –node
    Pst because
    massugu
    straight
    arukenaka-ttan
    couldn’t walk
    da.
    Part

  2. Mary:
    Peter
    Peter
    wa
    Top
    yopparatte
    drunk
    itan
    Pst
    da,
    Part
    to-iu-nowa
    reason for saying that
    massugu
    straight
    aruke
    walk
    nakatta
    couldn’t
    kara(-da).
    because

Node in (20a) remains within the scope of the clause ‘he couldn’t walk straight’ and literally equals ‘since’ or ‘because.’ On the other hand, in (18b) there is no counterpart to to-iu (=to say) in (20b). The complement of to-iu in (20b) is that Peter was drunk and the ‘because’ clause is supposed to be an explanation for what the speaker said. Therefore, the metarepresentations of (20a) and (20b) could be something like (21a) and (21b), respectively.

(21)

  1. [Mary says [Peter couldn’t walk straight because he was drunk]]

  2. [Mary says [Peter was drunk] because he couldn’t walk straight]

It is clear from the paraphrases of (21) that the use of ‘because’ in (18b) explains the reason why the speaker said ‘he was drunk.’ In other words, in (18a) the ‘because’ clause refers to the main verb phrase and in (18b) to the speech act component of the higher-level explicature ‘Mary says.’

The analysis discussed above may remind us of the ‘performative hypothesis’ proposed by Ross (1970)Ross, John Robert 1970 “On Declarative Sentences.” In Readings in English Transformational Grammar, ed. by Roderick A. Jacobs, and Peter S. Rosenbaum, 222–272. Waltham, MA: Ginn and Company.Google Scholar. Indeed, there are linguistic data evoking ‘performative analysis’ and appealing to English native speakers’ intuition, but one of the crucial defects of the hypothesis is that it was purely a syntactic analysis. As we saw, the complex mechanism involved is basically pragmatic rather than syntactic, and there are in fact a number of advantages to taking a pragmatic approach.66.Ross (1970Ross, John Robert 1970 “On Declarative Sentences.” In Readings in English Transformational Grammar, ed. by Roderick A. Jacobs, and Peter S. Rosenbaum, 222–272. Waltham, MA: Ginn and Company.Google Scholar, 254–258) does mention ‘pragmatic analysis’ as an alternative to the performative analysis and concludes that the pragmatic analysis would be a notational variant of his performative analysis. In his view of pragmatics, the utterance ‘Prices slumped’ has a ‘deep’ structure such as [Prices slumped] instead of [I say to you prices slumped], but, as we have seen, one of the explicatures of ‘Prices slumped’ would be [The speaker says [prices slumped]], where the speech act component is explicitly stated and is not to be deleted. The main reason that Ross didn’t give value to pragmatics at that time might be that pragmatics was not as cognitively oriented as it is today. See the explicature analysis proposed in Uchida (2011 2011 Goyoron no Shatei [The Scope of Pragmatics]. Tokyo: Kenkyusha.Google Scholar, 145–158). See Levinson (1983Levinson, Stephen 1983 Pragmatics. Cambridge, MA: Cambridge University Press. DOI logoGoogle Scholar, 246–263) for a detailed description of the performative analysis.

2.3Linguistic realizations of speech acts

Japanese sentence final particles have invited considerable discussion, but thus far, no one has proposed a metarepresentational approach. Let us focus on differences in realizing higher-level explicatures between Japanese and English concerning speech acts of utterances. First, take the simple English sentence in (22):

(22)

He is a linguist.

Sentence (22) could be a response to a number of questions and statements, including the following:

(23)

  1. Who is a linguist?

  2. What does Tom do?

  3. Tom is a lawyer.

  4. What did Jane say?

When (22) is a response to (23a), ‘he’ is to be stressed, while on the other hand, ‘linguist’ will receive the major stress if it is meant to be a reply to (23b). If (22) is uttered as a refutation of (23c), the contrastive stress will fall on ‘linguist.’ If it is a response to (23d), it is a report of what Jane said. (Reported speech is to be discussed in Section 3.1.) That is, prosody will be different depending on the situation, but the linguistic form ‘He is a linguist’ remains the same.

The Japanese responses corresponding to (23a) to (23d) would be (24a) to (24d), respectively.

(24)
  1. Kare
    he
    ga
    Nom77. Ga and wa are both nominative case-marking particles and wa also works as a topic marker.
    gengo gakusha
    linguist
    da.
    is

    ‘He is a linguist.’

  2. Kare
    he
    wa
    Top
    gengo gakusha
    linguist
    da.
    is

    ‘He is a linguist.’

  3. Kare
    he
    wa
    Top
    gengo gakusha
    linguist
    da-yo.
    is-Part

    ‘He is a linguist.’

  4. Kare
    he
    wa
    Top
    gengo
    linguist
    gakusha
    is
    da-tte.
    -(hear-say) Part

    ‘He is a linguist.’

Ga in (24a) implies that kare (=he) is new information while wa in (24b) introduces a topic in the sense of Kuno (1973)Kuno, Susumu 1973 The Structure of the Japanese Language. Cambridge, MA: MIT Press.Google Scholar. Da in (24a) to (24d) is a sentence final particle indicating the speaker’s judgement.88. Da and noda are both sentence final particles, but da follows after nouns and noda comes at the end of predicates. Yo in (24c) typically conveys new information, and tte in (24d) is an informal version of to itta (=said that …), marking the rest of the sentence as information reported by someone else.

The point here is that a single linguistic form like (22) can be expressed by at least four patterns in Japanese. In other words, the basic explicature ‘he is a linguist’ is the same in both languages but prosodic information will be added in English while case-marking particles and sentence final particles are required in Japanese.

Sentence final particles in Japanese can allow for manifest speech acts to be conveyed (Uchida 2013 2013 Kotoba-wo Yomu, Kokoro-wo Yomu: Ninchi Goyoron Nyumon [Reading Words, Reading Mind: Introduction to Cognitive Pragmatics]. Tokyo: Kaitakusha.Google Scholar, 95–96). Suppose that Mary says (25) to Peter.

(25)

Tom is coming toward us.

Utterance (25) can convey a number of speech acts. Some of them are given in (26) below (Uchida and Noh 2018Uchida, Seiji, and Eun-Ju Noh 2018 “Metarepresentational Phenomena in Japanese and Korean.” Memoirs of Nara University 46: 1–23.Google Scholar).

(26)

  1. [Mary informs [Tom is coming toward us]] (informing)

  2. [Mary warns [Tom is coming toward us]] (warning)

  3. [Mary confirms [Tom is coming toward us]] (confirmation)

  4. [Mary expects [Tom is coming toward us]] (expectation)

In English explicit linguistic devices suggesting each speech act are not necessarily required. Those different speech acts can be reflected in sentence final particles in Japanese, as in (27):

(27)

  1. Tom is coming toward us -yo.

  2. Tom is coming toward us -zo.

  3. Tom is coming toward us -ne.

  4. Tom is coming toward us -na.

Yo in (27a) tells Peter that Tom is coming while zo, typically used by men, warns him that Tom is coming. Ne in (27c) is often used to confirm information and na in (27d) suggests that the speaker expects something to happen.

Again, we can say that what is linguistically realized in English is the basic explicature of (25) as discussed in the case of (22), but in Japanese those speech acts of higher-level explicatures in English must be expressed with various sentence final particles.

An interesting phenomenon occurs in the use of ‘why’ (cf. Uchida 2011 2011 Goyoron no Shatei [The Scope of Pragmatics]. Tokyo: Kenkyusha.Google Scholar, 72–81). Consider the following exchanges.

(28)

Mary:I believe him.

Peter:Why?

(29)

Mary:I don’t believe him.

Peter:Why not?

It is logical that ‘Why?’ in (28) means ‘Why do you believe him?’ and ‘Why not?’ in (29) ‘Why don’t you believe him?’ Let us call these responses Verb Phrase Agreement (henceforth VPA).

Note, however, that there is another way to respond to ‘I don’t believe him.’

(30)

Mary:I don’t believe him.

Peter:Why?

‘Why?’ here is to be interpreted as ‘Why do you say so (=Why do you say that you don’t believe him)?’ We can call the response Speech Act Agreement (henceforth SAA) since it replies to what Mary just said. Consider the differences in interpretation by positing the following explicatures.

(31)

  1. [Mary says [Mary believes him]]

  2. [Mary says [Mary doesn’t believe him]]

It is clear that VPA is concerned with the basic explicatures [Mary believes him] in (31a) and [Mary doesn’t believe him] in (31b), while SAA is concerned with the higher-level explicature [Mary says].

If this analysis is correct, ‘Why?’ in (28) can be processed in the same way. That is, ‘Why?’ in (28) is ambiguous between VPA and SAA. When we put (30) into Japanese, two types of translation result, as in (32) and (33):

(32)

Mary:I don’t believe him.

Peter:
Do shite? / Naze ?
why
‘Why don’t you believe him?’ / ‘Why do you say that?’

(33)

Mary:I don’t believe him.

Peter:
Do shite
why
so
that
iu
say
no?
Part
/
 
Naze
why
so
that
iu
say
no?
Part
‘Why do you say that?’

The literal translation of ‘Why?’ is Do shite? or Naze? and these question forms can cover the two ways of agreement, VPA and SAA. On the other hand, Do shite so iu no? and Naze so iu no? are regarded as an SAA response to the utterance ‘I don’t believe him.’

Of interest is a parallel linguistic fact concerning VPA and SAA. Consider the exchanges in (34) and (35).

(34)

Mary:People make mistakes.

Peter:(Yes,) they do. / (No,) they don’t.

(35)

Mary:He didn’t say anything.

Peter:(No,) he didn’t. / (Yes,) he did.

The utterances ‘Yes, they do’ and ‘No, they don’t’ in (34) and ‘No, he didn’t’ and ‘Yes, he did’ in (35) respond to ‘People make mistakes’ and ‘He didn’t say anything’ respectively. It is VPA in that ‘do’ and ‘don’t’ are used to refer to the verb phrase ‘make mistakes.’ We can also receive the following response to both (34) and (35).

(36)

(Yes,) you’re right.

In (36) the speaker means that ‘what you’ve said is right.’ That is, Peter agrees with what is asserted in ‘People make mistakes’ and ‘He didn’t say anything.’ When the first utterance is an affirmative like (34), the interpretation of ‘Yes, you’re right’ semantically appears to be the same as the VPA of ‘Yes, they do,’ while the speech act interpretation of ‘Yes, you’re right’ would be quite different from the VPA when the preceding utterance is a negative, as in (35).

These two different interpretations can be explained in the same way as above. We can posit the candidates of higher-level of explicatures of Mary’s utterances in (34) and (35) as follows:

(37)

  1. [Mary says/asserts [People make mistakes]]

  2. [Mary says/asserts [He didn’t say anything.]]

It is clear that VPA corresponds to the basic explicatures [People make mistakes] in (37a) and [He didn’t say anything] in (37b), and SAA to [Mary says/asserts] in both (37a) and (37b).

In English there is no overt linguistic marker to distinguish between the two uses of agreement, especially when the preceding part is expressed in the affirmative. Recall ‘you’re right’ in (36), where the utterance can be used in both ways. In Japanese, on the other hand, we have several phrases to specify the two different types of agreement. Consider the following exchange:

(38)

Mary:He didn’t say anything.

Peter:
So. / So 99.This so here is a pronoun which refers to what is anaphorically mentioned and is quite different from -soo as an evidential marker, which Narrog and Yang (2018)Narrog, Heiko, and Wenjiang Yang 2018 “Evidentiality in Japanese.” In The Oxford Handbook of Evidentiality, ed. by Alexandre Y. Aikhenvald, 709–724. Oxford: Oxford University Press. DOI logoGoogle Scholar discusses.
really
da-ne.
was -Part
‘Didn’t he? / He didn’t.’

So in (38) refers to the utterance ‘He didn’t say anything.’ To put it another way, so works to indiccate grammatical agreement and cannot be used as a counterpart of ‘Yes, you’re right.’ Ossharu tori, on the other hand, would be used instead of So / So da-ne to indicate speech act agreement.

(39)

Mary:He didn’t say anything.

Peter:
Ossharu
say (Polite)
tori.
right
‘What you say is right.’

Ossharu tori implies that the speaker agrees with what the other party says. In other words, it responds to the higher-level explicature [The speaker says/asserts].

We also find similar phenomena in some uses of adverbs in English, as in (40) below:

(40)

Mary:This novel is worth reading.

Peter:Certainly.

Peter’s ‘Certainly’ in (40) is ambiguous between (41a) and (41b).

(41)

  1. This novel is certainly worth reading.

  2. Certainly you’re right.

(41a) agrees grammatically with Mary’s utterance in (40) and (41b) corresponds to the speech act of asserting of the utterance. However, if the first utterance is a negative sentence like (42) and Peter responds to Mary by saying ‘Certainly’ or ‘Certainly not,’ there is no ambiguity.

(42)

Mary:This novel isn’t worth reading.

Peter:Certainly.

(43)

Mary:This novel isn’t worth reading.

Peter:Certainly not.

That is to say, ‘Certainly’ in (42) and ‘Certainly not’ in (43) mean (44) and (45), respectively.

(44)

Certainly you’re right.

(45)

Certainly the novel isn’t worth reading.

Those behaviors of ‘certainly’ can be nicely accommodated in the explanations of VPA and SAA discussed above: i.e. (41a) and (45) are cases of VPA and (41b) and (44) those of SAA. Adverbs which behave like ‘certainly’ include absolutely, definitely, maybe, perhaps, probably, and so on.

We have seen in this subsection that Japanese sentence final particles are responsible for the speech acts performed which can be shown in the higher-level-explicatures of the utterances. In English, on the other hand, there is no counterpart that performs the same functions.

3.Realizations of first- and second-order metarepresentations

When we feel something or we want to do something, the original state of affairs is crucially involved. In other words, first-order metareprentation is involved when the original state of affairs comes from the speaker herself/himself, while if another person tries to convey someone else’s first-order metarepresentation it is a case of second-order metarepresentation. Depending on whether it is private or public, different linguistic realizations may result. We discuss such intrinsic and interesting linguistic behaviors in Japanese below.

3.1Reported speech

Reported speech is typical evidence of metarepresentation in that both first-order and second-order metarepresentations are basically involved, where the speaker attributes the source of the information to someone else. Suppose that Peter says ‘Jane loves Tom’ as a reply to Mary in (46).

(46)

Mary:What did Tom say?

Peter:Jane loves Tom.

There is no problem in (46), but consider a literal translation to Japanese. (47) is not an utterance which is equivalent to ‘Jane loves Tom’ intended as a reply to Mary in (46).

(47)
? Jane
Jane
wa
Top
Tom
Tom
wo
Ac
ai-shiteru.
love

‘Jane loves Tom.’

A reporting verb is required there, as in (48):

(48)

  1. Jane loves Tom -to itta.

  2. Jane loves Tom -tte.

    ‘Jane said she loves Tom.’

‘Jane said’ in ‘Jane said she loves you’ is a part of the basic explicature of (48). This is yet another example of the linguistic fact that there exists a characteristic difference in realizing basic and higher-level explicatures between English and Japanese.

Based on this observation, we find that ‘You’ve dropped your purse’ in (49) can be ambiguous, with both (50a) and (50b) as possible interpretations.

(49)

Mary:What did Jill say?

Peter:You’ve dropped your purse.

(50)

  1. Jill said you (=Mary) have dropped your purse.1010.Yet another interpretation, ‘Jill said you (=Peter) have dropped your purse,’ is also possible, but (50a) is a more likely interpretation.

  2. You (=Mary) have dropped your purse.

Possible Japanese translations corresponding to (50a) and (50b) would be (51a) and (51b) respectively. This can be shown as in (52).

(51)
  1. Kimino
    your
    saifu
    purse
    ga
    Nom
    ochita
    dropped
    to
    Quot
    itta / ochita-tte.
    said
  2. (Kimino)
    your
    saifu
    purse
    ga
    Nom
    ochita
    dropped
    yo.
    Part
(52)

  1. You’ve dropped your purse -to itta / -tte.

  2. You’ve dropped your purse -yo.

Notice that to itta and -tte in (52a) and yo in (52b) are attached to the original English sentence. To itta consists of the quotative particle to and the verb iu (=‘say’) and -tte is a hearsay particle, so that both to itta and -tte tell us that the preceding section is being reported. Yo indicates that the speaker is telling the hearer information that he believes is unknown to the hearer at the time of the utterance.

Let us look at (49) to (52) above from the viewpoint of metarepresentation. The two interpretations of the ambiguous utterance ‘You’ve dropped your purse’ in (49) can be shown as (53a) and (53b).

(53)

Peter:You’ve dropped your purse.

  1. [Peter says that [Jill says Mary has dropped Mary’s purse.]]

  2. [Peter says that [Mary has dropped Mary’s purse.]]

One of the interpretations, (50a), is given in (53a), incorporating the reporting verb phrase ‘Jill says.’ The other interpretation, (50b), would be (53b), which can be uttered in the case where Peter, the speaker, notices that Mary has just dropped her purse, leaving Mary’s question unanswered. In contrast, the Japanese counterparts explicitly encode each piece of information, as in (54) and (55):

(54)

  1. You’ve dropped your purse -tte / to itta.

  2. [Peter says [Jill says Mary has dropped Mary’s purse]].

(55)

  1. You’ve dropped your purse -yo.

  2. [Peter is telling Mary [Mary has dropped Mary’s purse]]].

As mentioned above, -tte or to itta is a reporting verb meaning ‘say’ and sentence final particle yo implies the transmission of new information and can be paraphrased by ‘I’m telling you.’ These items are obligatory in Japanese and (54a) and (55a) would sound odd if they were deleted.

Similar behaviors are observed in echo questions.1111.See Blakemore (1994)Blakemore, Diane 1994 “Echo Questions: A Pragmatic Account.” Lingua 94: 197–211. DOI logoGoogle Scholar and Noh (1995Noh, Eun-Ju 1995 “A Pragmatic Approach to Echo Questions.” UCL Working Papers in Linguistics 7: 107–140.Google Scholar, 1998 1998 “Echo Questions: Metarepresentation and Pragmatic Enrichment.” Linguistics and Philosophy 21 (6): 603–628. DOI logoGoogle Scholar, 2000 2000 Metarepresentation. Amsterdam: John Benjamins. DOI logoGoogle Scholar) for relevance theoretic approaches to echo questions. Noh (1995)Noh, Eun-Ju 1995 “A Pragmatic Approach to Echo Questions.” UCL Working Papers in Linguistics 7: 107–140.Google Scholar mentions that Korean and Japanese echo questions behave similarly. Iwata (2003)Iwata, Seizi 2003 “Echo Questions.” Linguistics and Philosophy 26: 185–254. DOI logoGoogle Scholar discusses an alternative analysis. Take (56) and (57) as examples.

(56)

Mary:You startled me!

Peter:Startled you?

(57)

Mary:He came by to see me.

Peter:He went to see you?

Peter echoes Mary’s utterances with deictic items modified accordingly, meaning ‘Did you say I startled you?’ and ‘Did you say he went to see you?’ in (56) and (57) respectively. Here again, the reporting clause ‘did you say’ is not necessarily realized. How about the Japanese counterparts?

(58)

Peter:
Kimi
you
wo
Ac
odorokaseta -tte / to iuno?
startled (you) Quot say
‘Did you say (I) startled you?’

(59)

Peter:
Kare
he
ga
Nom
kimi
you
ni
Obj
aini
came
kita
to
-tte/to
see (you)
iuno?
Quot to say
‘Did you say he came to see you?’

With appropriate rising intonation, the literal translations of Peter’s utterances in (56) and (57) are not impossible, but those with the reporting clause -tte or to iuno as in (58) and (59) sound much more natural. Metarepresentational information concerning basic or higher-level explicatures involved in echo questions is most likely to be expressed linguistically in Japanese.

So far we have pointed out that in reported speech the speech act component of higher-level explicatures tends to be explicitly realized in Japanese but not always in English. In the following section we consider a slightly more complicated linguistic fact related to inflection in Japanese.

3.2Metarepresentational information in inflection in Japanese

In this subsection we consider the fact that we can trace the source of information in inflection of some adjectives and verbs in Japanese and, furthermore, that the Japanese language has linguistic devices of inflection to convey metarepresentational information.

3.2.1Private predicates

Inner feelings or sensations such as joy, happiness, sorrow, grief, heat, cold, pain, and so on, can only be perceived by the individuals experiencing those sensations. That is to say, it is not possible for the speaker to describe the deeply rooted feeling of others, asserting that someone else is sad, hot or cold. But consider (60) below, where the experiencers are Sam in (60a) and Tom in (60b).

(60)

  1. Sam:I’m sad.

  2. Sam:Tom is sad.

In (60a) it is the speaker ‘I’ who experiences the emotion, but in (60b) Sam says that Tom is sad, not himself. The same predicate ‘be sad’ is used in these two cases.

In Japanese, on the other hand, we have to use different predications based on the subjects.

(61)

  1. Sam:
    Watashi
    I
    wa
    Top
    kana-shii.
    sad
    ‘I’m sad.’

  2. Sam:
    Tom
    Tom
    wa
    Top
    kana-shigatte iru.
    sad
    ‘Tom is sad.’

Kana-shii occurs with the first person subject and kana-shigatte iru with the third person subject. We cannot switch the forms, as seen below.

(62)

  1. Sam: * Watashi wa kana-shigatte iru.

  2. Sam: *Tom wa kana-shii.1212.This utterance is possible in the context of fictional discourse. See Kuroda (1973)Kuroda, S.-Y. 1973 “Where Epistemology, Style, and Grammar Meet: A Case Study from Japanese.” In A Festschrift for Morris Halle, ed. by Stephen R. Anderson, and Paul Kiparsky, 377–391. New York: Holt, Rinehart and Winston.Google Scholar and Uchida (2013 2013 Kotoba-wo Yomu, Kokoro-wo Yomu: Ninchi Goyoron Nyumon [Reading Words, Reading Mind: Introduction to Cognitive Pragmatics]. Tokyo: Kaitakusha.Google Scholar, 142–144). If no-da (sentence final particle) or chigainai (=epistemic ‘must’) follows, kana-shii is also possible, where noda functions as a kind of marker of interpretive use in relevance theory: Sam commits to the truth of the contents of the proposition that Tom is sad (cf. Uchida 1998Uchida, Seiji 1998 “ ‘(No)da’ – Kanrensei Riron karano Shiten [‘Noda’ in Relevance Theory].” In Gendai Eigo no Goho to Bunpo [English Usage and Grammar], ed. by the Editorial Board, 243–251. Tokyo: Taishukan.Google Scholar).

This particular linguistic fact has been mentioned in Kuroda (1973)Kuroda, S.-Y. 1973 “Where Epistemology, Style, and Grammar Meet: A Case Study from Japanese.” In A Festschrift for Morris Halle, ed. by Stephen R. Anderson, and Paul Kiparsky, 377–391. New York: Holt, Rinehart and Winston.Google Scholar, Kuno (1973)Kuno, Susumu 1973 The Structure of the Japanese Language. Cambridge, MA: MIT Press.Google Scholar, Aoki (1986)Aoki, Haruo 1986 “Evidentials in Japanese.” In Evidentiality: The Linguistic Coding of Epistemology, ed. by Wallace Chafe, and Johanna Nichols, 223–238. Norwood, NJ: Ablex.Google Scholar, and Narrog and Yang (2018)Narrog, Heiko, and Wenjiang Yang 2018 “Evidentiality in Japanese.” In The Oxford Handbook of Evidentiality, ed. by Alexandre Y. Aikhenvald, 709–724. Oxford: Oxford University Press. DOI logoGoogle Scholar, among others. Let us consider Aoki (1986)Aoki, Haruo 1986 “Evidentials in Japanese.” In Evidentiality: The Linguistic Coding of Epistemology, ed. by Wallace Chafe, and Johanna Nichols, 223–238. Norwood, NJ: Ablex.Google Scholar below.

Aoki (1986)Aoki, Haruo 1986 “Evidentials in Japanese.” In Evidentiality: The Linguistic Coding of Epistemology, ed. by Wallace Chafe, and Johanna Nichols, 223–238. Norwood, NJ: Ablex.Google Scholar discusses Japanese usages such as (61) and (62) from the evidential point of view, claiming that the gar 1313.In my view gar or garu is a suffix rather than a verb. of kana-shigatte iru is a verb which ‘has the function of expressing inference rather than experience’ (Aoki 1986Aoki, Haruo 1986 “Evidentials in Japanese.” In Evidentiality: The Linguistic Coding of Epistemology, ed. by Wallace Chafe, and Johanna Nichols, 223–238. Norwood, NJ: Ablex.Google Scholar, 225). According to him, Japanese evidentials, including gar, are not ‘grammaticized.’ However, the sentences in (62) are clearly ungrammatical, suggesting that there should be some restrictions.

Let me express (61a) and (61b) from the perspective of (meta)representation, as follows:

(63)

  1. [Sam says [Sam is sad]]

  2. [Sam says [Tom is sad]]

(63a) and (63b) show that when the speaker is the same person as the subject of ‘sad,’ (kana) shii is selected and that (kana) shigatte iru occurs when the speaker is different from the subject of ‘sad.’ We can therefore posit the following hypothesis.

(64)

Kana-shii co-occurs with the first person and kana-shigatte iru with the third person.1414.The status of second person as the other party of an exchange is complex concerning an experiencer of private feelings and is ignored here.

(64) describes a simple and basic case. However, take the case in which other persons intervene between the speaker and the subject. In other words, consider the situation where the speaker reports someone else’s inner feelings or sensations:

(65)

  1. Sam to Peter:
    Bill
    Bill
    wa
    Top
    kana-shii
    sad
    to
    Quot
    itta.
    said
    ‘Bill said he is/was sad.’

  2. Sam to Peter: *Bill wa kana-shigatte iru to itta.

In (65a) the subject of kana-shii and itta is the same person, Bill. Notice that ‘Bill’ is third person but kana-shii occurs where kana-shigatte iru should come according to the principle of (64). Why does kana-shii appear there?

Next, let us assume that the person who feels sad is a person other than Sam, the speaker, or Tom, the hearer. Suppose he is Gen.

(66)

Sam to Peter:
  1. Bill
    Bill
    wa
    Top
    Gen
    Gen
    ga
    Nom
    kana-shigatte
    sad
    iru
    be
    to
    Quot
    itta.
    said

    ‘Bill said Gen is/was sad.’

  2. *Bill wa Gen ga kana-shii to itta.

In the context of (66), we can explain straightforwardly that kana-shigatte iru occurs since Gen is third person, but what about the case of (67), where it is the speaker himself, Sam, who is ‘sad’?

(67)

Sam to Peter:
  1. Bill
    Bill
    wa
    Top
    watashi
    I (=Sam)
    ga
    Nom
    kana-shigatte
    sad
    iru
    be
    to
    Quot
    itta.
    said

    ‘Bill said I am/was sad.’

  2. *Bill wa watashi ga kana-shii to itta.

Here, the first person, watashi (=I) occurs with kana-shigatte iru, which seems counter to the constraint of (64).

Let us more closely examine ‘Bill’ in (65a) and watashi in (67a). In (65a) Bill is supposed to express his own sadness. That is, Bill is the direct source of his feeling and Sam, the speaker, tells Peter what Sam heard from Bill. In (67a), on the other hand, we see watashi, the first person, on the surface behaving as a third person since this watashi is followed by kana-shigatte iru. We will return to this point later.

3.2.2Desiderative predicates

Such phenomena that are sensitive to the original state of affairs can also be seen in the behavior of other private predicates in Japanese that express the desires of the subjects. Consider the examples in (68) and (69) (cf. Uchida 2011 2011 Goyoron no Shatei [The Scope of Pragmatics]. Tokyo: Kenkyusha.Google Scholar, 182–187).

(68)

  1. Sam:
    Watashi
    I
    wa
    Top
    Jill
    Jill
    to
    Part
    kekkon
    marry
    shi-tai.
    want
    ‘I want to marry Jill.’

  2. Sam: * Watashi wa Jill to kekkon shi-tagatteiru.

(69)

  1. Sam:
    Tom
    Tom
    wa
    Top
    Jill
    Jill
    to
    Part
    kekkon
    marry
    shi-tagatteiru.
    want
    ‘Tom wants to marry Jill.’

  2. Sam: *Tom wa Jill to kekkon shi-tai.

Kekkon here is the stem of the verb kekkon suru (=marry) and the utterances of (68) show that ‘want’ in English can be expressed in two ways, with a suffix -tai or -tagatteiru, in Japanese but the choice seems to be subject to some ‘principle.’ The simplest picture is that -tai appears when the subject of the verb is first person as in (68a), and -tagatteiru occurs with third person subjects, as in (69a). That is, in (68a) it is Sam, the speaker, who wants to marry Jill. In (69a), on the other hand, Tom, in the third person, is the one who wants to marry Jill. These linguistic facts can be represented in (70a) and (70b) respectively.

(70)

  1. [Sam says [Sam wants to marry Jill]]

  2. [Sam says [Tom wants to marry Jill]]

However, the whole picture is not so simple. In (71), for example, Bill is third person but co-occurs with -tai, and in (72) -tagatteiru is appropriate even if the subject is first person.

(71)

  1. Sam to Peter:
    Bill
    Bill
    wa
    Top
    Jill
    Jill
    to
    Part
    kekkon
    marry
    shi-tai
    want
    to
    Quot
    itta.
    said

  2. Sam to Peter: *Bill wa Jill to kekkon shi-tagatteiru to itta.‘Bill said he wants to marry Jill.’

(72)

  1. Sam to Peter:
    Bill
    Bill
    wa
    Top
    watashi
    I
    ga
    (=Sam)
    Jill
    Nom
    to
    Jill
    kekkon
    marry
    shi-tagatteiru
    want
    to
    Quot
    itta.
    said
    ‘Bill said I want /wanted to marry Jill.’

  2. Sam to Peter: *Bill wa watashi ga Jill to kekkon shi-tai to itta.

We can describe the metarepresentations of (71a) and (72a) as in (73a) and (73b) respectively:

(73)

  1. [Sam says [Bill said [Bill wants to marry Jill]]]

  2. [Sam says [Bill said [Sam wants to marry Jill]]]

It is clear that in (73a) the person who wants to marry Jill and the person who utters that fact are the same person, Bill, and that in (73b), on the other hand, the person who wants to marry Jill is different from the man who reported that Sam wants to marry Jill. These behaviors are parallel to those of private predicates.

In the two subsections above we have discussed those phenomena that are sensitive to the original state of affairs concerning private and desiderative predicates. If someone wants to express someone else’s feelings or desires, they have to report them based on what they heard from the direct source of information. In Japanese we have linguistic devices such as inflections of verbs or adjectives, in which direct or indirect information is explicitly realized. In English, on the other hand, there are no equivalent linguistic devices such as (kana)-shii / (kana) shi-gatte iru or -tai / -tagatte iru. Rather, a single form of ‘be sad’ or ‘want’ appears instead. We cannot trace back to the original state of affairs of the feelings or desires solely from linguistic information.

Just as word order is clearly a syntactic matter, ‘inflected metarepresentation’ as seen in the present paper also seems syntactic – but is it?

Let us sum up what has been discussed in Section 3 as follows:

(74)

  1. Sam to Tom:Bill said I(=Sam)’m sad *-shii/-shigatte iru to itta. (cf. [67])

  2. [Sam says [Bill said [Sam is sad]]]

(75)

  1. Sam to Tom:Bill said he wants to marry Jill -shi-tai/*-tagatte iru to itta. (cf. [71a])

  2. [Sam says [Bill said [Bill wants to marry Jill]]]

These behaviors of (kana)-shii/(kana)-shigatte iru and (shi)-tai/(shi)-tagatte iru can be pragmatically explained in the following way: in (74a), the first person ‘I’ behaves like a third person since it is followed by (kana)-shigatte iru instead of (kana)-shii, while in (75a), ‘he’ is a third person, but it behaves like the first person since (shi)-tai, not (shi)-tagatte iru, follows the subject. I’d like to call this first person in (74a) a ‘covert third person’ and the third person in (75a) a ‘covert first person.’ We could say that ‘I’ in (74a) and ‘he’ in (75a) are syntactically first person and third person respectively, but pragmatically, third person and first person. We may call this linguistic phenomenon ‘fake’ person.1515.I borrowed the term ‘fake’ from ‘fake past’ in Nishiguchi (2006)Nishiguchi, Sumiyo 2006 “Fake Past and a Monster.” Stony Brook Occasional Papers in Lingusitics I: 152–171.Google Scholar. See also Uchida (2013 2013 Kotoba-wo Yomu, Kokoro-wo Yomu: Ninchi Goyoron Nyumon [Reading Words, Reading Mind: Introduction to Cognitive Pragmatics]. Tokyo: Kaitakusha.Google Scholar, 90–93). This is a phenomenon similar to agreement of tense in English as seen in ‘was’ in ‘Mary said Jill was sick in bed.’ ‘Was’ here could be interpreted as ‘present’ and be called ‘covert present.’

In view of evidentiality, inner feeling, desire, thought, perception and so on can only be conveyed by the person concerned and those conveyed from other persons are treated as indirect evidence.1616.See Uchida (2022) 2022 “Metarepresentation kara mita evidentiality [Metarepresentation and Evidentiality].” Memoirs of Nara University 50: 129–145.Google Scholar for a cognitive pragmatic approach to evidentiality. Therefore, kana-shii or shi-tai appears when the subject regards the information as something that she/he knows or feels, and kana-shigatteiru or shi-tagatteiru is found when the subject obtains the information through someone else.

4.Concluding remarks

We have shown in the present paper that in Japanese some types of metarepresentational information must be overtly reflected by higher-level explicatures such as verbs indicating speech acts, speech act particles, and suffixes, while in English, they can be left implicit.

The metarepresentational approach advocated here has been developed from the concept of higher-level explicatures in relevance theory by Sperber and Wilson (1986/1995)Sperber, Dan, and Dierdre Wilson 1986/1995 Relevance: Communication and Cognition. Oxford: Blackwell.Google Scholar and is closely connected with cognitive aspects of language as their subtitle Interpretation and Cognition implies. Metarepresentational phenomena, as we have discussed, cover both linguistic affairs and cognitive information on how utterances are processed. Metarepresentation is inherently cognitive-oriented.

Generally speaking, comparative studies on languages tend to focus on the differences in linguistic forms or meaning between target languages, without paying particular attention to the cognitive side of interpretation. The cognitive processing of linguistic information by humans is supposed to be more or less the same in nature; what differs is linguistic realization in each language. In this sense the metarepresentational perspective reveals vast potential in the field of comparative linguistics. Uchida and Noh (2018)Uchida, Seiji, and Eun-Ju Noh 2018 “Metarepresentational Phenomena in Japanese and Korean.” Memoirs of Nara University 46: 1–23.Google Scholar shares this view and demonstrates that Japanese and Korean behave very similarly in that the speech act side of higher-level explicatures tends to be linguistically realized in both languages.1717.The paper also suggests that the reason for that might be that they have the same SOV word order, while English is an SVO language.

We suggested above that the alternation of (kana)-shii/(kana)-shigatte iru or (shi)-tai/(shi)-tagatteiru depends on pragmatic factors such as direct/indirect evidence and furthermore, that this conclusion can be applied to other languages where information on higher-level explicatures or metarepresentations is basically explicit as in Japanese and Korean or implicit as in English. We hope that the metarepresentational approach we have presented here will encourage further comparative linguistic studies.

Funding

The present paper has been partly supported by Japan Society for the Promotion of Science (18K00674).

Acknowledgements

I remember that my communication with Deirdre started in 1992. I sent an air mail letter – e-mail was not a popular communication tool yet – to invite her to an international conference in Osaka the following year on “New Perspectives from Relevance Theory,” meant to celebrate the publication of the Japanese translation of Relevance: Communication and cognition. She kindly accepted our offer and visited Japan with Dan Sperber and Robyn Carston. The two-day session was a success, resulting in the 1998 publication of Relevance Theory: Applications and Implications edited by R. Carston and S. Uchida. Ever since that time, my research interest has focused on relevance theoretic approaches to various linguistic phenomena both in English and in Japanese, perpetually inspired by Deirdre’s work. It is a great pleasure to contribute to this special issue of Pragmatics dedicated to Deirdre Wilson.

I would like to thank two anonymous reviewers for their invaluable comments and suggestions.

Notes

1.As noted by one of the reviewers, the paraphrases in (3) are examples of ‘second-order’ metarepresentaion in the sense of Wilson (2000)Wilson, Deirdre 2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar and Noh (2000) 2000 Metarepresentation. Amsterdam: John Benjamins. DOI logoGoogle Scholar. Metarepresentations, whether first-order or second-order, are linguistically related in some way in languages like Japanese, as we discuss below. We mention these two types of metareprentation in Section 3, where information from metarepresentation is crucial.
2.See Wilson and Sperber (1993)Wilson, Dierdre, and Dan Sperber 1993 “Linguistic Form and Relevance.” Lingua 93: 1–25. DOI logoGoogle Scholar, Ifantidou (2001Ifantidou, Elly 2001 Evidentials and Relevance. Amsterdam: John Benjamins. DOI logoGoogle Scholar, 97–118), and Carston (2002Carston, Robyn 2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. DOI logoGoogle Scholar, 120–125) for relevance theoretic approaches to sentence adverbs.
3. Blakemore (1996) 1996 “Are Apposition Markers Discourse Markers?” Journal of Linguistics 32: 325–347. DOI logoGoogle Scholar takes the same standpoint in that ‘in other words’ and ‘in short’ put constraints on higher-level explicatures as reformation markers.
4. Sweetser (1990)Sweetser, Eve 1990 From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar classifies ‘because’ clauses into three types; content, epistemic, and speech-act types. The use of the ‘since’ clause of (16b) corresponds to the speech act domain. cf. footnote 5.
5.According to Sweetser (1990Sweetser, Eve 1990 From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar, 76–78), (18a) and (18b) are cases of the content domain and the epistemic domain respectively, and Rutherford (1970)Rutherford, William E. 1970 “Some Observations Concerning Subordinate Clauses in English.” Language 46 (1): 97–115. DOI logoGoogle Scholar calls the ‘because’ clause of (18a) the restrictive use and that of (18b) the non-restrictive. However, neither Sweetser nor Rutherford mention the speech act performed by saying ‘Peter was drunk.’
6.Ross (1970Ross, John Robert 1970 “On Declarative Sentences.” In Readings in English Transformational Grammar, ed. by Roderick A. Jacobs, and Peter S. Rosenbaum, 222–272. Waltham, MA: Ginn and Company.Google Scholar, 254–258) does mention ‘pragmatic analysis’ as an alternative to the performative analysis and concludes that the pragmatic analysis would be a notational variant of his performative analysis. In his view of pragmatics, the utterance ‘Prices slumped’ has a ‘deep’ structure such as [Prices slumped] instead of [I say to you prices slumped], but, as we have seen, one of the explicatures of ‘Prices slumped’ would be [The speaker says [prices slumped]], where the speech act component is explicitly stated and is not to be deleted. The main reason that Ross didn’t give value to pragmatics at that time might be that pragmatics was not as cognitively oriented as it is today. See the explicature analysis proposed in Uchida (2011 2011 Goyoron no Shatei [The Scope of Pragmatics]. Tokyo: Kenkyusha.Google Scholar, 145–158).
7. Ga and wa are both nominative case-marking particles and wa also works as a topic marker.
8. Da and noda are both sentence final particles, but da follows after nouns and noda comes at the end of predicates.
9.This so here is a pronoun which refers to what is anaphorically mentioned and is quite different from -soo as an evidential marker, which Narrog and Yang (2018)Narrog, Heiko, and Wenjiang Yang 2018 “Evidentiality in Japanese.” In The Oxford Handbook of Evidentiality, ed. by Alexandre Y. Aikhenvald, 709–724. Oxford: Oxford University Press. DOI logoGoogle Scholar discusses.
10.Yet another interpretation, ‘Jill said you (=Peter) have dropped your purse,’ is also possible, but (50a) is a more likely interpretation.
11.See Blakemore (1994)Blakemore, Diane 1994 “Echo Questions: A Pragmatic Account.” Lingua 94: 197–211. DOI logoGoogle Scholar and Noh (1995Noh, Eun-Ju 1995 “A Pragmatic Approach to Echo Questions.” UCL Working Papers in Linguistics 7: 107–140.Google Scholar, 1998 1998 “Echo Questions: Metarepresentation and Pragmatic Enrichment.” Linguistics and Philosophy 21 (6): 603–628. DOI logoGoogle Scholar, 2000 2000 Metarepresentation. Amsterdam: John Benjamins. DOI logoGoogle Scholar) for relevance theoretic approaches to echo questions. Noh (1995)Noh, Eun-Ju 1995 “A Pragmatic Approach to Echo Questions.” UCL Working Papers in Linguistics 7: 107–140.Google Scholar mentions that Korean and Japanese echo questions behave similarly. Iwata (2003)Iwata, Seizi 2003 “Echo Questions.” Linguistics and Philosophy 26: 185–254. DOI logoGoogle Scholar discusses an alternative analysis.
12.This utterance is possible in the context of fictional discourse. See Kuroda (1973)Kuroda, S.-Y. 1973 “Where Epistemology, Style, and Grammar Meet: A Case Study from Japanese.” In A Festschrift for Morris Halle, ed. by Stephen R. Anderson, and Paul Kiparsky, 377–391. New York: Holt, Rinehart and Winston.Google Scholar and Uchida (2013 2013 Kotoba-wo Yomu, Kokoro-wo Yomu: Ninchi Goyoron Nyumon [Reading Words, Reading Mind: Introduction to Cognitive Pragmatics]. Tokyo: Kaitakusha.Google Scholar, 142–144). If no-da (sentence final particle) or chigainai (=epistemic ‘must’) follows, kana-shii is also possible, where noda functions as a kind of marker of interpretive use in relevance theory: Sam commits to the truth of the contents of the proposition that Tom is sad (cf. Uchida 1998Uchida, Seiji 1998 “ ‘(No)da’ – Kanrensei Riron karano Shiten [‘Noda’ in Relevance Theory].” In Gendai Eigo no Goho to Bunpo [English Usage and Grammar], ed. by the Editorial Board, 243–251. Tokyo: Taishukan.Google Scholar).
13.In my view gar or garu is a suffix rather than a verb.
14.The status of second person as the other party of an exchange is complex concerning an experiencer of private feelings and is ignored here.
15.I borrowed the term ‘fake’ from ‘fake past’ in Nishiguchi (2006)Nishiguchi, Sumiyo 2006 “Fake Past and a Monster.” Stony Brook Occasional Papers in Lingusitics I: 152–171.Google Scholar. See also Uchida (2013 2013 Kotoba-wo Yomu, Kokoro-wo Yomu: Ninchi Goyoron Nyumon [Reading Words, Reading Mind: Introduction to Cognitive Pragmatics]. Tokyo: Kaitakusha.Google Scholar, 90–93).
16.See Uchida (2022) 2022 “Metarepresentation kara mita evidentiality [Metarepresentation and Evidentiality].” Memoirs of Nara University 50: 129–145.Google Scholar for a cognitive pragmatic approach to evidentiality.
17.The paper also suggests that the reason for that might be that they have the same SOV word order, while English is an SVO language.

Abbreviations

Ac

accusative case

Nom

nominative case

Obj

objective case

Part

particle

Polite

polite form

Pst

past form

Quot

quotative particle

Top

topic marker

References

Aoki, Haruo
1986 “Evidentials in Japanese.” In Evidentiality: The Linguistic Coding of Epistemology, ed. by Wallace Chafe, and Johanna Nichols, 223–238. Norwood, NJ: Ablex.Google Scholar
① ② ③ ④
Blakemore, Diane
1994 “Echo Questions: A Pragmatic Account.” Lingua 94: 197–211. DOI logoGoogle Scholar
①
1996 “Are Apposition Markers Discourse Markers?” Journal of Linguistics 32: 325–347. DOI logoGoogle Scholar
①
Carston, Robyn
2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. DOI logoGoogle Scholar
① ②
Grice, Paul H.
1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press.Google Scholar
①
Ifantidou, Elly
2001 Evidentials and Relevance. Amsterdam: John Benjamins. DOI logoGoogle Scholar
①
Iwata, Seizi
2003 “Echo Questions.” Linguistics and Philosophy 26: 185–254. DOI logoGoogle Scholar
①
Kuno, Susumu
1973 The Structure of the Japanese Language. Cambridge, MA: MIT Press.Google Scholar
① ②
Kuroda, S.-Y.
1973 “Where Epistemology, Style, and Grammar Meet: A Case Study from Japanese.” In A Festschrift for Morris Halle, ed. by Stephen R. Anderson, and Paul Kiparsky, 377–391. New York: Holt, Rinehart and Winston.Google Scholar
① ②
Levinson, Stephen
1983 Pragmatics. Cambridge, MA: Cambridge University Press. DOI logoGoogle Scholar
①
Nishiguchi, Sumiyo
2006 “Fake Past and a Monster.” Stony Brook Occasional Papers in Lingusitics I: 152–171.Google Scholar
①
Narrog, Heiko, and Wenjiang Yang
2018 “Evidentiality in Japanese.” In The Oxford Handbook of Evidentiality, ed. by Alexandre Y. Aikhenvald, 709–724. Oxford: Oxford University Press. DOI logoGoogle Scholar
① ②
Noh, Eun-Ju
1995 “A Pragmatic Approach to Echo Questions.” UCL Working Papers in Linguistics 7: 107–140.Google Scholar
① ②
1998 “Echo Questions: Metarepresentation and Pragmatic Enrichment.” Linguistics and Philosophy 21 (6): 603–628. DOI logoGoogle Scholar
① ②
2000 Metarepresentation. Amsterdam: John Benjamins. DOI logoGoogle Scholar
① ② ③
2021 “On Linguistic Communication Based on Resemblance in Form.” Journal of Pragmatics 186: 20–32. DOI logoGoogle Scholar
① ②
Ross, John Robert
1970 “On Declarative Sentences.” In Readings in English Transformational Grammar, ed. by Roderick A. Jacobs, and Peter S. Rosenbaum, 222–272. Waltham, MA: Ginn and Company.Google Scholar
① ②
Rutherford, William E.
1970 “Some Observations Concerning Subordinate Clauses in English.” Language 46 (1): 97–115. DOI logoGoogle Scholar
①
Sperber, Dan
1994 “Understanding Verbal Understanding.” In What is Intelligence? ed. by Jean Khalfa, 179–198. Cambridge: Cambridge University Press.Google Scholar
①
ed. 2000 Metarepresentations: A Multidisciplinary Perspective. Oxford: Oxford University Press.Google Scholar
① ②
2000 “Metarepresentations in an Evolutionary Perspective.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 117–137. Oxford: Oxford University Press.Google Scholar
①
Sperber, Dan, and Dierdre Wilson
1986/1995 Relevance: Communication and Cognition. Oxford: Blackwell.Google Scholar
① ②
2002 “Pragmatics, Modularity and Mindreading.” Mind and Language 17: 3–23. DOI logoGoogle Scholar
①
Sweetser, Eve
1990 From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. DOI logoGoogle Scholar
① ②
Uchida, Seiji
1998 “ ‘(No)da’ – Kanrensei Riron karano Shiten [‘Noda’ in Relevance Theory].” In Gendai Eigo no Goho to Bunpo [English Usage and Grammar], ed. by the Editorial Board, 243–251. Tokyo: Taishukan.Google Scholar
①
2011 Goyoron no Shatei [The Scope of Pragmatics]. Tokyo: Kenkyusha.Google Scholar
① ② ③
2013 Kotoba-wo Yomu, Kokoro-wo Yomu: Ninchi Goyoron Nyumon [Reading Words, Reading Mind: Introduction to Cognitive Pragmatics]. Tokyo: Kaitakusha.Google Scholar
① ② ③
2022 “Metarepresentation kara mita evidentiality [Metarepresentation and Evidentiality].” Memoirs of Nara University 50: 129–145.Google Scholar
①
Uchida, Seiji, and Eun-Ju Noh
2018 “Metarepresentational Phenomena in Japanese and Korean.” Memoirs of Nara University 46: 1–23.Google Scholar
① ② ③ ④
Wilson, Deirdre
2000 “Metarepresentation in Linguistic Communication.” In Metarepresentations: A Multidisciplinary Perspective, ed. by Dan Sperber, 411–448. Oxford: Oxford University Press.Google Scholar
① ② ③ ④ ⑤ ⑥ ⑦
Wilson, Dierdre, and Dan Sperber
1993 “Linguistic Form and Relevance.” Lingua 93: 1–25. DOI logoGoogle Scholar
① ② ③

Address for correspondence

Seiji Uchida

Nara University

1500 Misasagi

Nara, 635-8502

Japan

[email protected]

Biographical notes

Seiji Uchida is now a specially appointed professor of Nara University and a professor emeritus of Nara Women’s University, Japan. His research is mainly concerned with metarepresentaional phenomena including deixis and text analysis. He is the lead translator of Relevance: Communication and Cognition and Robyn Carston’s Thoughts and Utterances: The Pragmatics of Explicit Communication. He was president of the Society of English Grammar and Usage, Japan, from 2012 to 2016.