Involvement and Attitude in Japanese Discourse: Interactive Markers by Naomi Ogi is a study of a subset of what is traditionally called ‘sentence-final particles’ (終助詞) in Japanese.Footnote ^1, Footnote ² This subgroup comprises what the author calls ‘interactive markers’. The book consists of eight chapters: 1. ‘Introduction’, 2. ‘Approaches to interactive markers’, 3. ‘Involvement, formality and gender in language use’, 4. ‘Involvement and the speaker's attitudes’, 5. ‘Incorporate markers ne and na’, 6. ‘Monopolistic markers yo and sa’, 7. ‘Monopolistic markers wa, zo and ze’, and 8. ‘Conclusion’. These chapters are followed by references, data sources, author index, and subject index. Ogi's study is based on two corpora and two other data sources, namely two manga series (as examples of written representations of spoken language).

In Chapter 1, Ogi defines the Japanese particles under investigation as ‘interactive markers’, since she insists that they are not simply related to ‘information state’ (p. 6 and Chapter 2) but have to do with ‘different attitudes of the speaker towards the hearer through involvement’ (p. 7). Ogi sees the Japanese markers as related to ‘discourse markers’ (p. 8), which also are ‘interpersonal’ (Halliday Reference Halliday1970), ‘interactional’ (Brown & Yule Reference Brown and Yule1983) and ‘non-referential’ (Silverstein Reference Silverstein, Basso and Selby1976, Maynard Reference Maynard1993) (footnote 5 on page 5)Footnote ³ but nevertheless are a distinct subgroup of them, since they function at an interactional level, not at a discourse-organizational level. Ogi refers to another, narrower class of ‘discourse particles’, which are ‘particular elements in German, Dutch and Norwegian, which seem to function differently from those elements in English that are generally referred to as Discourse Markers’ (p. 9). We could add that they occur in Danish and other languages, too.

In Chapter 2, Ogi reviews some previous studies on the seven markers studied by her – ne, na, yo, sa, wa, zo and ze – and she insists that while each of the earlier approaches (‘syntactical’ (pp. 24–27), ‘cognitive’ (pp. 27–37), ‘illocutionary force’ (pp. 37–42), ‘interactional’ (pp. 42–44), and ‘social’ (pp. 44–47) approaches) has made invaluable contributions, none has achieved an integrated analysis. She maintains that this can be done by exploring these seven targeted markers with the notion of ‘involvement’, which she defines as ‘the speaker's interactional attitude towards hearer’ (p. 47).

In Chapter 3, Ogi introduces three key notions related to her analysis: involvement, formality and gender. As to the notion of involvement, Ogi notes that the seven particles have several common features: ‘(they) frequently occur in face-to-face conversation’, ‘their use is almost mandatory for the success of conversation in Japanese’, ‘they are closely related to the non-referential (or interactional) meaning of language rather than referential (or propositional) meaning’ and ‘they do not affect the truth-condition of the propositional information of an utterance, but influence the hearer's interpretation of the utterance’ (p. 52). Regarding the notion of formality, Ogi states that the use of the seven markers also relates to ‘the interpersonal distance between the speaker and the hearer’, and that the use of them ‘is restricted by the interpersonal relationship between the speaker and the hearer such as the relative social hierarchy and/or age which play a significant role for Japanese speaker's choice of speech styles, i.e. more formal or informal’ (p. 57). In relation to the notion of gender, Ogi outlines how and why onna-kotoba (women's language) has developed in Japanese, and among the seven markers, wa is considered as ‘a feature of female speech’ while na, zo and ze are ‘a feature of male speech’ (p. 65).

In Chapter 4, Ogi introduces two subcategories of interactional markers: (i) incorporative and (ii) monopolistic. The notion ‘incorporative’ implies ‘the speaker's attitude of inviting the hearer's involvement through which he/she is committed to align with the hearer with respect to the content and feeling conveyed in the utterance’, and the notion ‘monopolistic’ implies ‘the speaker's attitude of inviting the hearer's involvement through which he/she is committed to enhance his/her position as a deliverer of the content and feeling towards the hearer’ (p. 73). Among the seven particles, ne and na are incorporative and ‘can roughly be glossed as “(. . .) Don't you think so?”’ (p. 73), while yo, sa, wa, zo and ze are monopolistic and ‘can roughly be paraphrased as “Listen. I have something to tell you”’ (p. 83).

In Chapters 5–7, Ogi investigates the seven interactive markers in detail. First in Chapter 5, she deals with the incorporative markers, ne and na. The difference between ne and na is that na has some restrictions in its use, e.g. the speaker's age, social status and gender. In Chapter 6, Ogi takes up two of the monopolistic markers, yo and sa, and in Chapter 7, the other three, wa, zo and ze. Besides her description of the other differences between the functions of the monopolistic interactive markers, she also notes that while yo and sa are used by both men and women, wa is exclusively used by women, and zo and ze by men.

In reviewing this book from a Scandinavian perspective, we will focus especially on the similarities and the differences between Ogi's interactive markers (IMs) and Danish dialogical particles (DPs) (Hansen & Heltoft Reference Hansen and Heltoft2011). We will begin by describing why we consider this comparison as being useful, and we will next argue how we could find the features shared by IMs and DPs.

A number of languages in different parts of the world have a class of ‘small words’, which have variously been called ‘modal particles’ (Arndt Reference Arndt1960), shading particles (‘Abtönungspartikel’, Weydt Reference Weydt1969), dialogical particles (‘dialogiske partikler’, Hansen & Heltoft Reference Hansen and Heltoft2011), ‘sentence-final particles’ (Simpson Reference Simpson, Huang, Audrey Li and Simpson2014), or ‘interactional particles’ (Morita Reference Morita2005). They are in some form found in all Germanic verb-second (V2) languages – i.e. all modern Germanic languages except English (Schubiger Reference Schubiger1965) – and also in Russian, Chinese (Putonghua, but especially in Cantonese) and Japanese. The terminological variation suggests that the phenomenon is heterogeneous, but we might say that the terminological variation is also a consequence of several researchers having developed descriptive categories for each language independently. The World Atlas of Linguistic Structures (WALS) has no entry for these ‘small words’ though, which might make one ask if there actually is a comparative category. Instead of keeping single-language descriptive categories strictly apart from comparative categories, as Haspelmath (Reference Haspelmath2010) has argued for, we would rather follow van der Auwera & Sahoo (Reference van der Auwera and Sahoo2015), who suggest that we can approach comparative categories by comparing a descriptive category in one language with another descriptive category in a different language; if there are sufficient similarities, this can be the point of departure for defining a comparative concept (adding more languages in the process). What is relevant are not only the similarities but also the differences we can find by comparing different descriptive categories. We can use the similarities to define a comparative category, and can proceed from the differences to describe what is specific to each language.

In his discussion of what he reluctantly calls ‘modal particles’ in Russian and German, Arndt (Reference Arndt1960) provides us with a useful starting point: he notes that the Russian ‘small words’ in question

1. (a) are short (typically monosyllabic);

2. (b) are unstressable, though often minimally distinguished from stressed doublets only by absence of stress;

3. (c) do not occur as complete utterances (between major pauses);

4. (d) are not elicitable, i.e. not correlated to interrogatives so that they can be elicited by questioning;

5. (e) are positionally restricted in comparison with adverbs, free in comparison with conjunctions;

6. (f) modify an utterance as a whole, and not with special reference to any constituent such as the predicate;

7. (g) are omissible, i.e. convey no element of the objective message (in its factual or cognitive function), but convey the subjective emotional or mental attitude of the speaker to his interlocutor, to the objective message content, or to another element of the linguistic situation (emotive function).

• (Arndt Reference Arndt1960:326)

We will leave aside point (e) because it refers to language-specific clause topology. Arndt himself examined to what extent these features are shared by German modal particles. He concluded that all features set up for Russian are shared with them except, not surprisingly, for point (e), where differences occur between German and Russian (Arndt Reference Arndt1960:332).

Before proceeding with any further discussion of IMs and DPs, we want to consider to what extent they are ‘small words’ according to Arndt's list.

1. (a) The IMs ne, na, yo, sa, wa, zo, ze, are all monosyllabic, while DPs are mono- or bisyllabic: jo, skam, mon, vel, nok, vist, nu, da, dog, måske, altså, ellers, etc. (Hansen & Heltoft Reference Hansen and Heltoft2011:1046).

2. (b) In her book, Ogi does not discuss stress, intonation or Japanese pitch accent, so we will not discuss these issues further, although we assume that IMs can neither carry emphatic nor contrastive stress.Footnote ⁴ DPs can clearly not be stressed (Davidsen-Nielsen Reference Davidsen-Nielsen1993:2).

3. (c) Neither IMs nor DPs occur as independent complete utterances. Ogi (p. 74) gives an example with ne alone as a complete utterance (from Lee Reference Lee2007:368), but we wonder if this ne is not a ‘stressed doublet’ in the sense of Arndt (point (b) above), hence not an IM proper. In the same way, Hansen & Heltoft (Reference Hansen and Heltoft2011:1130–1131) refer to altså and dog occurring as a complete utterance with emotional uses, but they categorize these as interjections, not as DPs.Footnote ⁵

4. (d) As a corollary of (c) just above, neither IMs nor DPs can be elicited by questions.

5. (f) From the rather rigid clausal topology of both IMs and DPs, it follows that they cannot have a single constituent of their clause as scope.

6. (g) Ogi states that an IM ‘does not affect the truth-conditions of the propositional information of an utterance’ (p. 4) and she calls this semantic or functional aspect of IMs ‘non-referential’ (p. 5). The same goes for DPs (Andersen Reference Andersen1982:89).

It emerges from Table 1 that IMs and DPs share a sufficient number of features beyond being ‘small words’ to consider a comparative category.

Table 1.

Shared features of Japanese interactive markers (IMs) and Danish dialogical particles (DPs).

In her introduction, Ogi (p. 7) outlines four main purposes of her study, cited here verbatim:

1. 1. Why are these markers used in spoken language, whereas rarely used in written language?

2. 2. What is the common property shared by these markers?

3. 3. What is the unique linguistic function of each marker?

4. 4. What is the relationship between the linguistic function of each marker and the speaker's expression of formality and/or gender within social contexts?

In the remaining part of this review, we will follow Ogi's agenda and take each of these points in turn in relation to Danish DPs.

Why are these markers used in spoken language, whereas rarely used in written language? Ogi states that the main common function of the seven markers is to ‘express the speaker's attitude of inviting the hearer's involvement, whereby the hearer pays his/her exclusive attention to the utterance and more actively participates in the conversation’ (p. 198). This is also the reason why the seven markers are used, in principle, only in spoken language in Japanese. To demonstrate that, Ogi takes an excerpt from Japanese newspaper article as an example. Inserting interactional markers, she shows that the manipulated text turns out to be ‘impossible’ (pp. 67–70). In contrast, such an insertion would work perfectly well in Danish. This is a first important difference: the Japanese particles only occur in informal spoken language and its representation in writing,Footnote ⁶ but not in expository prose at all. The term ‘dialogical particle’ highlights the fact that DPs primarily appear in actual or implicit dialogue, i.e. presuppose two voices, but they are not necessarily restricted to spoken language. That this is possible may be explained by an understanding of such prose as ‘implicit dialog’ (Gray Reference Gray1977): in expository prose, every sentence can be read as a reaction to an imagined question of the reader. Thus, different voices can be present even in prose. Why this explanation works for Danish but not for Japanese remains an intriguing question that requires further discussion. Our suggestion is that it might turn out that this is not primarily a question of the medium (spoken or written), but an issue of orientation towards the audience: the Japanese markers are sensitive to a specific social relationship between sender and audience which is only clearly defined in face-to-face interaction. An exception may be children's books. Obe & Haberland (Reference Obe and Haberland2018) found, among others, two cases of ne in the narrative prose of Ohata's Japanese translation of H. Chr. Andersen's fairy tale Den grimme ælling (The Ugly Duckling). There is the possibility that in Japan this fairy tale is considered as children's literature, hence as a story told to a given audience (i.e. to children), which would ‘fix’ the problem of an appropriate style of address to the audience.

What is the common property shared by these markers? Hansen & Heltoft's (Reference Hansen and Heltoft2011:1034) claim is that the Danish dialogical particles all make some assumptions about the structure of the communicative situation or interaction. This could also broadly be said about the Japanese markers; both Hansen & Heltoft (Reference Hansen and Heltoft2011) and Ogi emphasize interaction as crucial here. But while the range of functions of DPs and IMs seems to overlap, it is not identical. Especially the content of the ‘evidential’ subtype of dialogical particles (such as nok, vel, vistnok; Hansen & Heltoft Reference Hansen and Heltoft2011:1062) is typically not expressed in Japanese by members of the IM system, but by modal expressions like yooda, sooda, and daroo.Footnote ⁷

1. (1)

   

What is the unique linguistic function of each marker? Neither IMs ne and na nor DP jo co-occur with abrupt commands:

1. (2)

   

This is due to the conflict caused by the use of ne and na signaling the speaker's incorporative attitude on the one hand, and the use of abrupt commands expressing ‘the speaker's intention of unilaterally delivering the utterance towards the hearer’ (p. 106) on the other. The same conflict might be a reason why jo in Danish does not occur with imperatives. Jo is a Danish dialogical particle signaling that the speaker presupposes that the hearer will not argue with the speaker about the content of the utterance (Hansen & Heltoft Reference Hansen and Heltoft2011:1091). This is not compatible with an abrupt command.

Three of Ogi's seven markers, yo, ze and zo, can signal that the utterance conveys ‘an implied message’.

1. (3)

   

Ogi explains that ‘the use of yo is crucial here because it is this marker that functions to make the hearer understand this implied message’ (p. 148). Similarly, Ogi remarks (p. 188) that the two occurrences of ze in (4) signal that there is an implied message in B's utterance.

1. (4)

   

However, while ze in this utterance is necessary to signal this – without ze, no additional message is implied – ze and yo can also occur when no further message is implied: the occurrence of yo or ze does not automatically signal that there is an implied message. This is different with zo, which always signals an implied message (p. 176), such as in the following example:

1. (5)

   

Here, the surface message is ‘I'm going’, but what is implied is that the addressee, Kotaro, is supposed to go, too, i.e. ‘Let's go’ (pp. 177–178).

Danish argumentative dialogical particles such as ellers and altså also signal the presence of an implied message:

1. (6)

   

This example is taken from Jensen (Reference Jensen2000:100). Hansen & Heltoft (Reference Hansen and Heltoft2011:1088) explain that there is an implied message in the parent's reply in (6), marked as a premiss itself (‘There is bacon in it’) together with an implicit second premiss ‘and you like bacon’. Therefore, the second utterance works as a suggestion to the child to eat the sauce. However, although ellers and altså are dialogical particles, hence interactional, their content can be expressed in Japanese both by IMs and by other means:

1. (7)

   

Another example that shows that Danish DPs can underline an implied message but are not necessary to express it is given in (8):

1. (8)

   

Torben Andersen notes that altså in B's utterance signals that B would like to change A's plan to go to Copenhagen because Birgit is coming home tomorrow (Andersen Reference Andersen1982:91). However, apparently, there is no dialogical particle in Danish that always signals an implied message in an utterance, while, on the other hand, even if ellers in (6) above and altså in (8) are omitted, the utterances can still carry an implied message. This is different from the situation with Japanese, yo, ze and zo.

What is the relationship between the linguistic function of each marker and the speaker's expression of formality and/or gender within social contexts? Ogi states that ‘the use of these [i.e. interactive] markers is a direct presentation of the speaker's attitude towards the hearer through expressing his/her invitation of the hearer's involvement, thus it is restricted by the interpersonal relationship between the speaker and the hearer such as the relative social hierarchy and/or age which play a significant role for Japanese speaker's choice of speech styles, i.e. more formal or informal’ (p. 57), and she repeats at several places in her book that this feature of the seven markers is the most important reason why these markers are used in casual conversations and not e.g. in a newspaper article which is directed to a more diffuse audience. Furthermore, she remarks that ‘the inappropriate use of non-referential signs, such as interactive markers, may often be due to a lack of sociocultural knowledge, and may cause communication problems in cross-cultural settings’ (pp. 201–202). Ogi gives an example with yo, which can be paraphrased as ‘“Listen. I want to ensure that you understand what I say and how I feel” and “I recognize you as my conversation partner and wish to continue the conversation with you”’ (pp. 137–138), to illustrate a case where the utterance can be interpreted as impolite.

1. (9)

   

Ogi agrees with Hayashi's (Reference Hayashi2000:39) analysis of the utterance by Student 2 as sounding more impolite than Student 1’s utterance because of the use of yo, and Ogi explains that ‘in a Japanese context, students are generally expected to speak in a formal and polite way to keep a certain distance between teachers. The use of yo by Student 2 above, thus, inappropriately imparts an intimate nuance towards the teacher and would be interpreted as impolite’ (p. 137).

There has been much focus on how in Japanese, grammar and reference to interpersonal relationships are intertwined, but similar observations can also be made about the supposedly ‘egalitarian’ Danish speech community. The following is an example of a socially inappropriate use of the Danish DP lige (when addressing a teacher):

1. (10)

   

Hansen & Heltoft (Reference Hansen and Heltoft2011:1072) characterize the use of lige with imperatives as the speaker's presupposing that the action required is not against the hearer's willingness to perform it, and that the action is not difficult to perform (this is what Levisen & Waters (Reference Levisen and Waters2015:251) call ’routine activity’). However, there is more to it: the use of lige is only appropriate if the speaker has the right to impose an action on the hearer, i.e. among social equals or towards socially subordinates.

In conclusion, we hope that it is clear that we find Involvement and Attitude in Japanese Discourse: Interactive Markers to constitute a carefully argued and very useful account of a subset (= IMs) of the traditional Japanese sentence-final particles. The author does her best to make very complicated matters as clear as possible, in particular by explaining the various studies in, and approaches to, the field and the differences between each IM on several levels, semantically, pragmatically and sociolinguistically.

As we have seen, this book helps us to find that Japanese and the European ‘particle’ languages (of which we have taken Danish as an example) have elements that look similar and exhibit certain shared features but are far from being used in the same way. Thus, this book also gives us some interesting viewpoints, which are not discussed sufficiently in connection with Danish dialogical particles. Finally, it may be worth noting in passing that looking at differences as well as similarities between languages gives us new insight into the unity and diversity of human language.