REVIEW OF THE LITERATURE
ON MA-PARTICLE INTERROGATIVES
2.1 Introduction
In reviewing the literature on ma-particle interrogatives, one cannot help but be struck by linguists' explicit denial but implicit assertion of the complexity of these "yes-no" questions. Admittedly, there is a superficial simplicity to such constructions, as asserted by Li and Thompson (1981:547) to be the least complex of all the question types in structural terms. But this syntactic phenomenon masks a wealth of controversy covering the areas of pragmatics, semantics, phonology, and even basic typology of the ma-particle question.
This chapter will review some of the major works on such interrogatives from the past three decades. The positions of the various linguists will be presented first, with critical discussion provided in section 2.6. Additional material will be drawn from Bolinger (1972), Hudson (1975), Lyons (1977) and Levinson (1983) in an attempt to better understand the Chinese structure within a more language-universal framework. In particular, the distinction drawn by Lyons between question posing and question asking will be further developed as this study's framework for the analysis of the pragmatics of Mandarin interrogatives.
2.2 Classification of interrogatives
Without exception, ma-particle questions are commonly considered to be a distinct category of interrogatives in Mandarin Chinese. The specific categories vary in number from author to author, and even at the hands of the same author. Lyu's (1985) most recent categorization cites question-word or WH-questions, and yes-no questions (Chinese tezhi wenju 特 指 問 句 and shifei wenju 是 非 問 句 respectively) as the basic types. Sentences (1) and (2) below are examples of WH-questions and yes-no questions in Mandarin Chinese.
(1) Shei lai le who come ASP "Who came?" (2) Ta lei ma s/he be tired Q "Is s/he tired?"Lyu furthermore proposes that A-not-A and disjunctives (Chinese zhengfan wenju 反 正 文 句 and xuanze wenju 選 擇 問 句 respectively) are derived from the yes-no (that is, ma-particle) question. Examples of A-not-A and disjunctive questions are given in sentences (3) and (4) below.
(3) Ni he bu he jiu you drink NEG drink liquor "Do you drink?" (4) Ni qu haishi ta lai you go or he come "Will you go, or will he come?"However, as previously noted, Li and Thompson (1981:520-21) have collapsed A-not-A and disjunctive questions under the single heading of disjunctive questions, and add a fourth category of "tag questions," exemplified by sentence (5) below.
(5) Women zou hao bu hao we go good NEG good "Let's go, ok?"In contrast, Helen Lin (1981:41-43) maintains the A-not-A and disjunctive distinction, and adds a second separate A-not-A category for certain interrogatives utilizing the copular verb shi, e.g.,
(6) Ni jintian shi bu shi bu tai shufu you today be NEG be NEG too comfortable "You're not feeling too well today, are you?"In any event, the separate category for ma-particle questions seems relatively non-controversial.
However, what remains unclear is which of those forms the ma-type may be most closely aligned with. So as to make the various proposals outlined in this section easier to follow, Table 1 provides a basic outline of the connections among types of interrogatives proposed by some scholars in the Chinese linguistic literature.
Table 1. Ma-particle interrogative categorization in the literature ========================================================= Question types subcategorized Criteria Works with ma-particle questions A-not-A Semantic Elliott equivalencece (1965) A-not-A, disjunctive Historical G.Y. Huang evidence (1986) WH-word, intonation Phonological Rand intonation) (1969) None or Sentence Y.W. Lin A-not-A, disjunctive structure (1985) None explicitly stated ---- Chao (1968) =========================================================In the early transformational treatments, Elliott (1965) and Rand (1969) take sharply opposing positions. In his comparative study of interrogation in English and Mandarin Chinese, Elliott (1965:90) sees a grammatical and semantic equivalence to A-not-A questions for a large group of ma questions, barring negative ma questions such as
(7) Ni bu qu ma you NEG go Q "Aren't you going?"The exact nature of this exception pertains to semantic and pragmatic factors, and will be discussed in sections 2.3 and 2.4.
Among other Western linguists, Rygaloff (1973:53) treats particle interrogatives and A-not-A questions as examples of "primary interrogation (of the sentence)" (interrogation primaire (de phrase)). More recent writers such as Norman (1988:169) have also lumped particle questions and disjunctive questions together as simpler and more complex types of yes-no questions, in contrast to question word (WH-word) interrogatives.
This analysis is consistent with the work of native Chinese writers including Lyu (1985), as well as G.Y. Huang (1986) and Liu (1988). Both of the latter two authors cite historical evidence in vernacular literature from northern China in particular to support the connection between particle interrogatives and A-not-A questions. Within that vernacular literature, one type of interrogative included a negative marker as the sentence-final morpheme, either bu, fou or wu. An example is given in (8), from G.Y. Huang (1986:133).
(8) ...mo you dang zhi shi zhe wu no one have proper worth envoy NOM NEG ..."is there none worthy to serve as envoy?"The final wu is argued by Huang to be the equivalent of "not A," where "A" equals the verbal phrase immediately preceding it. Over time, various elements of linguistic change, both phonological and lexical, led to the substitution of the morpheme ma in that sentence-final position, albeit in a written form different from the present-day graph for interrogative ma.[1] Ultimately, G.Y. Huang theorizes a process sometime prior to the Qing dynasty of "empty-ization" (xuhua) -- that is, a loss of the "not A" semantic quality originally associated with the sentence-final morphemes bu, fou and wu, and a substitution of the then-predominant ma [2] with the more strictly interrogative marker that exists today.
In contrast, espousing a formal analysis independent of semantic consideration, Rand rejects Elliott's linking of disjunctive and ma-type questions for three reasons, two of them semantically motivated. Instead, Rand chooses to classify ma-type questions under the heading of "intonation/particle questions,"[3] which together with question-word interrogatives make up the larger class of "constituent replacement" questions. Crossing the boundary of syntax, Rand's analysis of Mandarin interrogatives is based in part on intonation. This aspect of Rand's analysis as well as the question of intonation in general with respect to the ma-particle question will be discussed in the next section.
Later linguists are more categorically cautious in their attempts to categorize Mandarin questions. While treating ma-particle and A-not-A interrogatives as essentially equivalent, Li and Thompson (1979) argue that despite the inter-language evidence (Moravscik 1971), Chinese ma-particle questions are structurally less complicated and pragmatically less constrained than A-not-A questions. Both Chao (1968) and Tang (1986) are more noncommittal and make no such claims.
On the other extreme, Y.W. Lin (1985), drawing to a great extent from the earlier work of Lyu (1942), takes pains to provide three separate bases for categorization. On the basis of relation to declarative sentence structure, Lin states that ma-particle interrogatives make up one sub-class and WH-word, disjunctives and A-not-A questions another. With respect to interrogative-internal structure, Lin considers WH-word questions to be one sub-class and the remaining three another. This same subcategorization is proposed when the means of response is used as a basis for categorization.
Before leaving this discussion of categorization, it is worth mentioning C.T. Huang's (1987) categorization in his government and binding treatment of interrogatives. By analyzing disjunctive and WH-word interrogatives as one type of question, there is once again an implication that particle questions are typologically (and otherwise) fundamentally different from other Mandarin interrogatives.
2.3 Phonological considerations
There are two phonological elements in ma-particle questions which have been treated in the literature. The first and most frequently noted regards the sentence intonation pattern, specifically as to whether the overall pitch for the question rises, falls or stays at approximately the same level. The other issue, discussed directly by G.Y. Huang (1986) and indirectly by Zhu (1985) and Lin (1988), concerns the historical development of the ma particle, and the present phenomenon in southern Chinese dialects of a phonologically distinct morpheme with different semantic and pragmatic effect than the ma particle used in standard Chinese.
As previously noted, Rand claims a common feature for all constituent replacement questions. His label of constituent replacement questions refers to particle, echo (intonation) and WH-word questions, all of which either maintain a level pitch throughout the question, or else fall rather sharply, with a rise commencing on the last syllable with strong stress. However, Rand himself concedes that "this is rather impressionistic, as I have no supporting instrumental phonetic data" (1969:65).
The acoustic evidence seen in Wu (1982, cited in Lin, 1985) confirms Rand's impressionistic observation. Wu discovered that all four types of interrogatives (that is, yes-no or ma-particle, question word, A-not-A, and disjunctive interrogatives) can have rising (note not falling) intonation for the purpose of emphasis. This provides empirical confirmation for the somewhat confusing characterization of Lyu (based in part on an earlier (1975) informally published work) that seems to ultimately confirm that interrogative intonation is not necessary for any question form, yes-no questions included (Lyu 1985:250).
Liu (1988) is in greater initial agreement with Rand's intuitive judgement in stating that questions normally rise at the end. However, this agreement is a tempered one, as she continues by saying that ma-particle interrogatives can have level, and even falling, intonation (1988:25). Like Lyu, her ultimate conclusion is that so-called "interrogation intonation" is not a necessary condition in the ma-particle question, and that the "interrogation information" (yiwen xinxi ) is not the intonation pattern, but the particle, i.e. ma (1988:25). By "interrogation information," Lyu is referring to the element of the sentence which specifically indicates the interrogative (versus declarative) status of the sentence
Most recently, Shen (1989:26) has provided acoustic phonetic evidence that ma-particle and echo questions share an intonation pattern distinct from all other question types as well as declarative sentences. Both of these yes-no questions were found to have a starting point and ending point higher in overall pitch register than declarative sentences. Shen also notes (1989:10) that this condition was first observed by DeFrancis in the textbook Beginning Chinese, coincidentally one of the three textbooks to be analyzed in Chapter Seven.
Finally, we note once again the research of Huang (1986) into the ma-particle interrogative. By his analysis, there is a phonological distinction in southern Chinese dialects such as Yue, Kejia and Min, whereby separate morphemes are utilized for more purely interrogative and rhetorical purposes, as shown in Table 2 below.[4]
Table 2: Southern dialect interrogative morphemes ================================ Dialect Neutral Rhetorical Yue mo mE Kejia mo ma (me) Min bo ma (me) ================================According to Huang, the Yue interrogative morpheme and the Kejia rhetorical interrogative are written with the same character ma 嗎 as is under consideration here.[5] In the northern standard Mandarin dialect, however, which is more directly relevant to the present study, the single morpheme [ma] is employed by speakers either with or without presupposition.[6]
2.4 Semantic considerations
Predictably, earlier Chinese linguists have based their treatment of ma-type question semantics on the interrogative typology discussed above. Thus Elliott calls a question such as (9) semantically equivalent to (10).
(9) Ni lei ma you tired Q "Are you tired?" (10) Ni lei bu lei you tired NEG tired "Are you tired?"Anne Hashimoto seemingly echoes this analysis as she notes "with many speakers, the V-bu-V question may have an alternative ma-question form...with the same meaning and implication" (1971:110-11).
However, negative ma-particle interrogatives, e.g.,
(11) Ni bu lei ma you NEG tired Q "Aren't you tired?"are treated by Elliott and Hashimoto as well as other writers (e.g., Chao 1968) as somehow functionally distinct from positive forms (as in (9)). The problem with this purported difference is the inconsistency as to whether its character is semantic, pragmatic or a combination of the two. We will attempt here to focus on the semantic consideration and leave the pragmatic one for the next section.
In that spirit only Li and Thompson (1979) provide a semantic oriented account. In what they refer to as a "universal semantic explanation," appeal is made to an interpretation whereby form constrains meaning and thus use. That is, while disjunctive questions present two alternatives, equivalent to presenting a set and its complement for the respondent to choose from, yes-no questions are simply a proposition with the addition of an interrogative signal (1979: 205). Yet even while claiming this to be a "semantic" explanation, the authors' arguments are fundamentally based on the speech context and speakers' presumptions. This places their analysis as much within the realm of pragmatics as semantics, in seeming contradiction to their treatise's title.
Interestingly enough, despite his claim that his analysis excludes semantic considerations, Rand does include semantic considerations for ma-particle questions. For him, the ma particle "indicates the features of the underlying...adverbial pro-form, that is, by virtue of the features of force, certainty, suggestion and the like" (emphasis mine; 1969:67). This not only appears ad hoc in terms of structural description, but also introduces vague semantic notions into the syntax.
In short, semantic considerations of ma-particle questions by previous authors appear to lack a clear methodology, mixing with the syntactic and pragmatic components. It is that latter area of pragmatics with which this study is most deeply concerned, and to which we will turn now.
2.5 Pragmatic considerations
2.5.1 Pragmatic constraints: Three categories
As presented in the literature of Chinese linguistics from the past thirty years, pragmatic constraints on ma-particle questions can be divided into three categories.
(1) The context of usage in terms of neutrality (2) The questioner's assumption of the actual, "real world" situation (3) The questioner's assumption of the listener's response.There is a good deal of overlap among the three categories. Nonetheless, we will try to concentrate on each separately, beginning with the issue of neutrality of context.
2.5.2 Neutrality of context
Li and Thompson (1981) stipulate a separation between neutral and non-neutral contexts for ma-type question usage based upon the speaker's assumptions. The type of assumption will be dealt with in sections 2.5.3 and 2.5.4. For the present purpose, we will content ourselves with the definition of a neutral context being one in which there are no assumptions with respect to the proposition that is under consideration. Questioners operating within such a context merely wish to know whether or not the proposition is correct (1981:550).
Other writers have couched the neutral/non-neutral dichotomy in terms of a straight interrogative/rhetorical question division. Lin (1981:276) is particularly succinct in this respect, simply stating that ma-particle questions can be used for rhetorical questions as well as more purely information-gathering interrogatives.
In a similar conceptual vein but with a unique analytic approach, G.Y. Huang (1986) proposes a descriptive framework intended for a formal semantic analysis of interrogatives but explicitly directed toward describing the usage of such questions. He proposes a 5-value logic system between 0 and 1, where 0 equals "false" and 1 equals "true." His sense of false and true does not indicate the truth of the proposition per se, but rather the degree of presuppositional belief held by the speaker as to the truth of the proposition -- in essence, the speaker's perception of the real world. By Huang's system, a ma-type interrogative with a presuppositional truth value of 1/2 (that is, halfway between the 0 false and 1 true values) is a neutral interrogative; all other values (0, 1/4, 3/4 and 1) are so-called "false questions" (jiawen ju ), i.e. rhetorical or non-neutral. Potentially, Huang's system could lead to a meaningful, quantified assessment of questions reflecting assumptions of the speaker either equivalent to or the opposite of the statement contained within the question. Within the scope of his article, however, he merely sets up this system, and uses it only in the broadest terms to delineate between neutral and false (i.e., rhetorical) questions rather than provide any finer lines of distinction. Based upon those criteria, however, Huang clearly demonstrates a steadily growing preponderant percentage of ma-particle interrogatives used for rhetorical purposes. He employs a data base beginning with Yuan dynasty zaju ("miscellaneous plays"), an early dramatic art form that shows signs of a growing use of the vernacular language. His data base then continues through into the mid-1950s mainland Chinese fiction. From this data, Huang notes ma-type question usage as low as 2% during the Yuan, rising through 20% in representative Ming dynasty fiction works, to no less than 66% and as high as 90% in the 20th century pieces examined.[7]
Lyu (1980) is also fairly flexible in citing ma-type questions used in either situations of absolutely innocent and ignorant interrogation, or for more pointed, rhetorical purposes. Where Lyu differs from Lin and conforms to many of his predecessors, such as Chao (1968), is by suggesting the connection between negative form (i.e., the use of a negative adverb bu or mei) and rhetorical content. Lyu, however, only speaks of a slight tendency thereto, focussing more on the lack of neutrality when the sentence has a negative adverb.
Other Chinese linguists including both Elliott (1965) and Anne Hashimoto (1971) allow for both the presence and absence of presupposition when the ma-question is in the positive form, as in sentence (9), repeated below.
(9) Ni lei ma you tired Q "Are you tired?"However, for Elliott and Hashimoto, presupposition exists in the speaker's mind when he or she uses the negative ma-type question, as in sentence (11), also repeated below.
(11) Ni bu lei ma you NEG tired Q "Aren't you tired?"More recently, Liu has also asserted that ma-type interrogatives generally reflect presupposition in the mind of the speaker, particularly when compared to A-not-A questions. Still, Liu (1988:28,31) maintains that such questions can be used without presupposition in the mind of the speaker, especially when compared to intonation questions.
Problematically, however, Y.R. Chao's well-respected A Grammar of Spoken Chinese makes no such reference to a potential usage in a neutral context, asserting only that positive statements with a sentence-final ma-particle contain "either a slight or considerable doubt about an affirmative answer, implying a probability of less than 50%. When the question contains a negative adverb, it is a rhetorical question, suggesting a reply to the contrary" (1968:800). More recently, Tang (1986) is equally absolute in asserting that the particle question is used only in a non-neutral context, albeit leaving the type of assumption (either positive or negative) more flexible, as will be discussed in the next section. We are thus left with a dichotomy between Lin, Lyu and Li & Thompson on the one hand, espousing either a neutral or non-neutral ma-particle question, and Chao and Tang on the other, asserting that such questions are always, to a certain degree, non-neutral, with Elliott, Hashimoto, G.Y. Huang and Liu somewhere in the middle.
As for the nature of such non-neutral contexts, the fundamental issue is the purported presence of presumptions on behalf of the speaker. There are two types of presumptions, one of which is dealt with fairly extensively, albeit controversially.
2.5.3 Questioner's assumption of actual situation
Chinese linguists have been evenly split on the question of how the speaker's perception of the real world is expressed in ma-particle questions. For some such linguists, the presence of a negative adverb controls whether presupposition exists in the first place. Elliott asserts that "for positive ma question...the presupposition is neutral, corresponding to the neutral presupposition of the A-not-A question." In contrast, for negative ma-type questions, "the presumption...is simply the statement left when the ma is removed" (1965:93). In other words, by saying
(11) Ni bu lei ma you NEG tired Q "Aren't you tired?"the assumption by the questioner is that you are not tired. Elliott's latter remark is echoed in Hashimoto's 1971 work, although she makes no absolute connection between positive questions and presence of neutrality. In her words,
For many speakers, a ma-question with normal rising intonation in the affirmative or negative form presumes what is implied in the sentence...an affirmative ma-question implies an affirmative presupposition of the speaker, while a negative one implies a negative presupposition (1971:17)It should be noted that while the main thrust of Hashimoto's treatise stresses a presumption-based framework for the usage of ma-particle questions, the door is left open for neutrality in affirmative ma-type interrogatives, as discussed in section 2.5.2. Nonetheless, both Hashimoto and Elliott assert a tendency that the asking of question (11) implies a belief that the person in question is not tired, and question (9)
(9) Ni lei ma you tired Q "Are you tired?"implies a belief that the person is tired.
Elliott and Hashimoto stand in sharp contrast to Chao implicitly and to Li and Thompson explicitly. The latter two contend that in a non-neutral context, the statement contained within a ma-type interrogative involves a conflict with the presumption of the questioner (Li and Thompson 1981:552). Thus, for a sentence such as
(12) Ni chi pingguo ma you eat apple Q "Do you eat apples?"and in discourse segments such as
(13) Ni haoxiang shou - le yidian you seem thin - ASP a little "You seem to have lost a little weight." Shi ma Ni kan wo shou ma Wo ziji dao bu juede be Q you see I thin Q I self (emph) NEG feel "Is that so? Do you think I've lost weight? I haven't noticed it."(Li and Thompson 1979:201), they suggest that the speaker's assumption is the opposite of the statement contained within the question. Thus, for question (12), Li and Thompson would propose the speaker presumed the hearer did not eat apples, and for discourse segment (13), that the second speaker viewed him/herself to have not lost weight. With respect to sentence examples (9) and (11), the situation is in essence exactly the opposite of that proposed by Elliott and Hashimoto; in other words, one asks sentence (9) "Are you tired?" if one assumes the person is not tired, and sentence (11) "Aren't you tired?" (sentence (11)) if one assumes the person is tired. Lin (1981:43-44), in her discussion of emphasis-producing rhetorical questions, including ma-particle forms, also has a view that is similar to that of Li and Thompson. She states that for a rhetorical question, a negative expression of that question stresses the affirmative notion and a positive expression emphasizes the negative notion.
More recently, Tang (1986) has managed to circumvent the problem altogether by claiming particle questions can be used with positive or negative assumptions on behalf of the speaker, thereby suggesting that syntactic form does not always constrain the pragmatic effect. He makes a case for the possibility that "the particle question can be used when the questioner has a tacit assumption concerning the proposition which is being questioned, regardless of whether the speech situation is contrary to his assumption or not" (1986:7). By claiming that the particle question is used exclusively in a non-neutral context, Tang's argument is the ultimate simplification of the question of both neutrality and speaker assumption. Indeed, Tang takes Li and Thompson's own examples to prove that a ma-particle question asker can have either a positive or negative assumption as to the question at hand. For the previously cited discourse stretch (13), Li and Thompson create a situation in which the ma-particle question asker believes the real world situation to be the contrary of what his or her audience views it to be -- in short, that he or she is not skinny. Tang takes the same initial statement as in (13) and replaces Li and Thompson's response as follows:
(14) Ni haoxiang shou - le yidian you seem thin - ASP a little "You seem to have lost a little weight." Shi ma Ni kan wo shou ma Wo ziji ye zheyang xiang be Q you see I thin Q I self also this way think "Is that so? Do you think I've lost weight? I think so, too."(1986:7). Tang thus argues the asker of the ma-particle question to have a real world perception identical to that statement within his ma-particle question, in direct opposition to Li and Thompson's assertion. Liu has also asserted that when presupposition exists in the mind of the speaker, that presupposition can either be positive or negative (1988:28). G.Y. Huang (1986) likewise provides for several possibilities of speaker assumption in his multi-value logic system.
2.5.4 Questioner's assumption of hearer's response
Of all the Chinese linguists examined herein, only Chao (1968) takes into consideration a speaker's presuppositions of the hearer's response to a ma-particle question. To recapitulate Chao's remarks, for such yes-no questions without a negative adverb, there exists in the speakers' mind either a mild or considerable doubt about an affirmative answer, with the implication of a probability of less than 50%. As for those with the negative adverb, as previously cited, the role is that of a rhetorical question, with a suggestion that the reply will be to the contrary (1968:800). Simply stated, Chao's notion is that whatever form one asks it in, the response to one's ma-type question is presumed to have a better than 50-50 chance of being the opposite of the statement embedded within the original question.
It is true that Li and Thompson (1979) make a passing reference to the issue of speaker's response raised by Chao; to wit, "Neither our intuitions nor the data on natural discourse, whether from our own field work of from film-scripts and novels, support Chao's claim that the particle question implies doubt about an affirmative answer" (1979: 199). However, Li and Thompson weaken their argumentation by not elaborating on their intuitions, and by citing data that are so formulaic in character as to potentially invalidate its empirical worth, e.g.,
(15) Ni hao ma you good Q "How are you?"As cited by Li and Thompson, Rygaloff (1973) also contends the particle question involves speaker's expectation of a negative response. But these two authors are the exception to the rule of focusing on speaker (versus hearer) expectation as to the response.
2.6 Discussion
2.6.1 Classification of interrogatives
The analyses of Huang, Liu and Lyu provide compelling historical evidence for the development of particle questions from A-not-A interrogatives. Certainly within the realm of language universals, as discussed by Moravcsik, there is a strong justification for such a link. Given such diachronic justification, as well as the synchronic evidence of semantic equivalence (to be further discussed in Chapter Three), it thus seems all the more appropriate to deal with ma-type questions in conjunction with A-not-A interrogatives.
However, any attempt to categorize ma-particle questions with other types of interrogatives potentially undercuts the fact that there are indeed distinct question forms for different purposes. Bolinger (1972, cited in Li and Thompson, 1979:200) notes that "every contrast a language permits to survive is relevant some time or another." If we accept Bolinger's assumption, the very presence of ma-type questions indicates that they must serve a role which is functionally distinct from that of disjunctive, question-word or any other type of interrogatives. Such a distinction must extend as well to the pragmatic features associated with the particle interrogative.
2.6.2 Phonological considerations
The research by Shen provides acoustic confirmation of a distinction between the intonation patterns of ma-particle questions and other sentence types. It is certainly a more convincing argument than either Rand's or Liu's more intuitive judgements. To a very limited extent, it is consistent with Wu's empirical evidence regarding rising intonation patterns in all interrogatives. And it serves to contradict Rand's categorization of WH-questions together with ma-type interrogatives, albeit validating his claim that particle and echo questions make up a common class.
However, Wu's claim was that rising intonation was for the purpose of emphasis, which Shen makes no reference to. As the present study must necessarily restrict its scope, it is hoped that further phonetic study may be carried out to more conclusively determine the relationship between patterns of intonation for interrogatives in both stressed and non-stressed forms.
Huang's "division of labor" between phonetically distinct labial nasal-initial consonants in the southern non-Mandarin dialects presents a potential problem in an analysis of Mandarin Chinese spoken on Taiwan. Because there are multiple dialectal influences on Taiwan Mandarin, including Kejia and Min (two of the dialects discussed in Huang's study), we may well expect complications in an analysis of the ma-particle interrogative. This issue will be addressed in more detail in section 4.1.2.
2.6.3 Semantic considerations
Given the previously cited difficulties with the typology of ma-particle questions, an analysis based in large part on such kind of categorization is at best tenuous. Such is the case for the semantics of ma-type questions as presented up to this time in the literature. One can certainly find the neutral contexts in which ma-type and V-not-V questions are interchangeable. But interchangeability does not seem a sufficient means by which to characterize the meaning of ma-particle questions, entailing as it does an essentially circular argument.
It is beyond the scope of this study to construct a formal semantic analysis of ma-type interrogatives. However, one suggestion for a framework is that of sincerity conditions, as proposed by Hudson (1975:7) in his analysis of the syntax, semantics and pragmatics of English questions. Sincerity conditions for questions would entail not believing the proposition is true, but believing it may be true, and furthermore believing that the addressee is at least as likely to know whether or not it is true. However, at least three problems are apparent. First, as the data provided by Hashimoto and Elliott demonstrate, there are cases where the speaker of a ma-particle question does believe that the proposition is true. For example, from Elliott,
(7) Ni bu qu ma you NEG go Q "Aren't you going?"the presupposition that the listener is going is arguably correct, at least in certain contexts. Secondly, because it is designed to embrace rhetorical as well as more purely information-gathering interrogatives, Hudson's definition of sincerity conditions is so open-ended that there is little conceptual content to the notion. Finally, one senses that such sincerity conditions cannot be divorced from pragmatic influences of context. For if pragmatics is, as Levinson (1983) suggests, the relationship between signs and interpreters (that is, the users), then these sincerity conditions cannot be purely semantic.
In summary, what has passed for semantic analysis of ma-particle interrogatives is in fact pragmatics.
2.6.4 Pragmatic considerations
Whatever limitations their analysis may otherwise possess, Li and Thompson's neutral/non-neutral contextual dichotomy is a beginning for uncovering the pragmatic conditions for ma-particle question usage. But blanket generalizations such as Tang's, albeit well-motivated for its contextual orientation, are over-simplified and lack the empirical evidence that there is no case where the ma-type interrogative is used in a neutral context. Similarly, Chao's assertion that the ma-particle question is "not quite equivalent to the V-not-V question, which is completely noncommittal to the answer (1968:800)," oversimplifies the characterization of both types of interrogatives. There are in fact examples of actual speech of A-not-A questions being used in a clearly non-neutral context, as will be discussed in the next chapter.
There is a twofold problem that has emerged from the presentation here regarding pragmatic conditions surrounding ma-particle question usage. The first difficulty is the almost complete balance between opposing stands. This obviously presents a lack of consensus of opinion. Secondly, all of those authors previously cited, with the exception of Huang and Liu (and to a limited extent, Li and Thompson in their 1979 article), have carefully constructed discourse contexts providing purported proof as to the presuppositions of the ma-particle question asker in particular. By creating rather than gathering examples, and in some cases creating examples specifically for the purpose of disproving another linguist's theory, the data bases are clearly contrived. Such contrivance is not necessarily undesirable. However, if there are alternative sources of language extracted from natural speech, it is desirable to examine such sources.
As for Huang and Liu, the former in particular does provide substantial historical evidence of general trends in the development of particle interrogatives in both standard Chinese and the Chinese dialects (for example, Yue, Kejia and Min). The primary limitations on both Huang and Liu is that they have concentrated on the written language, and for the most part on literary works prior to 1950. As such, their value is limited for descriptions of the spoken language, which is the focus of the present study.
Beyond these more methodologically-oriented issues, there is an even more fundamental problem underlying the previous analyses of the pragmatics of Mandarin interrogatives. That problem has to do with the focus on speaker assumption as a basis for neutrality. It is impossible to know what presumptions the speaker holds when asking a question unless the speaker explicitly expresses those presumptions. Such explicit expression, while not unseen, is the exception rather than the rule. Therefore, barring a psycholinguistic probe (or at the least, a lie detector test) into the speaker's mind, how can one objectively prove a lack of presumptions concerning the proposition being questioned? That is to say, how can one empirically ascertain the presence of what would be called a "neutral context?" If the presence of presumptions cannot be absolutely ascertained, how can the quality of such an assumption (either positive or negative) be determined? Such issues make the presupposition-based definition of neutrality subject to reconsideration.
Despite such a limitation in previous analyses, there is a dual need to better characterize some sense of neutrality in Mandarin interrogatives. The first need is more purely theoretical, and is motivated by the divergence of opinion that has been presented in section 2.5. The second need is pedagogical, and has been spoken of in section 1.5. So long as our textbooks are unresolved as to whether a ma-particle interrogative reflects the presence or absence of speaker presupposition, we will not be able to teach our students as well as we potentially might.
2.7 The question posing/asking approach to neutrality
It is proposed that neutrality be analyzed in a spirit consistent with Lyons' distinction between posing a question and asking a question. Quoting Lyons (1977:755): "When we pose a question, we merely give expression to, or externalize, our doubt; and we can pose questions which we do not merely expect to remain unanswered, but which we know, or believe, to be unanswerable."[8] This element of doubt is crucial not only to the category of questions, but also with respect to the distinction between questions and declarative sentences. Lyons notes the similarity between yes-no questions (equivalent to ma-particle interrogatives in Mandarin) and declarative statements to be very strong. This suggests that for those languages in which such a distinction is drawn in the verbal component of sentences, the difference between declaratives and interrogatives lies precisely in the grammaticalization of the feature of doubt (1977:754).
In contrast, Lyons proposes "to ask a question of someone is both to pose a question, and in doing so, to give some indication to one's addressee that he is expected to respond by answering the question that is posed" (1977:755). It is important to note that the very same question could serve as both an instance of question posing and question asking, subject to the presence of the indication to one's addressee that he is expected to respond. One could take the question below
(16) Will there ever be a world without war?and hypothesize at the least two situations. One is that the speaker is externalizing his or her doubt regarding the possibility of an earth forever at peace, knowing full well that the average listener may not -- and indeed, can not -- provide a definitive answer. The other is that the speaker is actually seeking to elicit an answer from the listener, regardless of the inherent difficulty or even impossibility in answering the question. Lyons notes no such restrictions on answerability with respect to question asking. If the speaker is merely wondering as to the possibility of world peace, we may call it "question-posing;" if an answer is being sought from some listener, it is "question-asking." What seems crucial is that there is that unspecified indication eliciting a response directed at the listener. Yet Lyons himself is none too explicit as to where that indication lies; indeed, Lyons says the indication is not part of the question itself (1977:755). Nor does he provide any further elaboration as to just where that indication lies.
The nebulous quality of this answer-inducing indication notwithstanding, there are advantages to this approach. Lyons himself comments:
The advantage of this analysis of questions is that it is more general than their analysis as mands. It covers, not only information-seeking questions, but various kinds of rhetorical and didactic questions without obliging us to treat these as being in any respect abnormal or parasitic upon information-seeking questions. (1977:755)By divorcing the analysis of questions from unverifiable elements of presupposition and speaker intent, and in focusing on the discourse organization, a unique and more unified characterization of the question class as a whole may thus be obtained.
Turning now to the pragmatic implications, consider the two possible interpretations of the definition of question-asking. One is that we consider it solely in terms of the speaker, in that the speaker externalizes his doubt, and seeks a response. The other moves a step further and incorporates the listener as an indispensable part of the question-asking process. In fact, there is no way to distinguish a question-posing and question-asking interrogative unless the indication that one's addressee is expected to respond to the question has been successfully communicated, so that the listener has understood the indication, and responded. In other words, one cannot have a clear case of question-asking without a listener response, although for any number of reasons, the listener may not detect the indication that he or she is expected to respond.
In a pragmatic analysis, investigating the communicative use of language, it seems inadequate to talk only about a single speaker. Without the listener's response, in the form of an answer (or in the case of question-posing, no answer), we are not really seeing language in communicative use. It thus seems prudent that the listener must also be incorporated into any systematic investigation of interrogatives, as has been done for this study.
Thus, while Lyons' original framework for a question-posing/question-asking dichotomy was more semantic in orientation, we may extend it in such a way so as to be truly a discourse-oriented, pragmatic analytical tool. It is this concept that will serve as the analytical heart of the present study. Moreover, we can improve on Lyons' analysis by proposing that there are contextual correlates within the discourse which may indicate that the addressee is expected (or not expected) to respond.
Specifically, the following three contextual correlates are proposed. They were selected for their comparatively objective quality for purposes of identification as well as their reflection of the presence or absence of speaker presupposition.[9]
- No pause by questioner following question. By not allowing the listener time to immediately answer the question, one may well argue that the question asker has no compelling need or desire for a response, and as such poses, rather than asks, the question.
- Speaker responding to his or her own question. There were in fact very few instances of a question asker answering his or her own question -- specifically, two involving ma-particle questions, and one in an A-not-A question. However, it did seem to be a clear contextual indicator of a non-neutral, question-posing situation.
- Listener failure to verbally respond to the question. We will not attempt here to discuss the full range of possibilities signified by a lack of listener response to the question. Indeed the case may well be made that lack of any audible answer is as good pragmatically as a spoken answer, particularly when the presence of body language complicates the situation. However, such metalinguistic factors are at best difficult and at worst impossible to objectively gauge. Moreover, the very fact that a listener does not verbally respond suggests, at the least, a context exists wherein the listener either does not want to, or does not feel he or she needs to respond.
By analyzing speech for the presence of these three contextual correlates, we are led to a more quantitatively verifiable assessment of neutrality of discourse organization of the Mandarin interrogative, whereby we may characterize a question-asking interrogative as one which both seeks and elicits a response, regardless of the assumption held by the speaker or hearer. In terms of both pragmatic description and pedagogical necessity, the discourse organizational sense of neutrality along the question-posing/question-asking axis seems every bit as important as that relating to the speaker's assumptions.
This notion of question-asking does not of course solve the problem of analyzing the nature of assumptions in question-asking. Nor is this to suggest that assumptions are not important within pragmatic description. It is merely that in terms of empirical verifiability, they are not a viable line of inquiry at this time, as the review of the literature on ma-particle interrogatives presented in this chapter has demonstrated.
NOTES 1. Well into the early years of the twentieth century, the written form 么, homophonous with the morpheme ma 嗎 under consideration in this chapter, was used in written forms of what we would call "ma-particle interrogatives." See Liu (1988) for specific examples from the twentieth century playwright Cao Yu in particular.
2. That is, 么, as discussed in note 1 above.
3. The ongoing research of Robert Sanders also treats intonation and particle questions under a single classificatory heading.
4. However, Marjorie Chan (personal communication, 1990) says that Huang's transcription of the neutral interrogative morpheme in Yue as [mo] is incorrect, and should in fact be [ma]. See also note 5 below.
5. The picture is actually more complex. Marjorie Chan (personal communication, 1990), has pointed out that there is a separate morpheme [mou] 'not have' in Yue used in interrogatives asking "Have you eaten?" or "Did you eat?" This morpheme is graphically distinct from the ma-particle that Huang states is used in Cantonese (Yue) non-rhetorical interrogatives. The ma-particle in Cantonese is analyzed by Chao (1968:807) as a fusion of negative [m] 'not' plus sentence-final particle [a]. [mou], on the other hand, shares some of the same functions as those of Kejia and Min in Table 2. This discussion further emphasizes the complex division of labor that exists among Cantonese sentence particles. Chan also comments that textbooks that treat ma-particle questions as neutral may reflect the influence of different speakers with different dialect backgrounds where ma is used in neutral contexts. Another possibility is that they reflect an attempt to simplify the language by making such interrogatives a close correspondent to English yes-no questions.
6. Huang also proposes a second phonologically-based distinction between "pure" (i.e., non-rhetorical) and rhetorical questions. The former category can not have the final particle ma replaced with a, or its phonetic variants wa and ya. The variants are conditioned by the preceding syllable, as in hao wa or lai ya. Additionally, such non-rhetorical questions cannot be transformed into an intonation question. That is, they cannot be transformed into a question without a final particle but still clearly conveying interrogative meaning and purpose. He cites the following example (1986:132).
wo xiang liaojie de shi ta dong yingwen ma I think know NOM be he understand English Q "What I want to know is: does he understand English?"In the above sentence, replacement of the final clause with either *ta dong yingwen (with only interrogative intonation) or *ta dong yingwen a/na would be unacceptable if the question is non-rhetorical.
In contrast, a rhetorical question such as the one below
ni xiang dao tian'anmen guangchang qu ma you think to Tian'anmen square go Q "Do you want to go to Tian'anmen Square?"(1986:132) may be restated as either Ni xiang dao tian'anmen guangchang qu (intonation only) or Ni xiang dao tian'anmen guangchang qu ya(a). However, Chan (personal communication, 1990) observes that the first ma-particle interrogative above is an embedded question. Thus, it may be that the differences between the two are due not to a distinction between pure and rhetorical questions, but rather a distinction between direct and indirect questions.
7. As previously mentioned in note 1, there were graphs for the particle ma distinct from the one in general use today (that is, 嗎). However, historical evidence from G.Y. Huang and Liu (1988), among others, suggests that those graphs are simply written variants of the same basic spoken morpheme.
8. The question has been raised as to what the distinction is between this so-called "question-posing" and "rhetorical questions." I have decided to use Lyons' unique label (as I have also attempted to move away from the previously-used "neutral/non-neutral" dichotomy) so as minimize the reader's presuppositions concerning what is meant by a "rhetorical question." Such presuppositions have been unavoidably derived from our exposure to works on literary criticism, rhetoric, and so forth. It is true that "rhetorical questions" are in many cases a subset of "question-posing" interrogatives. But as will be made explicit, the definition of "question-posing" (and "question-asking") is discourse-oriented -- that is, attendant to the presence of certain features on behalf of both the speaker and the listener. "Rhetorical question," on the other hand, is sometimes defined as "a question to which no answer is expected" (this from The American Heritage Dictionary, Second College Edition (1982). But other definitions only make reference to a question "being used for rhetorical effect" without clarifying the matter of speaker response (the latter definition from Webster's New Students Dictionary (1964)). For purposes of clarity, I have thus selected the terminology in Lyons.
9. It may be noted that the correlates which are presented here are not part of the question per se, but rather features of the discourse immediately following the question. This is consistent with the analysis of Lyons, who stated that the indication that a response is required of the speaker was not part of question itself (1977:755).
[ MC's Homepage ]
[ Or click "BACK" on your web browser to return to the previous web page. ]
To cite this page:
McGinnis, Scott G. 1990 A Pragmatic Analysis of Mandarin Interrogatives: Data from Modern Taiwan Drama. Ph.D. dissertation, Ohio State University. (Excerpt: Chapter 2, pages 16-51.)
<http://people.cohums.ohio-state.edu/chan9/ling/mcginnis/chapter2.htm> [Accessed DATE]
Prepared by M. Chan on 9 October 1999.
Copyright (c) 1999 Scott G. McGinnis. All rights reserved.
URL:
http://people.cohums.ohio-state.edu/chan9/ling/mcginnis/chapter2.htm