HAL Id: hal-01774821
https://u-paris.hal.science/hal-01774821
Submitted on 24 Apr 2018
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Questions as indirect speech acts in surprise contexts
Agnès Celle
To cite this version:
Agnès Celle. Questions as indirect speech acts in surprise contexts. Tense, Aspect, Modality, and
Evidentiality: Crosslinguistic perspectives, John Benjamins, pp.211-236, 2018. �hal-01774821�
[Pre-final version. Please check with author before quoting. In press. Ayoun Dalila, Celle Agnès
& Lansari Laure (eds.) Tense, Aspect, Modality, and Evidentiality, Crosslinguistic perspectives,
pp. 211-236. Amsterdam, Philadelphia: John Benjamins [Studies in Language Companion Series
197]
Chapter 10
Questions as indirect speech acts in surprise contexts
1
Agnès Celle
Université Paris Diderot and University of Colorado
Abstract
This chapter offers an analysis of two types of interrogatives used as indirect speech acts in
surprise contexts in English – unresolvable questions and rhetorical questions. The function of
these questions is not to request information that is unknown to the speaker. It is argued that
surprise-induced unresolvable questions are expressive speech acts devoid of epistemic goals.
Surprise-induced rhetorical questions are shown not to suggest an obvious answer, but to request
a commitment update from the addressee. Adopting a schema-theoretic approach to surprise, it is
shown that unresolvable questions and rhetorical questions can express mirativity, the former at
the initial stage of the cognitive processing of unexpectedness, the latter at the last stage.
Keywords: rhetorical questions, conjectural questions, unresolvable questions, commitment,
mirativity, surprise, expressivity
1
The research leading to these results has received funding from the People Programme (Marie Curie
Actions) of the European Union’s Seventh Framework Programme (FP7/2007-2013) under REA grant
agreement n. PCOFUND-GA-2013-609102, through the PRESTIGE programme coordinated by Campus
France. This paper was made possible by the data annotation carried out with Hakima Benali, Anne
Jugnet, Laure Lansari and Emilie L’Hôte. I wish to thank Anne Jugnet, Laure Lansari and Tyler Peterson
for the discussions we had on questions. I am especially grateful to Anne Jugnet and Laure Lansari for
their feedback on previous drafts. I would also like to warmly thank the three anonymous reviewers for
their valuable comments and suggestions. Any remaining errors are my own.
1. Introduction
This chapter offers an analysis of interrogative structures used as indirect speech acts in surprise
contexts in English. Refining Littell et al.’s (2010) typology, I distinguish two different types of
interrogative structures: those that are mapped on the default interpretation of interrogatives, that
is, requests for information (analysed in Celle et al. forthc.), and those that are linked to other
speech acts. The latter correspond to indirect speech acts and include unresolvable questions and
rhetorical questions. These are the object of the present chapter.
In recent years, there has been a flurry of research into conjectural questions in
connection either with the conditional, the epistemic future and the subjunctive in Romance
languages (Diller 1977; Haillet 2001; Celle 2007; Rocci 2007; Dendale 2010; Bourova &
Dendale 2013; Azzopardi & Bres 2014) or with evidentials in languages that grammatically
encode evidentiality (Littell et al. 2010; San Roque et al. 2017). In Romance languages,
conjectural questions with the conditional, the epistemic future or the subjunctive are reported to
put forward an inference that the addressee is asked to evaluate. In Amerindian languages, the
insertion of an evidential into a question seems to give rise to a different meaning. Littell et al.
(2010: 92) claim that in three Amerindian languages with an evidential system, “the insertion of
a conjectural / inferential into a question creates a non-interrogative utterance, roughly
translatable using ‘I wonder’”. In this chapter, it is argued that emotive modifiers in English
cancel the interrogative force of a question in a similar way to those evidentials. Conjectural
questions in English are to be understood as expressive questions expressing wondering and
uncertainty. However, the label ‘conjectural’ may be misleading in this case, as these questions
do not form a conjecture, but rather implicate that it is impossible to resolve the question.
Therefore, I propose to label these questions unresolvable rather than conjectural.
Unresolvable questions and rhetorical questions do not constitute requests for
information. In the surprise contexts under scrutiny, unresolvable questions function as outcome-
related and speaker-oriented utterances expressing wonder and disbelief. In English, their
expressive function is marked by interjections, emotive modifiers and deictic items. Rhetorical
questions stand as argumentative tools questioning some prior surprising discourse entity or
extralinguistic event. Rhetorical questions in surprise contexts highlight the connection between
surprise and negatively-valenced emotions such as anger and disappointment.
The aim of this chapter is to determine the function of those questions that do not request
an answer in a surprise context. What do they tell us about the speaker-addressee relationship?
What is the relation between surprise and questions used as indirect speech acts? Are these
questions the linguistic expression of mirativity, and how do they relate to the cognitive
integration of unexpected new information? Section 2 presents the data. Section 3 is devoted to
surprise-induced unresolvable questions and section 4 to surprise-induced rhetorical questions.
2. The data
This study is part of a large-scale project on surprise
2
. It presents an analysis of verbal reactions
to surprising situations in the scripts of three movies (Ed Wood, War of the Worlds, Dr.
Strangelove)
3
. All surprising episodes were coded using the annotation tool Glozz
4
, based on the
same annotation scheme as Celle et al. (forthc.). The data used are enacted data. I am aware that
this type of data may bias the expression of emotions, actors being prone to overemphasise some
cues (Scherer et al. 2011: 409) as they relive an emotional experience of their own
5
. However,
enacted data allow recognizing emotions in a reliable way.
First, stage directions from the movie scripts can provide important environmental
information about the context and the experiencer’s emotional state. Second, emotions in movies
can be detected and identified through patterns of observable vocal, facial and bodily cues. On
the basis of experimentally-induced surprise reactions, Reisenzein (2000: 29) stresses that
surprise faces most frequently display only one of the facial components associated with
2
This study originates from the Emphiline project (ANR-11-EMCO-0005), “la surprise au sein de la
spontanéité des émotions: un vecteur de cognition élargie”, a project funded by the National Research
Agency from 2012 to 2015.
3
These movies provide a wealth of surprising episodes. The reasons for this choice are spelled out in more
detail in Celle et al. (forthc.).
4
Glozz is an annotation tool designed by Yann Mathet and Antoine Widlöcher. The annotation scheme
relies on units, relations and schemas. http://www.glozz.org
5
Nonetheless, on the basis of experimental studies comparing enacted and naturally-induced vocal data,
Scherer et al. (2011: 409) stress that the two procedures yield similar results, which suggests that “it may
not matter very much whether emotional expressions are enacted or experimentally induced, at least for
some major emotions”.
surprise: eyebrow raising, eye widening, or mouth opening, while two- or three-component
displays are less frequent. He further points out that this finding is in keeping with Carroll and
Russell’s (1997) enacted data based on the facial displays of surprise shown by movie actors.
Even if the present chapter does not aim to analyze intonation and gestures, those parameters
were taken into account in our annotation scheme and facilitated emotion recognition. The
semasiological perspective adopted is thus combined with an onomasiological approach, that is,
only interrogatives occurring in surprise contexts were considered.
Interrogative clauses used as indirect speech acts are questions that do not request an
answer from the addressee, although they may call for some response from the addressee. They
amount to 13% of all interrogatives in our sample (26 utterances out of a total of 146
interrogatives). These interrogative clauses are subdivided into rhetorical questions (n = 12),
unresolvable questions (n = 5), and clarification requests (n = 9)
6
. Like the interrogative clauses
used as direct speech acts examined in Celle et al. (forthc.), the interrogative clauses found in our
sample are triggered by some surprising event
7
. Unlike their counterparts used as direct speech
acts, however, they have no force of inquiry. The connection between surprise and interrogative
clauses used as indirect speech acts needs to be accounted for; so does the nature of the speaker-
addressee relationship when no answer is requested. I first examine unresolvable questions
before moving on to rhetorical questions.
3. Unresolvable questions
Unresolvable questions are the least frequent category of surprise-induced questions in our
sample (n= 5). This category is based on the conjectural question type put forward by Littell et
al. (2010) to account for the wonder effect produced by the insertion of a conjectural / inferential
evidential into a question in three Amerindian languages. Littell et al. maintain that conjectural
questions are wonder-like statements, although formally, they are wh-interrogatives. The claim
6
Clarification requests straddle the border between direct and indirect speech acts. As shown by Celle et
al. (forthc.), clarification requests may be used as indirect speech acts in a purely expressive way.
7
Embedded interrogatives are left aside in this paper. For a comparison of the uses of root interrogatives
and embedded interrogatives, see Celle (2009).
made in this paper is that a similar wonder effect is produced by the insertion of emotive
modifiers into a question in English. However, strictly speaking, these questions do not express a
conjecture in the sense that they cannot be rephrased using “I surmise / I presume“ + content
clause
8
. It is the proposition as a whole that is a matter of conjecture. In and of itself, this
question type reflects the speaker’s ignorance rather than their conjecture. Therefore, I propose
to dub these questions unresolvable rather than conjectural. Emotive modifiers indicate that the
situation is appraised as violating the speaker’s expectations to such an extent that the question-
answer presupposition is cancelled.
In the surprise contexts that were examined for the present study, unresolvable questions
may contain interjections (such as ‘shit’, ‘gosh’), emotive modifiers (such as ‘on earth’ or ‘the
hell’)
9
or deictics, that is, items that mark the speaker’s emotional involvement and context-
boundedness (see Ameka 1992: 108). Some of these items are swearwords (‘gosh’, ‘shit’, ‘the
hell’) used cathartically (see Pinker 2007), that is, they serve an intra-individual function by
reducing the stress associated with the utterance situation once it has been appraised as
discrepant. Both interjections and emotive modifiers take the utterance situation as the source of
surprise. As noted by Vingerhoets, (2013: 290), “cathartic swearing is regarded as an
adaptation,
10
especially meant to communicate that the situation we are confronted with deeply
affects us, as evidenced by the display of strong emotions”.
8
In French, the inferential conditional used in an interrogative clause produces a conjectural question,
and not an unresolved question: Or cet enfant venait d’être volé par un inconnu. Quel pouvait être cet
inconnu ? Serait-ce Jean Valjean ?[The child had just been stolen by an unknown man. Who could that
unknown man be? Could it be Jean Valjean?] (Hugo, 1862, Frantext). As stated by Dendale (2010: 297;
302), the “interlocutive function” of the question is affected by the conditional. However, the reason for
the weakening of the interrogative force is that the speaker believes the proposition to be true. The
conjectural question may be considered a mitigated assertion that can be rephrased as “I surmise that p”
(Je suppose que c’est Jean Valjean). As shown by Diller (1977: 3-4), the conditional conveys a
presupposition of evidence that is superimposed on the question, which reduces its interrogative force.
She argues that a conjectural question in the conditional asserts a presupposition. I claim that a
conjectural question seeks the addressee’s commitment (Celle 2007). By contrast, expressives in
unresolvable questions are triggered by defective evidence. They implicate that no value can instantiate
the question variable, which precludes assertion. This can be paraphrased as ”I don’t know if p ; I don’t
know where / what / how …”.
9
The distinction between interjections and emotive modifiers is borrowed from Huddleston & Pullum
(2002: 916).
10
My emphasis. The adaptation marked by swearing may be regarded as the verbal expression of the
cognitive and emotional adaptation to unexpected events that underlies surprise (see Darwin 1872/ 1965).
Adjustment to direct evidence is a feature shared by questions with emotive modifiers in
English and conjectural questions in the languages that have evidentials or inferential
conditionals. It gives credence to the claim that unresolvable questions are a cross-linguistic
phenomenon that can be extended to English. Indeed, emotive modifiers point to defective
evidence about the potential answers to the question, so that the addressee cannot be expected to
provide an answer.
The fact that these highly emotional questions are systematically content questions
suggests that a correlation can be established between speaker perspective and wh-questions (as
shown by Celle & al. (forthc.) in the case of interrogatives used as direct speech acts, and by San
Roque et al. (2017) in the case of evidential questions). This correlation is all the more striking
as the most frequent questions in standard communication contexts in English are polar questions
(see Stivers 2010; Siemund 2017). This suggests that the more emotional a question is, the more
open-ended the set of answers will be. Unresolvable questions used in surprise contexts are about
a salient open proposition. The clash between the speaker’s expectations and the incongruous
character of the situation makes it impossible for the speaker to assign a value to the question
variable and to expect the addressee to be able to do so.
In Littell et al’s (2010) typology, conjectural questions differ from rhetorical questions in
that the speaker does not know the answer; they also differ from ordinary questions in so far as
they do not require an answer from an addressee. This holds true for unresolvable questions in
English:
(1) Ed : Whoa, look at this camel, this is a real camel, Gosh, where’d they get a real
camel?
(2) Bela: Oh, there's my bus. [he checks his pockets] Shit, where's my transfer?!
Ed: Don't you have a car?
(3) EXECUTIVE 1 What the hell is this?!
EXECUTIVE 2 Is this an actual movie?!
These questions can be rephrased as follows:
I don’t know / wonder where they got a real camel.
where my transfer is.
what this is.
The source of wondering is the unexpected presence, absence or location of some element in the
utterance situation. In other words, the cause of surprise is an extralinguistic event that violates
the speaker’s expectations. It is some new environmental information (see Peterson 2017) that
may be surprising to both speaker and addressee. As such, it constitutes defective evidence, to
the point that the experiencer cannot make inferences about the situation (Stein & Hernandez
2007: 302). The experiencer is forced to revise his / her previously held beliefs. In (1), there is a
real camel on stage although the speaker assumes there should not be one; in (2), the transfer is
not in the speaker’s pocket although it should be there; in (3), the properties of the movie defy
the speaker’s ability to characterize it. At the same time, the specific contribution of the
interjection or the emotive modifier is that evidence is so defective that the addressee is not
expected to know the answer.
These questions may even be self-addressed as in (1), where there is no addressee. In (2)
and (3), no answer is provided by the addressee who responds by asking a biased question or a
rhetorical question, and the interchange is perfectly felicitous. In (3), the follow up polar
question restricts the set of possible values for the question variable and specifies the nature and
quality of the entity that both speaker and addressee find surprising.
Pragmatically, unresolvable questions are speaker-oriented, like exclamative utterances.
Huddleston and Pullum (2002: 916) stress that emotive modifiers “express surprise or
bafflement, and hence suggest that the speaker does not know the answer to the question. They
tend to emphasise the open-endedness of the set of possible values for the question variable.”
Noteworthy is the fact that they may be followed by a question mark as well as an exclamation
mark. However, these questions have the syntax and the semantics of interrogatives.
Syntactically, they require subject auxiliary inversion. Semantically, they are concerned
with the identification and the appraisal of an incongruous situation
11
, not with degree. Unlike
most exclamatives, unresolvable questions carry no explicit “scalar implicature” (Michaelis and
Lambrecht, 1996: 378)
12
. However, they do imply an implicit scale by suggesting that the actual
state of affairs violates the speaker’s expectations or norms in an extreme way.
11
The concept of “incongruity judgement” was originally coined by Kay & Fillmore (1999).
12
A parallel may be drawn here between unresolvable questions and the WXDY construction as defined
by Kay and Fillmore (1999: 25-26). However, in the case of unresolvable questions, mirative meaning is
related to their deictic nature. As shown by Kay and Fillmore, WXDY constructions can express a sense
of incongruity independently of the situation of utterance as they can be embedded. See also Celle &
Unresolvable questions containing interjections may be differentiated from those
containing emotive modifiers for two reasons. First, the valence associated with interjections
may be either positive (1) or negative (2), while emotive modifiers tend to be associated with a
negative valence (3). Second, emotive modifiers convey a stronger expressive meaning than
interjections, which has implications on the function of the speech act. Unresolvable questions
containing interjections allow continuation with an informative answer, although they do not
request an answer:
(1’) A - Whoa, look at this camel, this is a real camel, Gosh, where’d they get a real
camel?
B - In the Sahara.
(2’) A - Oh, there's my bus. Shit, where's my transfer?!
B – You must have left it at home.
As stated by Ameka (1992: 107) interjections “encode speaker attitudes and communicative
intentions and are context-bound”. They express the speaker’s emotional reaction to some
unexpected event: the presence of a real camel in (1), the absence of the transfer in (2). The
transfer is not where it is expected to be, the real camel is unexpected in this setting, hence the
unresolvable questions about the origin of the camel in (1) and about the location of the transfer
in (2). The answer may well increase the speaker’s knowledge by assigning a value to the place
variable in (1’) and (2’). However, the aim of the question is not to increase the speaker’s
knowledge, but to express the emotional reaction of the speaker faced with an unexpected
discrepant situation. Whatever the answer, it does not eliminate the sense of surprise.
Emotive modifiers provide questions with a strong expressive force, which overrides
referential meaning. ‘The hell’ systematically follows the wh-word and the sequence “wh-word
the hell” is a semi-fixed phrase. In (3), the speaker’s question is triggered by visually perceived
incongruous evidence. As the speaker is witnessing the situation, the addressee’s answer is
redundant in the sense that it does not increase the speaker’s knowledge:
(3’) ‘What the hell is this?!’
‘This is an actual movie.’
Lansari (2015) on aller + infinitive in one of its uses. This calls for further research into the connection
between mirative constructions (possibly unrelated to the speaker’s here and now) and mirative utterances
(deictically related to the speaker).
Paradoxically, this referential answer is also insufficient because it fails to account for the
incongruous character of the state of affairs. Providing a value for the variable is not enough to
account for the incongruity of the state of affairs. The unresolvable question is not about the
identification of the situation the speaker is witnessing. It conveys a negative assessment of the
film movie that is being watched because the movie does not meet the standards of an actual
movie. The answer can eliminate neither the sense of incongruity nor the negative assessment.
Huddleston and Pullum (2002: 916) rightly note that ‘which’ cannot substitute for ‘what’
when a what-question contains an emotive modifier. According to them, the open-endedness of
the set of possible values implied by emotive modifiers accounts for that restriction: “as a result,
they are hardly compatible with which, for this involves selection from an identifiable set”. This
reasoning can be carried one step further. I argue that these questions are not concerned with the
identification of a referential entity. The emotive modifier the hell cancels the question-answer
presupposition by suggesting that whatever the answer, the situation violates the speaker’s
expectations to such an extent that their state of surprise and negative assessment cannot be
altered
13
. Given the extreme character of the situation, any informative answer is epistemically
pointless. This type of question constitutes an expressive speech act devoid of any epistemic
goals (see Zaefferer 2001: 224).
Interjections and emotive modifiers impart a mirative meaning to unresolvable questions,
although mirativity is not encoded morphosyntactically in English. As stressed by DeLancey
(2001: 377-378), mirativity is a “covert semantic category” in English, as opposed to other
languages. However, interjections and emotive modifiers do encode the speaker’s surprise and
relate it to new environmental information. Strikingly, all the unresolvable questions in our
sample are induced by new environmental information and not by a surprising discourse entity.
13
In the surprise contexts studied in this paper, the wh-word-‘the-hell’ phrase indicates that the cause of
surprise is the utterance situation. In such contexts, I argue that the phrase cancels the question-answer
presupposition. However, the wh-word-the-hell phrase may be used in requests for information that
simultaneously carry an instruction of unresolvedness, especially when reference is made to a past event.
In the following example borrowed from den Dikken & Giannakidou (2002: 32), the interrogative is an
information question about the identity of the buyer: ‘Who the hell bought that book?’ As stated by den
Dikken and Giannakidou (2002: 32), the wh-word-‘the-hell’ phrase conveys the presupposition that
‘Nobody was supposed to buy the book’. This results from the instruction of unresolvedness (i.e. the
speaker's failed attempt to resolve the question), which is nonetheless compatible with a genuine
information question about the identity of the buyer as the event is presumed to have occurred in the past.
By contrast, questions used as direct speech acts in surprise contexts (i.e. clarification requests,
ordinary questions and inferential questions) are mainly induced by a surprising discourse entity
(see Celle & al. forthc.) and therefore do not qualify as mirative utterances (see Peterson 2017:
68).
The incompatibility of emotive modifiers (such as ‘the hell’, ‘on earth’) with echo
questions Fillmore (1985: 82) and with which-questions observed by Huddleston & Pullum
(2002: 916) and Pesetsky (1987: 111) substantiates this claim by suggesting that the surprising
element cannot be traced back to the previous discourse. To quote Pesetsky (1987: 111), ‘the hell
forces “a non-discourse-linked reading”, while “which-phrases are discourse-linked”. It is the
context-boundedness of emotive modifiers that allows for the indexical mirative meaning of
unresolvable questions.
The mirative meaning of this type of question can be probed using the undeniability test
(Rett & Murray 2013: 455; Celle & al. 2017: 220):
(3’’) A. ‘What the hell is this?!’
B. # ‘You are not surprised’.
The sense of surprise cannot be denied, which reveals that this utterance is an expressive speech
act (see Potts 2005: 157). As such, the mirative speech act can only reflect the speaker’s
emotional state, and the addressee cannot deny that emotional state.
Interestingly, Alcázar (2017: 37) points out that in Basque, the mirative particle ote often
collocates with swearwords equivalent to ‘the hell’ in questions of the type ‘Can’t-find-the-
value-of-x’. There seems to be typological evidence that the use of emotive modifiers in
unresolvable questions pertains to mirativity. Like evidentials in some languages, emotive
modifiers may shift the interpretation of questions, which take on an ignorance meaning.
14
This
shows that they have not only an intensifying or emphatic function in questions (Hoeksema &
Nicoli 2008). They also have an illocutionary effect on interrogatives.
4. Surprise-induced rhetorical questions
14
However, more research is needed to elucidate the relation between indexicals, evidentials and emotive
modifiers. Some scholars (see for example Korotkova 2016: 224-226) argue that evidentials are
addressee-oriented in questions, in contrast to indexicals which remain speaker-oriented.
4.1. Expectation violation
Formally, rhetorical questions resemble questions used as direct speech acts. Surprise-induced
rhetorical questions can enter into the same syntactic patterns as direct questions, except the
declarative pattern. Rhetorical questions have an interrogative syntax in a more systematic way
than ordinary questions. Wh-questions with subject-auxiliary inversion account for half of all the
interrogative clauses used as indirect speech acts in my sample. The different types of rhetorical
questions are classified in Table 1.
[Insert Table 1 here]
Table 1: Types of rhetorical questions
Wh-question: Wh-word +
subject-auxiliary inversion
What kind of sick mind would operate like that?
5
Yes/no question
Can you imagine what that guy would be like in
a movie?
4
Reprise fragment
15
Stronger? You see! You see!! You stupid minds!
Stupid!
1
Wh- question: clause
fragment
And what about this so-called “Barbara”
character? It’s obviously ME!
1
Reprise sluice
Since when?
1
In total, wh-questions (including a reprise sluice and a verbless sentence) are found in 8 out of 12
rhetorical questions. This confirms the correlation established above as well as in Celle et al.
15
This term is borrowed from Ginzburg (2012).
(forthc.) between surprise and wh-questions, that is, questions that denote a set of possible
answers. Wh-words include ‘what’ (n = 3), ‘when’ (n = 2), ‘how’ (n = 3). ‘How’-questions are
always associated with a modal auxiliary in their rhetorical reading. Among the elements that
facilitate a rhetorical reading are also deictic items (‘like that’) and degree words (‘so casual’).
Semantically, rhetorical questions define either a closed set of possible answers or an
open set of possible answers, the latter being the most frequent case in my sample.
Informationally, rhetorical questions are complete utterances, as opposed to ordinary questions.
Pragmatically, they express a biased position by pointing towards an obvious answer (Rohde
2006:149). As pointed out by Caponigro and Sprouse (2007: 131), “[r]hetorical questions are not
asked to trigger an increase in the amount of mutual knowledge”, nor do they assert anything
new. This raises two questions that are addressed below. First, how can rhetorical questions
qualify as questions? Second, why are rhetorical questions used to express surprise? Such
expressions reveal an epistemic asymmetry, the speaker’s expectations conflicting with the
addressee’s beliefs or with the state of affairs. In her analysis of responses to rhetorical
questions, Rohde (2006: 142) notes that rhetorical questions “generate very little surprise”:
“[T]he case of complete lack of surprise corresponds to rhetorical questions because the answer
is predictable to both the Speaker and the Addressee. The answer is so unsurprising that it need
not be uttered at all.” (ibid: 147).
My analysis of responses to rhetorical questions yields similar results: the answer to the
rhetorical questions found in surprise contexts need not be uttered
16
. However, it may not be
because the answer is unsurprising. Under Rohde’s analysis (2006: 152), the fact that rhetorical
questions generate little surprise is “evidence of their uninformativity”. Adopting Gunlogson’s
(2001) Common Ground theoretical framework, Rohde (2006: 152) views this uninformativity as
indicative of the fact that rhetorical questions “require no update to participants' commitment
sets”. This view is challenged in the present chapter. It is also argued that informativity and
update of the participants’ commitment sets should be distinguished. In surprise-generated
rhetorical questions, no informative answer is requested, but rhetorical questions are uttered in
reaction to some unexpected linguistic information or incongruous situation and involve a two-
16
Rohde (2006) uses naturally-occurring language data drawn from the Switchboard corpus, a corpus of
telephone speech. Her approach is purely semasiological - i.e. her analysis is not focused on rhetorical
questions occurring in surprise contexts.
fold update. First, they signal the speaker’s attempt to cognitively integrate unexpected new
information. Second, they request a commitment update on the part of the addressee. With
respect to surprise, Rohde only examines responses to rhetorical questions without considering
how these questions may lend themselves to the expression of surprise. I contend that rhetorical
questions may be used to express surprise precisely because the nature of the speech act they
convey allows for the expression of conflicting views in a questioning process whereby the
addressee is asked to update his / her commitment.
My claim is that this pragmatic commitment update process is highly congruent with the
appraisal process that underlies the surprise reaction on the psychological level. Within a
schema-theoretic framework, Meyer et al. (1997: 253) characterize the surprise-induced
appraisal process as follows:
[S]urprise-eliciting events initiate a series of processes that begin with the appraisal of a
cognized event as exceeding some threshold value of schema-discrepancy (or
unexpectedness), continue with the occurrence of a surprise experience and,
simultaneously, the interruption of ongoing information processing and reallocation of
processing resources to (i.e. the focusing of attention on) the schema-discrepant event,
and culminate in an analysis and evaluation of this event plus – if deemed necessary – an
updating, extension, or revision of the relevant schema.
The questioning process encoded by rhetorical questions necessarily requests a commitment
update from the addressee. Furthermore, surprise-induced rhetorical questions appear to be much
more complex emotionally than surprise-induced ordinary questions. Surprise may be tinged
with anger in rhetorical questions that are typically asked to express disbelief and disagreement.
In that case, expectation violation is coupled with the violation of standards and the thwarting of
the experiencer’s goals, which correspond to the ingredients of anger as defined by Ortony and
al. (1988: 152-153).
4.2. Informative answers
Rhetorical questions are generally said to be semantically equivalent to statements because they
contain the answer to the question they ask and do not request an answer from the addressee. The
view upheld in the present chapter is that rhetorical questions necessitate a pragmatic account.
Even when the rhetorical intent is obvious from the pragmatic context and the semantic
construction, the addressee may fail or deliberately refuse to recognize it. In that case, an
informative answer may be provided:
(4) Reverend Lemon: Mr. Wood? What do you think you're doing?!
Ed: I'm directing.
Reynolds: Not like THAT, you're not!
(5) Rachel : I'm allergic to peanut butter.
Ray : (laughs) Since when?
Rachel : (with a snotty look) Birth!
These questions fail the tests designed by Caponigro & Sprouse (2007) to reveal information-
seeking questions
17
:
(4’) # I’m really curious: What do you think you're doing?!
# I really don’t know: What do you think you’re doing?!
(5’) # I’m really curious: Since when (have you been allergic to peanut butter)?
# I really don’t know: Since when (have you been allergic to peanut butter)?
In addition to these tests, rhetorical meaning is also revealed by certain interrogative phrases.
When as a complement of the preposition since is less likely to request an informative answer
that selects a temporal starting point. It suggests a sudden start that may not be relevant to some
states, such as being allergic
18
. As noted by Huddleston & Pullum (2002: 905), “‘since when’ is
often used sarcastically, with cancellation of the presupposition”. In the following example,
‘since when’ points to a rhetorical reading in a similar way:
(6) Robbie: I don’t have a license.
Ray: Since when has that stopped you?
The ‘since when’-question is not about the starting point of the stopping process but cancels the
question-answer presupposition ‘That has stopped you for some time’. The rhetorical question
implies ‘That has never stopped you’. Consequently, Robbie’s statement ‘I don’t have a license’
loses its argumentative force as a justification for not driving.
Contrast with a ‘how long’-question:
(7) A - I’m allergic to peanut butter.
17
Except in sarcastic contexts where they are felicitous, as pointed out by an anonymous reviewer.
18
As rightly pointed out by an anonymous reviewer.
B - How long have you been allergic to peanut butter?
This question is accompanied by a question-answer presupposition (‘you have been allergic to
peanut butter for some time’) and seeks to assign a value to a variable in an open proposition
(‘you have been allergic to peanut butter since X’).
Although the questions in (4) and (5) are unequivocally rhetorical, they are followed by
an informative answer. Such answers are not requested. However, they are perfectly compatible
with rhetorical questions syntactically and conversationally because rhetorical questions are
questions, not assertive statements. They are made possible by the semantic nature of these
questions, which denote a set of potential answers (‘playing’, ‘working’, ‘directing’, etc. in (4),
‘since birth’, ‘since 1960’, ‘since the war’, etc. in (5)). Nonetheless, such answers refute the
rhetorical scenario on the pragmatic level and express disagreement. Instead of acknowledging
the rhetorical intent whereby the speaker points to some answer supposedly obvious to both
speaker and addressee (i.e. ‘whatever you think you’re doing is wrong, you have never been
allergic to peanut butter’), the addressee may well choose to assign a value to the variable as if
the question were information-seeking.
According to Rohde (2006: 161), rhetorical questions are understood as such by virtue of
their properties of “answer obviousness” and “similarity”, that is, the answer to a rhetorical
question is obvious to speaker and addressee and they supposedly both share a commitment to an
answer similar in nature. However, an unco-operative addressee may reject or ignore the
rhetorical effect imposed by the speaker even if it is recognized as such. It is noteworthy that
informative answers are found in the case of second-person utterances – or at least in questions
involving the second person as in the reprise sluice in (5). These rhetorical questions are indirect
speech acts that challenge either what the addressee is doing (as in (4)), or what the addressee
has just said (as in (5)) because speaker and addressee do not share the same standards or the
same beliefs. The sense of absurdity conveyed by the rhetorical questions may not be shared by
the addressee. Providing an informative answer amounts to assigning a value to a variable as in
the case of an ordinary question. In this way, the addressee avoids committing to the proposition
that is indirectly asserted by the rhetorical question (‘You are messing up’ in (4), ‘You have
never been allergic to peanut butter’ in (5)). The rhetorical question then fails to update the
addressee’s commitment.
4.4. The surprise-induced rhetorical scenario
Rhetorical questions may be followed by an informative answer, although they do not invite such
an answer. More often than not, they are followed by a response. Each case is examined in turn.
(8) Dolores: Ugh! How can you act so casual, when you're dressed like that?!
Ed: It makes me comfortable.
The answer can be construed as the causal explanation for acting so casually. It indicates that the
question is taken to carry a presupposition (‘you act very casually when you’re dressed like that’)
while the implied assertive statement, that is, the implicit evaluative judgment implied by the
question (‘you shouldn’t act so casual when you’re dressed like that’), is ignored, which foils the
rhetorical strategy. The rhetorical strategy fails in a similar way in the following constructed
examples:
(9) A - Ugh! How can you act so casual, when you're dressed like that?!
B – Thanks to my talent.
(10) A - How can you just walk around like that, in front of all these people?
B – With a walking stick.
(11) A - Goldie, how many times have I told you guys that I don't want no horsin'
around on the airplane?
B – Just once.
These answers signal that the addressee deliberately ignores the rhetorical scenario imposed by
the speaker. By contrast, responses do not attempt to undermine the rhetorical strategy, but take
disagreement for granted:
(12) Dolores: How can you just walk around like that, in front of all these people?
Ed: Hon', nobody's bothered but you.
(13) Kong: Goldie, how many times have I told you guys that I don't want no horsin'
around on the airplane?
Goldie: I'm not horsin' around, sir, that's how it decodes.
In (12) and (13), the responses signal the addressee’s disagreement with the evaluative
judgments expressed in the rhetorical questions. As such, these rhetorical questions force the
addressee to accept the implied assertive statement, and do not request a response. First, I define
the nature of the discrepancy conveyed by rhetorical questions before taking up the issue of
addressee commitment.
Surprise-induced rhetorical questions express a conflict between the speaker’s epistemic
domain and the actual state of affairs: in (13), the rhetorical question implies an indirect assertive
statement: ‘I have told you so many times that I don’t want no horsin’ around on the airplane’.
The source of the speaker’s surprise is the soldier’s preceding answer, which confirms surprising
information. This answer is mistakenly construed as a joke, i.e. as an act of disobedience. The
rhetorical question indicates that the speaker is epistemically unprepared to face a totally
unexpected turn of events and is therefore unable to behaviorally adapt to it. Discrepancy arises
from the speaker’s failure to correctly interpret an unexpected answer. Some fact is directly
perceived but misinterpreted.
When rhetorical questions contain modal auxiliaries, they typically express a conflict
between realis and irrealis:
(14) What kind of sick mind would operate like that?
The modal ‘would’ occurs with question-answer presupposition cancellation, that is, the
presupposition that some value can be supplied for the subject variable is cancelled by
modality.
19
The rhetorical question implies an indirect assertive statement, namely that ‘no sound
mind would operate like that’. In reaction to an actual situation for which the speaker is
epistemically and morally unprepared, a hypothetical stance is adopted that strips the surprising
situation of its realis quality. Surprise is related to the irrealis domain (see Akatsuka 1985). This
sense of reality denial is particularly striking with how-rhetorical questions, which always
contain the modal auxiliary can in my sample (as in (8) or (12)).
As noted by Desmets and Gautier (2009 : 109), comment ‘how’-rhetorical questions in
French (such as Comment peux-tu déambuler de cette façon?) contain two contradictory pieces
of information (tu déambules de cette façon and tu ne peux pas déambuler de cette façon). The
same analysis holds for how-rhetorical questions in English. Indeed, (8) and (12) carry a
presupposition that p (‘you are walking around like that, with such accessories’ in (12), and ‘you
are acting so casual’ in (8)) that conflicts with the negative comment implied by the rhetorical
19
For a detailed account of would in questions, see Celle (in press) and Celle & Lansari (2014) and
(2016).
questions (‘you can’t walk around like that, in front all these people’ in (12) and ‘you can’t act so
casual, when you’re dressed like that’ in (8)).
Counterfactual evidence runs counter to the assertion of p. It is supplied by a variety of
markers, such as a temporal when-clause, a degree modifier (‘so casual’), deictic items (‘like
that’), and a spatial PP (‘in front of’ …). These markers all signal a violation of the speaker’s
expectations and are conducive to the indirect assertion of non (modality) p, although p is
presupposed
20
. This discrepancy accounts for the mirative meaning of these rhetorical questions.
3.5. Addressee commitment
In surprise contexts, rhetorical questions may be considered mirative utterances
21
not only
because they express surprise, but also because the discrepancy they convey triggers a specific
stance on the part of the speaker. Modalised rhetorical questions deny reality either by cancelling
the question-answer presupposition
22
or by relying on counterfactual evidence
23
. The speaker
directly perceives some event, which should lead to the assertion of p, the status of p being in no
doubt. And yet, the speaker does not commit to the truth of p, because p runs counter to her
expectations. The function of rhetorical questions is then to question the grounds that made p
possible. In a rhetorical question, the speaker selects an answer and requests the addressee to
commit to the truth of that proposition
24
. Rhetorical questions are biased because they do not
20
This notation is borrowed from Desmets and Gautier (2009). They argue that comment-rhetorical
questions in French conflate a question about the modal operator (pouvoir) of the proposition and an
assertion that negates both the modal operator and the proposition. How-rhetorical questions in English
with the modal auxiliary ‘can’ behave in a similar way.
21
Mirativity is defined by several authors (among others, Guentchéva (2017), Peterson (2017)) as
resulting from a discrepancy between what is observed and what is expected.
22
On cancellation of question-answer presupposition, see Huddleston and Pullum (2002: 900-901).
23
My translation of “indice contrefactuel”, a concept borrowed from Desmets and Gautier (2009: 111).
24
I agree with Desmets & Gautier (2009) and Beyssade & Marandin (2009) that the speaker requests the
addressee’s commitment to a proposition in rhetorical questions. However, rhetorical questions do not
necessarily request a verbal response, let alone an answer. They seek the addressee’s alignment, but it is
not clear whether the addressee will actually commit to the proposition suggested by the speaker. If
rhetorical questions are not generated by surprise, anger or disagreement, the assertion they imply is
uncontroversial and the answer to the question is obvious. The absence of an answer from the addressee
may then be interpreted as tacit agreement. In emotion-induced rhetorical questions, however, a
commitment update on the part of the addressee is requested. In half of the examples of my sample,
leave any choice to the addressee with respect to the selection of a variable. However, they are
questions in the sense that they request the addressee’s commitment to a proposition. They can
be preceded by the discourse marker tell me, which, as shown by Reese (2007: 51) co-occurs
with questions that request a response, but not with assertions. Asked to commit to a proposition
that stands in contrast to the state of affairs or to her beliefs, the addressee may not respond:
(15) Ed: This is my way of telling you –
Dolores: [furious] What, by putting it in a fuckin' script, for everyone to see?!
What kind of sick mind would operate like that?
[Ed is terribly hurt. Dolores shakes that script.]
Dolores: And what about this so-called "Barbara" character? It's obviously ME!
I'm so embarrassed! This is our life!
The rhetorical question allows the speaker to express disapproval without committing to the
proposition ‘You are a sick mind’. The addressee is asked to commit to the proposition ‘No
sound mind would operate like that’. This indirect insult may reach its goal emotionally, as
specified in the stage direction (‘Ed is terribly hurt’). From an interactional perspective, however,
it is a dead end.
25
Unless an unlikely response such as ‘You are right’ or ‘I know’ is uttered, no
commitment update is possible on the part of the addressee.
In (8) and (12) a response is provided by the addressee, but it does not update the
common ground in the way expected by the speaker: what is presented as counterfactual
evidence according to the speaker’s standards is said to be normal behavior in the addressee’s
response. As a result, the contradiction conveyed by the rhetorical question is cancelled and the
speaker is forced to accept p, even if the rhetorical question requests the addressee to commit to
non (modality) p. In (13), the addressee’s response forces the speaker to revise his appraisal of
the situation.
rhetorical questions are not followed by a response or an answer. In the absence of any explicit verbal
reaction from the addressee, it is difficult to determine whether the speaker and addressee’s common
ground is eventually updated.
25
Although surprise-induced rhetorical questions may take different forms and do not as such constitute a
construction, a parallel may be drawn here with the Split Interrogative construction analysed by Michaelis
& Feng (2015). The conversational dead end is typical of what Michaelis & Feng (2015: 149) call
sarcastic syntax. Under their analysis, ironic utterances are “counterfeit speech acts” that “do not advance
the conversation” because their function is “disruptive”.
The analysis of emotion-induced rhetorical questions shows that the response to a
rhetorical question is not obvious to both speaker and addressee when they have different
expectations. Contra Rohde (2006: 149-150), I argue that the addressee may be committed to a
proposition that contradicts the speaker’s bias. The common ground may then be updated, but in
a way that is not anticipated by the speaker, especially when the speaker denies an unexpected
but actual event that s/he has failed to cognitively integrate.
The mirative nature of emotion-induced rhetorical questions has important theoretical
implications. As argued by Alcázar (2017: 37), mirative rhetorical questions express “antithesis
of the Common Ground”, which goes against standard treatments of rhetorical questions. In line
with Alcázar, I believe that the deictic essence of the rhetorical questions under study accounts
for their mirative meaning
26
. Evidence of this deictic component may be supplied by their
resistance to the embeddability test. Mirative meaning is lost in an embedded clause (see Rett
and Murray 2013, Celle et al. 2017):
(15’) Dolores asked Ed what kind of sick mind would operate like that.
This embedded sentence cannot express Dolores’s surprise, contrary to the rhetorical question in
(15). Even if this sentence is turned into an exclamation, it can only express the speaker’s
surprise, not Dolores’s:
(15’’) Dolores asked Ed what kind of sick mind would operate like that!
However, it should be stressed again that in English, mirativity is not marked as such
morphosyntactically
27
. Predictably, the form and structure of mirative rhetorical questions do not
differ from those of non-mirative rhetorical questions. Combining a semasiological approach
with an onomasiological perspective offers a means to detect a meaning that might otherwise go
26
I view mirativity as an epistemic stance adopted in reaction to an unexpected event. Mirativity consists
in the expression of surprise, not in the description or assertion of surprise (Celle & al. 2017). Along this
line of reasoning, expressions like ‘I am surprised, ‘he was surprised’ are not mirative utterances. These
are assertions of surprise (see Rett & Murray 2013: 455). They need not be anchored to the time of
utterance or to the first and second persons. Such expressions can be embedded without any change in the
surprise meaning.
27
Nonetheless, there is typological evidence in support of mirative rhetorical questions. In Basque, for
example, mirative rhetorical questions are marked by ote, a mirative conjunction (see Alcázar 2017). In
Ashéninka Perené, Mihas (2014: 213-216) also shows that the enclitic =ma~=taima that commonly encodes
inference can occur in content questions to express mirative meaning, both in direct and indirect speech
acts (including rhetorical questions).
unnoticed for lack of a dedicated morpheme. It also enables us to enrich our understanding of
mirativity in English. DeLancey (2001: 377-378) suggests that mirativity is a “covert semantic
category” in English mainly expressed intonationally
28
. I further argue that rhetorical questions,
i.e. questions used as indirect speech acts, may serve a mirative function. One might wonder why
there is such an affinity between mirativity and rhetorical questions. From a schema-theoretic
perspective, surprise induces cognitive processing that starts with the search for a cause and
ultimately ends with belief revision (see Meyer et al. 1997: 253; Miceli & Castelfranchi 2015:
52). This highly adaptive psychological pattern is ideally mapped on rhetorical questions, which
call upon the addresse for commitment update to validate belief revision. Note that mirativity
and rhetorical questions have non-commitment in common. The epistemic stance adopted by a
speaker in reaction to unexpected information is typically one of non-commitment, as shown by
Zeisler (2017) on the use of ḥdug in Ladakhi
29
. Zeisler stresses that “SPEAKER ATTITUDE (or
STANCE) primarily deals with the relation between the speaker and the content of the utterance
and between the speaker and the addressee”. Rhetorical questions are also primarily addressee-
oriented as they request the addressee’s commitment, the speaker putting forward an answer
without committing to the truth of the proposition.
This leads us to refine the epistemically-based typology of questions proposed in Littell
et al. (2010). Questions are not only based on the speaker’s knowledge and beliefs.
30
They also
contribute to dialogue in a dynamic way by requesting the addressee’s commitment. Therefore,
an epistemically-based typology of questions should accommodate the request for commitment
update that distinguishes rhetorical questions from assertions. In a rhetorical question, the
speaker believes that the addressee knows the answer because the answer is suggested by the
question itself, although speaker and addressee may differ in their beliefs and appraisals
31
. In the
28
Mirativity in English is often associated with an exclamation intonation (Rett & Murray 2013).
Questions used as indirect speech acts do have an exclamation intonation and can convey mirative
meaning. As explained in Celle et al. (forthc.), surprise in questions used as direct speech acts is generally
induced by a discourse entity rather than by new environmental information. Although such questions do
express surprise, they should not be considered mirative.
29
Ḥdug is an auxiliary used to encode visual perception and non-commitment. Zeisler (2017) argues that
ḥdug has parasitic mirative connotations.
30
In addition, speaker’s knowledge is grounded in dialogue and articulated to indexical cues, as shown by
Du Bois (2007: 157).
31
By contrast, in an ordinary question, the addressee is asked to commit to her own answer.
case of mirative rhetorical questions, the commitment update cannot be taken for granted as it
may be hindered by disagreement. These features are summed up below in Table 2.
[Insert Table 2 here]
Table 2. An epistemically- and dialogically-based typology of questions
Speaker knows
the answer
Speaker believes that the
addressee knows the answer
Speaker requests
addressee’s commitment
Ordinary
Questions
No
Yes
Yes
Rhetorical
Questions
Yes
Yes
Yes
Unresolvable
Questions
No
No
No
5. Conclusion
This chapter is part of a study of all question types used in reaction to surprising information. As
shown in Celle & al. (forthc.) 83% of the interrogatives found in our corpus of movie scripts are
questions used as direct speech acts. These questions request information that aims to increase
the speaker’s knowledge and are generally discourse-linked, that is, surprise is generated by
unexpected new linguistic information.
This chapter focused on the remaining 17% surprise-induced questions used as indirect
speech acts. It showed that in English, mirativity can be conveyed by two different types of
interrogatives used as indirect speech acts. Both surprise-induced rhetorical questions and
unresolvable questions take the form of interrogatives but do not request information. Rhetorical
questions are resolved questions, whereas unresolvable questions implicate that no resolution can
be reached. Their mirative meaning is conveyed by a discrepancy between what is observed and
what was expected. However, being generated by different types of surprising situations, these
questions exhibit mirative meaning at different stages of the cognitive assimilation of unexpected
new information.
Surprise-induced rhetorical questions are typically generated by counterfactual evidence.
By using such an indirect speech act, the speaker distances him/herself from the discrepant actual
state of affairs. Rhetorical content questions are found in reaction to some behavior for which the
addressee is held responsible. Their aim is to persuade the addressee to modify that behavior in
order to meet the speaker’s expectations. From a schema-theoretic perspective, rhetorical
questions may be viewed as the last stage of the surprise-induced appraisal process (see Meyer et
al. 1997: 253 cited above): they aim at a commitment update on the part of the addressee. Using
concepts borrowed from the schema-theoretic framework not only allows bridging the gap
between psychology and linguistics. It also sheds light on the shades of mirativity by correlating
them with different stages of the cognitive processing of new information induced by surprise.
Rhetorical questions appear to express mirativity at the semantic-pragmatic level as mirative
meaning is not associated with a specific morphosyntactic form in English. However, a
constellation of grammatical and lexical items (modal auxiliaries, deictics, degree words) is the
hallmark of mirativity.
Surprise-induced unresolvable questions tend to be triggered by directly perceived
evidence. The addressee’s agency is not involved – at least in my data – and a judgment of
incongruity is formed once the situation has been appraised either positively or negatively.
Whatever the answer – if any – it cannot account for the sense of incongruity attached to the
situation. This type of expressive speech act does not carry an epistemic goal. Rather, it carries
an instruction that no variable can be provided to instantiate the salient open proposition.
Mirativity projects from the initial stage of the cognitive assimilation of unexpectedness: “the
appraisal of a cognized event as exceeding some threshold value of schema-discrepancy (or
unexpectedness)” (Meyer et al. 1997: 253). At that stage, the surprise process produces a state of
ignorance and wonder. It is expressed by specific lexemes, that is, interjections and emotive
modifiers that provide the indirect speech act with an expressive force.
Pragmatically, rhetorical questions express a biased position and are generally said to
point to an obvious answer. The present chapter offers an alternative analysis that accounts for
the apparent paradox of rhetorical questions being used in reactions of surprise. Contra Rohde
(2006), I argue that informativity and update of the participants’ commitment sets should be
distinguished. In surprise-generated rhetorical questions, although no informative answer is
requested, a two-fold update is expected. First, rhetorical questions signal the speaker’s attempt
to cognitively assimilate new environmental information, the actual state of affairs being
counterexpectational. Second, they request a commitment update on the part of the addressee in a
questioning process triggered by the speaker’s and the addressee’s conflicting views. Rhetorical
questions show that surprise, in association with other emotions it contributes to generating (such
as anger), can be exploited within complex argumentative strategies, as evidenced in other
research works on the lexicon of surprise (see Tutin 2017; Celle et al. 2017). Emotion-induced
rhetorical questions serve an argumentative function whereby the addressee is asked to commit
to a proposition that the speaker does not commit to in a direct way. Rhetorical questions offer a
pragmatic means to attempt to reduce the belief discrepancy associated with the experience of
surprise. However, if the belief discrepancy cannot be reduced - in case of strong disagreement -
they take on a challenging function.
Unlike rhetorical questions, unresolvable questions are generated by evidence judged
incongruous. They are speaker-oriented, the speaker attempting to emotionally adapt to an
incongruous situation without expecting an answer or even a response of the addressee. In
English, interjections and emotive modifiers encode the mirative meaning of unresolvable
questions. Further investigations are needed to better assess their respective contributions to
mirativity. The claim made in this chapter is that in English, expressives can change the
illocutionary force of a sentence in the same way as evidentials in other languages. This can be
explained on the grounds that evidentials and expressives share common features. Emotive
modifiers and interjections are illocutionary modifiers triggered by direct evidence. They encode
the speaker’s emotional experience, while evidentials encode the speaker’s perceptual or
cognitive experience. In sum, both evidentials and expressives encode speaker perspective
32
.
Further investigations are needed to better assess the respective contributions of expressives and
evidentials to interrogatives.
This chapter offers a refinement of Littell et al.’s (2010) typology of questions by
including the commitment update parameter. It also proposes to distinguish between conjectural
questions and unresolvable questions. In conjectural questions, the speaker knows the answer
32
See San Roque et al. (2017). Potts (2007: 173) argues that expressives may be embedded and involve
“perspective dependence” rather than strictly speaker perspective.
and only seeks the addressee’s commitment to the truth of the proposition. In unresolvable
questions, emotive modifiers change the illocutionary force by implicating that neither speaker
nor addressee can provide an answer.
These findings also suggest that unresolvable questions are generated by outcome-related
surprise, while rhetorical questions are generated by person-related surprise
33
. I leave it to future
research to determine whether this distinction is reflected in the appraisal pattern and whether it
generates differences in linguistic responses. The complex relation between evidence-induced
mirative utterances and discourse-based topic-comment constructions is also an avenue for future
research.
References
Akatsuka, Noriko. 1985. Conditionals and the epistemic scale. Language 61(3): 625-639.
Alcázar, Asier. 2017. A syntactic analysis of rhetorical questions. In Proceedings of the 34th
West Coast Conference on Formal Linguistics, Aaron Kaplan, Abby Kaplan, Miranda K.
McCarvel & Edward J. Rubin (eds), 32-41. Somerville, MA: Cascadilla Proceedings
Project.
Ameka, Felix. 1992. Interjections: the universal yet neglected part of speech. Journal of
Pragmatics 18(2) : 101-118.
Azzopardi, Sophie & Bres, Jacques. 2014. Futur, conditionnel, et effets de sens de conjecture et
de rejet en interrogation partielle. SHS Web of Conferences 8 : 3003-3013.
Beyssade, Claire, & Marandin, Jean-Marie. 2009. Commitment: une attitude dialogique. Langue
française 162: 89-107.
Bourova, Viara & Dendale, Patrick. 2013. Serait-ce un conditionnel de conjecture? Datation,
33
Van Dijk and Zeelenberg (2002) distinguish between outcome-related vs. person-related
disappointment. Ortony et al. (1988) make a similar distinction between event-based and agent-based
emotions. This distinction can be fruitfully applied to surprise and helps account for the connection
between surprise and anger observed in the case of rhetorical questions, anger being a person-related
emotion.
évolution et mise en relation des deux conditionnels à valeur évidentielle. Cahiers
Chronos 26 : 183-200.
Caponigro, Ivano & Sprouse, Jon. 2007. Rhetorical questions as questions. In Proceedings of
Sinn und Bedeutung 11, E. Puig-Waldmüller (ed.), 121-133. Barcelona: Universitat
Pompeu Fabra.
Carroll, James M. & Russell, James A. 1997. Facial expressions in Hollywood’s portrayal of
emotion. Journal of Personality and Social Psychology 72 : 164-176.
Celle, Agnès. 2007. Analyse unifiée du conditionnel de non prise en charge en français et
comparaison avec l’anglais. Cahiers Chronos 19, Etudes sémantiques et pragmatiques
sur le temps, l’aspect et la modalité : 43-61.
Celle, Agnès. 2009. Question, mise en question: la traduction de l’interrogation dans le discours
théorique. Revue Française de Linguistique Appliquée XIV/1: 39-52.
Celle, Agnès. in press. Epistemic evaluation in factual contexts in English. In Epistemic Modality
and Evidentiality in a Cross-Linguistic Perspective, Zlatka Guentchéva (ed.). Berlin:
Mouton de Gruyter.
Celle, Agnès & Lansari, Laure. 2014. Uncertainty as a result of unexpectedness. Language and
Dialogue 4(1) 7-23.
Celle, Agnès & Lansari, Laure. 2015. On the mirative meaning of aller + infinitive compared
with its equivalents in English. Cahiers Chronos 27, Taming the TAME systems: 289-
305.
Celle, Agnès, & Lansari, Laure. 2016. L’inattendu et le questionnement dans l’interaction
verbale en anglais: les questions en why-would et leurs réponses. In Représentations du
sens linguistique: les interfaces de la complexité, Olga Galatanu, Ana-Maria Cozma &
Abdelhadi Bellachhab (eds.), 235-248. Bruxelles: Peter Lang.
Celle, Agnès, Jugnet, Anne, Lansari, Laure & L’Hôte, Emilie. 2017. Expressing and describing
surprise. In Expressing and Describing Surprise, Agnès Celle & Laure Lansari (eds),
215-244. Amsterdam: John Benjamins.
Celle, Agnès, Jugnet, Anne, Lansari, Laure & Peterson, Tyler (forthc.) Interrogatives in surprise
contexts in English. In Surprise at the Intersection of Phenomenology and Linguistics,
Natalie Depraz & Agnès Celle (eds.). Amsterdam: John Benjamins.
Darwin, Charles. 1965. The Expression of the Emotions in Man and Animals. Chicago:
University of Chicago Press.
DeLancey, Scott. 2001. The mirative and evidentiality. Journal of Pragmatics 33: 369-382.
Dendale, Patrick. 2010. Il serait à Paris en ce moment. Serait-il à Paris? A propos de deux
emplois épistémiques du conditionnel: grammaire, syntaxe, sémantique. In Liens
linguistiques. Études sur la combinatoire et la hiérarchie des composants, Castro
Camino Alvarez, Flor Marίa Bango de la Campa & Marίa Luisa Donaire (eds), 291-317.
Bern: Peter Lang.
Den Dikken, Marcel & Giannakidou, Anastasia. 2002. From hell to polarity: “aggressively non-
D-linked” wh-phrases as polarity items. Linguistic Inquiry 33(1): 31-61.
Desmets, Marianne & Gautier, Antoine. 2009. “Comment n’y ai-je pas songé plus tôt ?”
Questions rhétoriques en comment. Travaux de linguistique 58 : 107-125.
Diller, Anne-Marie. 1977. Le conditionnel, marqueur de dérivation illocutoire. Semantikos 2: 1-
17.
Du Bois, John. 2007. The stance triangle. In Stancetaking in Discourse, Robert Englebretson
(ed.), 139-182. Amsterdam: John Benjamins.
Fillmore, Charles. 1985. Syntactic intrusions and the notion of grammatical construction.
Proceeding of the Eleventh Annual Meeting of the Berkeley Linguistics Society: 73-86.
Ginzburg, Jonathan. 2012. The Interactive Stance: Meaning for conversation. Oxford: Oxford
University Press.
Guentchéva, Zlatka. 2017. An enunciative account of admirativity in Bulgarian. Review of
Cognitive Linguistics, The linguistic expression of mirativity, Special Issue 15(2): 539-
574.
Gunlogson, Christine A. 2001. True to Form. Rising and Falling Declarative as Questions in
English. PhD Dissertation, University of California, Santa.
Huddleston, Rodney & Pullum, Geoffrey. 2002. The Cambridge Grammar of the English
Language. Cambridge: Cambridge University Press.
Haillet, Pierre-Patrick. 2001. A propos de l’interrogation totale directe au conditionnel. In Le
conditionnel en français, Recherches linguistiques 25, Université de Metz, Patrick Dendale &
Liliane Tasmowski (eds.), 295-330.
Hoeksema, Jack & Napoli, Donna Jo. 2008. Just for the hell of it: a comparison of two taboo-
term constructions. Journal of Linguistics 44: 347-378.
Kay, Paul, & Fillmore, Charles. 1999. Constructions and linguistic generalizations: the what’s X
doing Y? construction. Language 75(1): 1-33.
Korotkova, Natalia. 2016. Heterogeneity and Uniformity in the Evidential Domain. PhD
Dissertation, University of California, Los Angeles.
Littell, Patrick, Matthewson, Lisa & Peterson, Tyler. 2010. On the semantics of conjectural
questions. In Evidence from Evidentials, Tyler Peterson & Uli Sauerland (eds), 89-104.
University of British Columbia Working Papers in Linguistics (UBCWPL).
Meyer, Wulf Uwe, Reisenzein, Rainer & Schützwohl, Achim. 1997. Toward a process analysis
of emotions: the Case of Surprise. Motivation and Emotion 21(3): 251-274.
Miceli, Maria & Castelfranchi, Cristiano. 2015. Expectancy and Emotion. Oxford: Oxford
University Press.
Michaelis, Laura & Feng, Hanbing. 2015. What is this, sarcastic syntax? Constructions and
Frames 7(2): 148-180.
Michaelis, Laura & Knud, Lambrecht. 1996. The exclamative sentence type in English. In
Conceptual Structure, Discourse and Structure, Adele Goldberg (ed.), 375-389. Stanford:
CSLI.
Mihas, Elena. 2014. Expression of information source meanings in Ashéninka Perené (Arawak).
In The Grammar of Knowledge, A Cross-Linguistic Typology, Alexandra Aikhenvald,
Robert Dixon & Malcolm Ward (eds), 209-226 Oxford: Oxford University Press.
Ortony, Andrew, Clore, Gerald L. & Collins, Allan. 1988. The Cognitive Structure of Emotions.
Cambridge: Cambridge University Press.
Pesetsky, David. 1987. Wh-in-Situ: movement and unselective binding. In The Representation of
(in)Definiteness, Eric Reuland & Alice ter Meulen (eds), 98-129. Cambridge, MA: MIT
Press.
Peterson, Tyler. 2017. Grammatical evidentiality and the unprepared mind. In Expressing and
Describing Surprise, Celle Agnès & Laure Lansari (eds), 51-89. Amsterdam: John
Benjamins.
Pinker, Steven. 2007. The Stuff of Thought: Language as a window into human nature. New
York: Penguin.
Potts, Christopher. 2005. The Logic of Conventional Implicatures. Oxford: Oxford University
Press.
Potts, Christopher. 2007. The expressive dimension. Theoretical Linguistics 33(2): 165-198.
Reese, Brian. 2007. Bias in Questions. PhD Dissertation, The University of Texas at Austin.
Reisenzein, Rainer. 2000. Exploring the strength of association between the components of
emotion syndromes: the case of surprise. Cognition and Emotion 14(1): 1-38.
Rett, Jessica & Murray, Sarah. 2013. A semantic account of mirative evidentials. In Proceedings
of SALT 23: 453-472.
Rocci, Andrea. 2007. Epistemic modality and questions in dialogue. The case of Italian
interrogative constructions in the subjunctive mood. In Tense, Mood and Aspect:
Theoretical and descriptive issues, Louis de Saussure, Jacques Moeschler & Genoveva
Puskas (eds), 129-153. Amsterdam: Rodopi.
Rohde, Hannah. 2006. Rhetorical questions as redundant interrogatives. San Diego Linguistics
Papers, Vol. 2, University of California, San Diego: 134-168.
San Roque, Lila, Floyd, Simeon & Norcliffe, Elisabeth. 2017. Evidentiality and interrogativity.
Lingua 186-187: 120-143.
Scherer, Klaus, Clark-Polner, Elizabeth, Mortillaro, Marcello. 2011. In the eye of the beholder?
Universality and cultural specificity in the expression and perception of emotion.
International Journal of Psychology 46(6): 401-435.
Siemund, Peter. 2017. Interrogative clauses in English and the social economics of questions.
Journal of Pragmatics 199: 15-32.
Stein, Nancy L. & Hernandez, Marc W. 2007. Assessing understanding and appraisals during
emotional experience. In Handbook of Emotion Elicitation and Assessment, James A.
Coan & John J. B. Allen (eds), 298-317. Oxford: Oxford University Press.
Stivers, Tanya. 2010. An overview of the question-response system in American English
conversation. Journal of Pragmatics 42(10): 2620-2626.
Tutin, Agnès. 2017. Surprise routines in scientific writing: a study of French social science
articles. In Expressing and Describing Surprise, Agnès Celle & Laure Lansari (eds), 153-
172. Amsterdam: John Benjamins.
Van Dijk, Wilco W. & Zeelenberg, Marcel. 2002. What do we talk about when we talk about
disappointment? Distinguishing outcome-related disappointment from person-related
disappointment. Cognition and Emotion 16(6): 787-807.
Vingerhoets, Ad, Bylsma, Lauren & De Vlam, Cornelis. 2013. Swearing: a biopsychosocial
perspective. Psychological Topics 22 (2013), 2, 287-304.
Zaefferer, Dietmar. 2001. Deconstructing a classical classification: a typological look at Searle’s
concept of illocution type. Revue internationale de philosophie 2001/2 (n° 216), 209-225.
Zeisler, Bettina. 2017. Don’t believe in a paradigm that you haven’t manipulated yourself ! -
Evidentiality, speaker attitude, and admirativity in Ladakhi. Review of Cognitive
Linguistics, Special Issue 15(2): 515-538.
