Ellipsis in inquisitive semantics

Scott AnderBois
Scott [email protected]

September 13, 2016

1 Introduction

Recent years have seen a proliferation of approaches to ellipsis which make crucial ref-

erence to the semantic interpretations of ellipsis sites and their antecedents in various

ways. At the same time, recent decades have witnessed a sea change within the field of

semantics, with many researchers treating sentence meanings not in terms of mere truth-

conditions, but rather in terms of a broader notion of Context Change Potential (CCP)

or Information Exchange Potential. It seems natural, then, to ask the question of how

this broader notion of semantic content can be brought to bear in the analysis of ellipsis.

In this chapter, we engage this question by focusing on one particular branch of se-

mantic theories with this broader conception of meaning: inquisitive semantics (Groe-

nendijk (2007), Groenendijk & Roelofsen (2009), AnderBois (2012a) inter alia, see Cia-

rdelli et al. (2013) for a recent overview). Inquisitive semantics holds that sentence mean-

ings for both declarative and interrogative sentences consist of (or at least determine) sets

of alternative propositions. For the study of ellipsis, then, the hypothesis is that seman-

tic conditions on certain ellipsis processes will (or at least may) make reference to this

broader, alternative-rich notion of semantic content rather than to mere truth conditions.

The outline for this chapter is as follows: §2 introduces inquisitive semantics; §3 briefly

presents the most fully fleshed out account of an ellipsis process using inquisitive seman-

tics: AnderBois (2014); §4 concludes by addressing the questions of Structure, Recov-

erability, and Licensing from the perspective of the account in §3 and discusses how

1
 Ellipsis in inquisitive semantics

they might di↵er under other possible ways of incorporating inquisitive semantics into a

theory of ellipsis.

2 Inquisitive semantics

2.1 What makes a semantics ‘inquisitive’ ?

The core intuition behind inquisitive semantics is that the meaning/CCP of sentences not

only includes truth-conditional information, but also includes the issue(s) that it raises,

i.e. its inquisitive content. This has been long recognized, of course, for the CCPs of in-

terrogative sentences. Inquisitive semantics extends this idea to capture the intuition that

assertions, especially those containing widest scope disjunctions and indefinites, also raise

issues in discourse. For example, (1) introduces two alternatives – ‘that it will rain’, ‘that

it will snow’ – and thereby makes this issue salient in the output context in a way that

truth-conditionally equivalent sentences such as ‘it will precipitate’ do not (assuming that

rain and snow are the only forms of precipitation).

(1) (Either) it will rain or it will snow.

Inquisitive semantics therefore builds on a number of recent works on disjunction

and indefinites (e.g. Kratzer & Shimoyama (2002), Simons (2005), Alonso-Ovalle (2006))

in what has been called ‘Hamblin’ or ‘Alternative’ semantics. These works hold that in-

definites and disjunctions introduce alternatives into semantic composition. While the

name ‘inquisitive semantics’ refers to a family of related semantic/pragmatic theories

(see e.g. Groenendijk (2011) for discussion), there are two fairly consistent ways in which

inquisitive semantics di↵ers from Hamblin semantics. First, the two di↵er in the ways

in which alternative-rich meanings are composed and what their formal properties are.

These properties have no clear importance for the study of ellipsis, so we refer the reader

to Ciardelli et al. (2013) and references therein for detailed discussion. Second, alternative-

evoking (or lack thereof) is treated as an aspect of the top level meaning of assertions and

2
 Ellipsis in inquisitive semantics

questions alike and therefore of their contribution to discourse as well.1

2.2 Inquisitive semantics across sentence types

We turn now to give a concrete version of inquisitive semantics for both questions and

assertions. Given our present purposes, the presentation here is necessarily informal (see

Groenendijk & Roelofsen (2009) and AnderBois (2012a) for more detailed formal presen-

tations).

The core formal shift in inquisitive semantics is to treat sentence meanings not as

sets of possible worlds (i.e. propositions), as is done classically, but rather as sets of sets

of possible worlds (i.e. sets of propositions). This move itself has a precedent in Hamblin

(1973)’s semantics for interrogatives, but di↵ers in that expressions other than questions

will make use of these richer meanings. In particular, we assume that disjunctions, indefi-

nites, and other forms of existential quantification also contribute alternatives into seman-

tic composition. Following Groenendijk & Roelofsen (2009), we will call a sentence inquis-

itive if its interpretation contains more than one alternative. The idea, then, is that not

only interrogatives, but also declarative sentences may be inquisitive in this sense. Fur-

thermore, if covert existential quantification is also taken to be inquisitive (as we suggest

in §3.3), the inquisitivity of declaratives will be quite regular.

Taking (1) as an example, we illustrate this idea in informal set notation in (2a) and

graphically in the diagram (2b). Diagrams like (2b) provide a pictorial representation of

the interpretation of the formula in a toy model. We assume a toy model containing four

possible worlds (w00 , w01 , w10 , w11 ), represented visually by the four named circles. The

names here correspond to the truth values of two propositions, p and q. In our current

example, then, p = ‘that it will rain’ and q = ‘that it will snow’. The boxes, then, repre-

sent alternatives in the interpretation of the sentence.

3
 Ellipsis in inquisitive semantics

(2) Classical disjunction: Jp _ qKcl Inquisitive disjunction: Jp _ qKinq

a. { ‘that it will precipitate’ } { ‘that it will rain’, ‘that it will snow’ }

b.

11 10 Precipitation 11 10 Rain

01 00 Snow! 01 00

The classical semantics for disjunction is not inquisitive since it produces only a sin-

gle ‘alternative’, the proposition that it will precipitate. In contrast, an inquisitive seman-

tics for disjunction produces the alternative-rich interpretation with distinct alternatives

for the di↵erent forms of precipitation mentioned in the sentence. While these two inter-

pretations di↵er in their inquisitive content, they contain the same informative content,

i.e. determine the same truth-conditions. It is the same set of possible worlds which ap-

pear in some alternative or other in the interpretation of the two formulas, just struc-

tured di↵erently.

Inquisitive semantics therefore allows us to distinguish truth-conditionally equivalent

formulas on the basis of the alternatives they evoke. To continue with our toy example,

then, we assign a sentence with a disjunction like (1) the alternative-rich interpretation

on the right of (2), while assigning (3) the di↵erent (yet truth-conditionally equivalent)

semantic representation on the left of (2).

(3) It will precipitate.

Extending such a semantics to indefinites and other forms of existential quantifica-

tion is fairly straightforward at an intuitive level.2 Whereas a disjunction specifies alter-

natives one by one, an indefinite produces a set containing one alternative per individual

in the restrictor set. For example, a sentence like (4) will receive an interpretation with n

alternatives, one per each of the n individuals in the interpretation of ‘student’.3

4
 Ellipsis in inquisitive semantics

(4) Prof. Ramı́rez met with a student.

(5) { ‘that R met with Al’, ‘that R met with Bella’, ‘that R met with Chad’, . . . }

In this section, we have introduced the basic conception of sentence meaning in in-

quisitive semantics with a focus on the two main alternative-evoking elements: disjunc-

tion and existential quantifiers such as indefinites. Simple sentences containing these el-

ements make salient a set of alternatives and simultaneously contribute the information

that the world of evaluation lies within some alternative(s) in this set. In §2.4, we provide

a semantics for questions which makes the parallel with indefinites and disjunction ex-

plicit, while also capturing the di↵erence between the two classes. Before doing so, how-

ever, we introduce a class of operators which interact with the inquisitive component of

formulas to which they apply.

2.3 Negation and other operators

Thus far, we have given an informal introduction to an inquisitive semantics for disjunc-

tion and existential quantification. This semantics holds that sentences containing these

elements make salient in the discourse a set of alternatives in a way that truth-conditionally

equivalent sentences may not. While this intuition seems fairly clear for the simple sen-

tences we have looked at thus far, it turns out only to hold of sentences which contain

wide-scope disjunctions and existential quantifiers. For example, a disjunction within the

scope of negation, as in (6a), seems to be no more inquisitive than its non-disjunctive

counterpart, (6b).

(6) a. It’s not the case that it rained or snowed.

b. It’s not the case that it precipitated.

Moreover, this behavior in fact follows from the way negation is naturally defined

in inquisitive semantics (see Roelofsen (2013) and references therein for detailed discus-

sion of its mathematical foundations). Since sentence meanings are sets of alternatives,

5
 Ellipsis in inquisitive semantics

negation rejects each of these alternatives as in ((7), middle), returning the maximal al-

ternative which does not overlap with any of these. One important consequence of this is

that, while it preserves truth conditions, double negation is no longer semantically vac-

uous since it eliminates the alternative-rich structure of the formula to which it applies,

((7), right).

(7) J9x.'(x)K J¬9x.'(x)K J¬¬9x.'(x)K

11 10 11 10 11 10

01 00 01 00 01 00

It follows, then, that not just any sentence containing an inquisitive element will be

inquisitive, but rather those where the inquisitive element takes widest scope. We have

illustrated this here for negation, other operators may also have this property, such as the

Comma operator found in appositive relative clauses. This fact is parallel to the obser-

vation (e.g. Chung et al. (1995), Romero (1998)) that inner antecedents for sluicing must

take wide scope (indeed, licensing of sluicing is often used as a diagnostic for scope of in-

definites).4

2.4 Questions

While inquisitive semantics assigns a more question-like semantics to sentences with wide-

scope disjunctions and existential quantifiers, we still need to distinguish these latter el-

ements from questions. The basic approach in inquisitive semantics – at least in matrix

clauses – is to di↵erentiate the two in terms of their informative potential. For ease of ex-

position, we will simply use the term ‘informativity’ below, though in all cases the rel-

evant notion is possible informativity rather then actual (see AnderBois (2012a) for de-

tailed discussion). The inquisitive semantics for disjunction in the right side of (2) not

6
 Ellipsis in inquisitive semantics

only introduces a set of alternatives, it also includes the truth-conditional information

that some alternative or other holds. That is to say that a declarative with a wide-scope

disjunction rules out the possibility that none of the alternatives hold. Intuitively, and

perhaps definitionally so, questions are not possibly informative in this way.

The literature on inquisitive semantics has seen two ways to cash out this insight,

which we can call absolute and relative uninformativity. We can illustate these two ap-

proaches for the wh-question in (8), recalling that we keep the pictures to two positive

alternatives. The absolute uninformativity approach is exemplified by Groenendijk &

Roelofsen (2009), who propose a Q(uestion) operator which adds in the ‘no one’ alter-

native, as in the left picture of (9). The second option, due to AnderBois (2012a), is to

claim that questions have an existential presupposition and that the alternative set of the

question is uninformative only relative to this presupposition. We can indicate this picto-

rially by shading out the worlds presupposed not to be live options (just world 00 in this

case):

(8) Who told John about the party?

(9) Two di↵erent inquisitive semantics for questions:

11 10 11 10

01 00 01 00

[Absolute uninformativity] [Relative uninformativity]

Here, we will follow AnderBois (2014) in adopting this latter option. Ultimately,

however, the decision between these two approaches is an empirical one, resting largely

on the longstanding question of whether questions contribute existential presuppositions.

Beyond the unresolved nature of this question for English, it is of course possible that

7
 Ellipsis in inquisitive semantics

both options are needed across languages or across question constructions within a single

language (e.g. argument vs. adjunct wh-questions, wh- vs. alternative questions).

3 Sluicing is sensitive to inquisitive content

Having introduced inquisitive semantics, we turn now to apply it to the analysis of ellip-

sis and, in particular, AnderBois (2010), AnderBois (2014)’s account of both merger and

sprouting subtypes of sluicing. Finally, §4 will conclude by exploring other possible ways

of incorporating the core insights of inquisitive semantics into a theory of ellipsis, drawing

on related theories such as dynamic semantics and QUD approaches to discourse.

3.1 The need to move beyond truth conditions

As a theory of the semantic content of questions and assertions, inquisitive semantics is

only of direct relevance to ellipsis to the extent that semantics itself is (or pragmatics

which is sensitive to semantics). In principle, it would be consistent for inquisitive seman-

tics to provide an appropriate theory of semantic content, yet for ellipsis to be resolved

in a purely syntactic (or LF-syntactic) way (as in, e.g. Sag (1976), Chung et al. (1995)).

However, there is a large body of work across many di↵erent frameworks arguing that se-

mantics/pragmatics do play a crucial role in ellipsis phenomena (Sag & Hankamer (1984),

Hardt (1993), Ginzburg & Sag (2001), Merchant (2001), Culicover & Jackendo↵ (2005),

Chung et al. (2011) among many others). Assuming that this is right in some way, inquis-

itive semantics naturally raises the question of whether this condition will be sensitive not

only to truth-conditional information, but also to the inquisitive aspect of semantic con-

tent.

There are several di↵erent kinds of data which have been argued to support the need

for the semantic condition on sluicing to be sensitive to inquisitive content. Except where

noted, we focus on data from English, though there is no reason to expect the observa-

tions we make to not be more general. In this section, we focus on data from the subtype

8
 Ellipsis in inquisitive semantics

of sluicing which Chung et al. (1995) dub ‘merger’, i.e. cases where there is an overt ‘in-

ner antecedent’ in the A(ntecedent) clause corresponding to the wh-phrase in the E(lided)

clause. In §3.3, we turn to cases of Chung et al. (1995)’s ‘sprouting’, i.e. cases where

there is no overt inner antecedent, with the wh-phrase instead corresponding to an im-

plicit argument or adjunct in the A-clause.

Perhaps the most fundamental observation supporting the relevance of inquisitive

semantics for sluicing is the role played by inquisitive elements as inner antecedents. We

see this clearly in the contrasts between the felicitous sluices in (10a), whose antecedents

are inquisitive sentences, and the infelicitious ones in (10b), whose antecedents are non-

inquisitive. Notice that in both cases the corresponding full-clausal versions are felicitous.

(10) a. She said she had spoken to {someone/a student/John or Bill}, but Harry

didn’t know who.

b. #She said she had spoken to {everybody/most students/the student/him/John

and Bill}, but Harry didn’t know who.

In addition to this basic observation, it is well-known that these elements must take

wide scope in order to serve as inner antecedents (e.g. Chung et al. (1995), Romero (1998),

Barker (2013)). As we will see below in a moment, this generalization follows straight-

forwardly from an inquisitive semantic approach to sluicing. In fact, from the inquisi-

tive semantic perspective, the ability of disjunctions and indefinites to serve as inner an-

tecedents is simply another manifestation of the interrogative-indefinite-disjunction affin-

ity (e.g. Haspelmath (1997), Bhat (2000), Haida (2008)).

While the privileged role of disjunctions and indefinites is of course quite suggestive,

this alone leaves somewhat open the possibility that it is some other aspect of these ex-

pressions which is crucial. For example, focusing primarily on indefinites, Chung et al.

(1995) argue that it is the logical form of these elements (i.e. the fact that they contribute

a variable in the Heimian view, Heim (1982)) which is crucial. Merchant (2001) argues

that it is the truth-conditions following existential closure of the A- and E-clauses which

9
 Ellipsis in inquisitive semantics

is relevant. One kind of data which is problematic for the former view at least (as Chung

et al. (1995) note) are disjunctive inner antecedents where the disjunction is not of argu-

ments, as was the case in (10a), but of entire clauses, as in (11).

(11) a. (Either) Freddie is baking a cake again or something is on fire, but I can’t tell

which (one). AnderBois (2014)

b. Russ is in the back or Ali is working alone, but I can’t tell which (one). An-

derBois (2014)

As for the latter view, there are several cases of expressions which are truth-conditionally

equivalent to overt widest-scope indefinites, yet do not license sluicing (i.e. cannot serve

as an inner antecedent). Such cases are unexpected if truth-conditions are all that the se-

mantic condition cares about, but entirely expected from the view of inquisitive semantics

since they plausibly have di↵erent inquisitive content. Perhaps the most straightforward

case where truth-conditions alone prove inadequate are examples like (12) where we see

that indefinites and NPIs with double negation and negative quantifiers with single nega-

tion (in Standard American English) fail to license sluicing. The would-be A-clauses have

counterparts with no negation (e.g. ‘Someone left.’ for (12b)) which are true in the same

circumstances but di↵er in their ability to license sluicing.

(12) a. #It’s not the case that Bill didn’t bring a dish, but I don’t know which (one).

AnderBois (2014)

b. A: It’s not the case that no one left.

B: #Who?

c. #It’s not true that Bill didn’t talk to anyone, and Jane just asked me who.

The second case discussed by AnderBois (2014) are indefinites that occur inside ap-

positive (non-restrictive) relative clauses. In contrast to restrictive relative clauses, the

content of appositive relative clauses is generally thought of as having sentence-level scope

(or perhaps as being in some sense ‘scopeless’), but having no other truth conditional im-

pact. Therefore, a truth-conditional account predicts that inquisitive elements inside of

10
 Ellipsis in inquisitive semantics

appositive relative clauses ought to readily serve as inner antecedents for sluicing. As seen

by the infelicity of (13), based on examples in AnderBois (2014), this prediction is not

borne out.

(13) a. #The valiant knight, who defeated {a masked enemy/someone} in a duel, still

wonders who.

b. #Joe, who once killed {a man/someone} in cold blood, doesn’t even remem-

ber who.

More recently, this generalization has been investigated experimentally by Collins

et al. (t.a.) who argue that such examples can be improved by (1) using a wh-phrase of

the form which + NP in the E-clause rather than a bare wh-word like who, and (2) mak-

ing the issue raised by the E-clause (or a related one) salient in the discourse preceding

the target sentence. Given the space limitations of the present work, we will leave a de-

tailed discussion of these issues to future work.5 However, it is worth noting that both of

these manipulations are ones which plausibly raise the salience of the E-issue in the ambi-

ent discourse and therefore in our view should not be seen as evidence that the indefinite

inside the appositive can in fact serve as the inner antecedent, so much as casting doubt

on the logically prior question of whether a linguistic antecedent is strictly necessary in

the first place (i.e. whether sluicing is an instance of ‘surface’ or ‘deep’ anaphora in the

terms of Hankamer & Sag (1976)), as the authors point out. At the same time, the exis-

tence of contrasts like (10) make clear that inquisitive elements at least can play a priv-

ileged role in licensing sluicing, one which is unexpected for Ginzburg & Sag (2001), as

discussed in §4.

As an anonymous reviewer notes, there is one aspect of Merchant (2001)’s account

that one might expect could help capture such data: the requirement that the antecedent

be salient. While salience is often not fleshed out in much detail, it seems a priori plau-

sible that double-negation and apposition reduce the salience of the antecedent and that

this is the reason they impede sluicing. However, we find that other elements typically

11
 Ellipsis in inquisitive semantics

thought to be sensitive to salience such as pronouns and even VP-ellipsis can readily find

an antecedent in these environments as in (14). Therefore, it seems that the e↵ects we are

seeing cannot be straightforwardly attributed to salience of the sort relevant for anaphoric

processes more generally.

(14) a. It is not true that John didn’t bring an umbrella. It was purple and it stood

in the hallway. Krahmer & Muskens (1995)

b. John, who helps people if they want him to, kisses them even if they don’t

want him to help them. AnderBois et al. (2015)

A third case where truth-conditional equivalence proves insufficient to license sluicing

are certain cases of bare noun incorporation, such as those discussed in detail by Collins

(2013) for Samoan. Collins observes that, unlike other truth-conditionally equivalent

indefinite-like expressions in the language, incorporated bare NPs do not license sluic-

ing (nor do they license pronominal anaphora) as seen in (15). Collins (2013) pursues an

analysis in the closely related framework of dynamic semantics, though clearly this could

potentially be treated as a di↵erence in inquisitivity in the current context.

(15) a. Sā fa‘apāgotā-foamea le ta‘ita‘i

Past arrest-thief the leader
‘The chief thief-arrested’

b. #‘o ai?
who
Intended: ‘Who?’

Thus far, we have seen several cases in which sentences with identical truth condi-

tions to those provided by overt indefinites are unable to license sluicing. We turn now

to one further kind of support for the relevance of ‘issues’ to sluicing: the fine-grained

patterns of variation across di↵erent types of nouns and wh-words investigated by Bar-

ros (2013) (see also Dayal & Schwarzschild (2010), Barker (2013) for related observa-

tions). Barros observes that the felicity of sluicing varies depending on complex6 inter-

actions between the nominal content of the inner antecedent and the properties of the

12
 Ellipsis in inquisitive semantics

wh-remnant. For nouns in the A-clause, Barros claims that the felicity of such sentences

depends on whether the noun is a ‘basic-level’ noun or not (in the sense of Brown (1958),

Cruse (1977) and others). Basic-level nouns are nouns that have a privileged status tied

to their encoding a ‘neutral’ level of specificity (i.e. presumably for non-linguistic rea-

sons). For example, out of the blue it would sound more natural to talk about my ‘cat’

than my ‘mammal’ even though clearly both are equally truthful descriptors for a cat.7

(16) a. #Jack has a profiterole, but Fred doesn’t know what (exactly).

b. #Jack ordered an éclair, but Fred doesn’t know what (exactly).

c. #Jack got a cat, but Fred doesn’t know what (exactly).

(17) a. Jack had a drink, but Sally can’t recall what (exactly).

b. Jack ordered food, but Sally doesn’t know what (exactly).

c. Jack ordered an appetizer, but Sally can’t recall what (exactly).

Beyond variations based on the specificity of the noun itself, Barros argues that the

animate wh-word who di↵ers from the inanimate what in allowing more ‘specific’ nouns in

the inner antecedent:

(18) Sally met with a {scientist/geologist/seismologist}, but I don’t know who.

Glossing over important details, the basic idea Barros pursues is that wh-words lex-

ically specify a particular level of specificity, that is, they specify an issue whose alter-

natives have a particular level of granularity. For sluicing to be felicitous, then, the de-

scriptive content of the noun or other inner antecedent material must not be more specific

than the level specified by the wh-word. Who and what di↵er in the level of specificity

they specify, leading to the asymmetry seen above. Barker (2013) makes similar observa-

tions and proposes the generalization in (19):

(19) The Answer Ban: the antecedent clause must not resolve, or even partially re-

solve, the issue raised by the sluiced interrogative.

13
 Ellipsis in inquisitive semantics

We return to these observations in more detail in a moment once we have presented

AnderBois (2014)’s account. For now, let us remark that these observations fit naturally

in a view where semantic content is alternative-rich.

3.2 An account based on symmetric inquisitive entailment

Inquisitive semantics is a theory of semantic content, and as such can be implemented

within a variety of di↵erent theoretical approaches to ellipsis. In this section, we present

the most worked out inquisitive semantic account of an ellipsis process, AnderBois (2014)’s

account of sluicing (AnderBois (2010), AnderBois (2011) present earlier versions of more

or less the same approach). AnderBois (2014)’s approach builds on the approach of Mer-

chant (2001), but incorporates the inquisitive semantic conception of semantic content.

Given this, we very briefly review the major features of Merchant (2001)’s account.

Merchant (2001)’s theory of ellipsis assumes, along with many other authors, that

pronounced material in the E-clause (i.e. the wh-remnant), arises from a fully articulated

clausal version, as in (20). From this starting point, an additional mechanism specifies the

non-pronunciation of the redundant material in this clause, indicated in strikethrough, an

approach commonly known as PF-deletion.

(20) [John talked to someone]A , but I don’t know [whoi John talked to ti ]E .

For Merchant, this PF-deletion operation is subject to the condition in (21), which

ensures that there is a semantically identical antecedent salient in the surrounding con-

text.8 The existential type-shifting portion of the definition existentially quantifies over

missing arguments in order to be able to apply the definition to ellipsis processes which

operate over parts of clauses (e.g. existentially quantifying over the subject in order to

compute entailment between verb phrases for VP-Ellipsis).

(21) Merchant (2001)’s e-Givenness condition: An IP ↵ can be deleted only if ↵

is e-Given.

14
 Ellipsis in inquisitive semantics

(22) F-closure9 The F-closure of ↵, written F-clo(↵), is the result of replacing F-

marked parts of ↵ with 9-bound variables of the appropriate type.

(23) e-Givenness An expression E counts as e-Given i↵ E has a salient antecedent A

and, modulo existential type-shifting,

a. A entails F-clo(E), and

b. E entails F-clo(A).

AnderBois (2014) departs from this basic setup in two ways, only the latter of which

is directly related to inquisitive semantics. First, following Chung (2006) and many sub-

sequent works, the work adopts a ‘hybrid’ approach which supplements the semantic con-

dition with a lexical identity condition in (24). The numeration is the minimalist term for

the list (technically a multiset) of lexical items that comprise the sentence.

(24) No New Morphemes: Every lexical item in the numeration of the sluice that

ends up (only) in the elided IP must be identical to an item in the numeration of

the antecedent CP.

This lexical identity condition is primarily relevant for handling cases of sprouting

(see §2.3), but is aimed at capturing the infelicity/ungrammaticality of examples like

(25), where the A and E-clauses clearly have the same truth-conditions, yet sluicing is not

possible. While it is not impossible to imagine that certain such cases can be captured se-

mantically (e.g. AnderBois (2010) attempts such an account for (25a)), we set aside this

possibility here. It should be noted, however, that even accounts of ellipsis which do not

posit silent linguistic material (e.g. accounts of ‘fragment answers’ by Ginzburg & Sag

(2001) and Jacobson (2013)) similarly treat such data as arising from a minimal (morpho-

)syntactic condition regarding the syntactic category of the E-remnant.

(25) a. #[The cake was eaten]A but we don’t know [who ate the cake]E .

b. #[John is jealous]A , but I don’t know [who John is jealous of]E

15
 Ellipsis in inquisitive semantics

The second di↵erence from Merchant (2001) is that the symmetric entailment condi-

tion analogous to (21) is formulated with respect to inquisitive semantic contents, rather

than just truth-conditions. Groenendijk & Roelofsen (2009) give the formal definition for

entailment in (26). To unpack this definition a bit, recall first that the interpretations of

any given formula will be sets of alternatives. Given this, to see if ' entails , we check

to see if each alternative in J'K is a subalternative of some alternative or other in J K.

Thinking in terms of the pictorial representations above, then, entailment checks to see if

each box in the interpretation of ' fits inside some box or other in the interpretation of

. Symmetric entailment then, means that the interpretations of two formulas have the

same alternatives.

(26) Entailment: ' entails i↵ 8↵ 2 J'K, ↵ is such that 9 2 J K such that ↵ ✓

One important thing to note about this definition is that it does not take into ac-

count the presuppositions of either formula, just the alternatives in the proposed output

state10 . As we saw in §2.4, the existential presupposition is the only thing distinguishing

the interpretation of an interrogative and a corresponding declarative with a wide-scope

indefinite. This therefore allows for the condition on sluicing to be formulated as follows:

(27) Symmetric Entailment Condition on sluicing: Given a structure CPE ,

C+Q IPE
IPE can be elided only if there is some salient antecedent CPA such that:

a. CPE entails CPA , and

b. CPA entails CPE

This condition essentially adapts Merchant’s e-givenness condition to the inquisitive

setting (we set aside issues related to F-closure, see footnote 9). Given the deep seman-

tic connection between interrogatives and indefinites/disjunctions in inquisitive seman-

tics, however, it can be stated over the entire clause (CP), rather than just the deleted

portion, IP. Since the E-clauses in sluicing are necessarily interrogative, the inquisitive

16
 Ellipsis in inquisitive semantics

entailment condition in (27) derives what AnderBois (2014) calls the ‘inner antecedent

generalization’ in (28).

(28) Inner antecedent generalization: An expression ↵ can serve as an inner an-

tecedent for sluicing only if ↵ makes an inquisitive contribution.

Beyond this, as argued by Barros (2013), Barker’s ‘answer ban’, (19), follows as a

particular case of (27). If the A-clause resolves (or partially resolves) the issue raised by

the E-clause, the E-clause by definition does not entail the A-clause and therefore fails

the condition in (27). Accounting for all of the data of this sort does require a more fine-

grained semantics for wh-words than we will give here, so we again refer the reader to

Barros (2013) for further details.11

With this in place, we now show how the account tackles basic cases of sluicing as

well as the infelicitous cases discussed in §3.1. First, let’s consider a basic case where the

inner antecedent is an overt indefinite ‘someone’. The inquisitive interpretation of the

A-clause someone left will be a set of alternatives of the form ‘x left’, (31, left). The in-

terpretation of the E-clause, who left, consists of the same set of alternatives (31, right).

While the E-clause additionally includes a (non-inquisitive) existential presupposition,

the entailment condition in (27) ignores this, and so, the symmetric entailment condition

is met and (29) is correctly predicted to be felicitous. We can visually verify that (27) is

met by looking at the diagrams in (31) and seeing that the alternatives on the two sides

are the same. Disjunction behaves the same, di↵ering only in the that the specific na-

ture of the antecedent obliges the use of the D-linked which in the E-clause (see Dayal &

Schwarzschild (2010), Barros (2013) for discussion of the conditions on which).

(29) [Someone left]A , but I don’t know [who left]E

(30) a. (29)A 9x.leave0 (x)

b. (29)E 9x.leave0 (x) (Presupposes: !9x.leave0 (x))

17
 Ellipsis in inquisitive semantics

(31) J(29)A K , J(29)E K

11 10 11 10

01 00 01 00

Turning to the infelicitous cases, we look first at double negation. Above, in (7), we

saw that double negation in inquisitive semantics preserves truth-conditions, as one would

expect, yet eliminates the inquisitive content of the formula to which it applies. The re-

sult is that an attempted sluice in (32) is interpreted as in the picture in (33). Applying

the inquisitive entailment condition, we find that the E-clause does entail the A-clause

since each of the alternatives of the form ‘that x left’ is a subalternative of some alter-

native or other in the interpretation of the A-clause (namely, the single alternative ‘that

someone or other left’). In the reverse direction, however, the entailment does not hold.

Given the single alternative in the A-clause, we cannot find any super-alternative in the

E-clause. Since the symmetric entailment condition fails, the account correctly predicts

that sluicing will not be possible in this case.

(32) #[It’s not the case that no one left]A , but I don’t know [who left]E .

(33) J32A K ; J32E K

11 10 11 10

01 00 01 00

18
 Ellipsis in inquisitive semantics

Finally, let’s look at the case of the indefinite inside an appositive relative clause.

A number of recent works have argued in one form or another that appositive relative

clauses have a special discourse status of one sort of another (e.g. Potts (2005), Ama-

ral et al. (2007), Simons et al. (2011), AnderBois et al. (2015)). One aspect of this spe-

cial status is that appositives represent purely informational updates which do not inter-

act with the Questions Under Discussion (QUDs) in any direct way. AnderBois (2014)

proposes that in order to capture this aspect of their meaning, appositives ought to be

treated as lacking the alternative-rich structure inquisitive semantics assumes for at-issue

assertions, and instead be assigned a single classical proposition of type st. Composition-

ally, this is achieved through the Comma operator in (34):

(34) JComma(')K = {w | there is some ↵ 2 J'K s.t. w 2 ↵ }

Since inquisitive meanings are captured as non-singleton sets of alternatives, the

Comma operator ensures that the formula to which it applies – as it enters the discourse

record – will not be inquisitive regardless of its internal composition. As in the case of

double negation, then, appositives deliver truth-conditionally equivalent interpretations,

yet lack the alternative-rich meanings needed to meet the symmetric entailment condi-

tion.

Summing up, we have seen in this section that an account of sluicing based on sym-

metric entailment defined over inquisitive semantic interpretations captures both the data

which motivated Merchant (2001)’s semantic approach (and indeed previous approaches

dating back at least to Sag & Hankamer (1984)) as well a number of other sets of data,

including several cases where sluicing fails despite truth-conditional equivalence.

3.3 Sprouting

Having examined cases of ‘merger’ sluicing with overt indefinite or disjunctive inner an-

tecedents, we turn now to cases of so-called ‘sprouting’ where the wh-phrase has no overt

correlate in the A-clause. In some cases, of course, there is good reason to believe that,

19
 Ellipsis in inquisitive semantics

despite the lack of an overt inner antecedent, there is nonetheless an implicit argument

present in the A-clause. For example, a large body of literature dating back to Fillmore

(1969) holds that apparently intransitive uses of verbs like eat in (35a) include an existen-

tial/indefinite implicit argument (see AnderBois (2012b) for recent discussion of sluicing

and the typology of implicit arguments). Beyond this, there are cases like (35b) where

there is clearly an existential entailment, and arguably also (contextually restricted) exis-

tential quantification depending on one’s semantics for tense.

(35) a. [Fred ate]A , but I don’t know [what Fred ate]E .

b. [Fred baked a cake]A , but I don’t know [when Fred baked a cake]E .

Such data, therefore, have given rise to the claim that in a certain sense, there is no

sprouting, but rather that there is always an implicit argument either syntactically (e.g.

Fortin (2007), Fortin (2011)) or semantically (Merchant (2001)). While this approach is

potentially viable for the above data, Chung (2006) points out that there are other cases

where even an existential entailment (let alone a true indefinite) is clearly not present,

(36). People can finish projects on their own, and Seth can arrive by car, bike, helicopter,

etc. Such cases, therefore, present somewhat the opposite puzzle to what we have seen

in §3.1 for merger. Here, the A-clause is not even truth-conditionally equivalent to the

E-clause, yet sluicing is possible.

(36) a. [He finished the project]A , but we don’t know [with whose help he finished

project]E . Chung (2006)

b. [Seth arrived]A , but I don’t know [on which bus Seth arrived]E .

c. A: [Fred learned French]A

B: [For who did Fred learn French]E ?

For the examples in (36), there is a clear intuition that these instances involve some

kind of accommodation. For (36a), this accommodation is pretty easy given the nature

of projects. For (36b), the ease of accommodation seemingly depends on various kinds

20
 Ellipsis in inquisitive semantics

of world knowledge – Is Seth someone who is likely to take the bus? Are there multiple

buses which he could have taken? etc. While B’s question in (36c) sounds fairly odd out

of the blue, it sounds quite natural in the admittedly unusual context where it is known

that Fred only learns languages to impress foreign visitors.

Unconstrained, however, accommodation runs the risk of overgenerating and predict-

ing that sprouting should be possible quite generally. However, there are many cases, as

in (37), where sprouting remains infelicitous even though it would seem quite plausible

given world knowledge.

(37) a. #Agnes wondered how John could eat, but it’s not clear what. Chung et al.

(1995)

b. A: The fact that Seth arrived was surprising. B: #On which bus?

These sorts of restrictions on sprouting were first noted by Chung et al. (1995), who

attribute them specifically to the presence of syntactic islands, claiming that in contrast

to the well-known island-insensitivity of sluicing more generally, sprouting is sensitive to

islands. However, subsequent work by Romero (1998) and Merchant (2001) argues on the

basis of pairs like (38) that this di↵erence is not limited to islands (since the non-elliptical

control in (38b) is grammatical), and is therefore best captured by appealing to inde-

pendently observable narrow scope of implicit existential quantification. For example,

whereas an indefinite and clausemate negation ordinarily give rise to a scope ambiguity,

an existential implicit argument like the one in ‘Sally didn’t eat’ unambiguously takes

narrow scope relative to negation (i.e. ‘Sally didn’t eat’ does not a have a reading para-

phraseable with ‘There is a thing/meal x such that John didn’t eat x.’).

(38) a. *Ramon is glad that Sally ate, but I don’t remember which dish.

b. I don’t remember which dish he is glad that Sally ate. Romero (1998)

For cases like these with an implicit argument, then, the merger account can be ex-

tended straightforwardly. Felicitous sprouting as in (35) is possible because the A-clause

21
 Ellipsis in inquisitive semantics

has an indefinite implicit argument and therefore has an interpretation which is inquisi-

tive. In examples like (37a) and (38a), there is still an indefinite implicit argument, but

one which cannot take widest scope. Therefore, the whole sentence’s interpretation is

non-inquisitive and sluicing is correctly predicted to be infelicitous.

What, then, about examples like (36) in which we have seen following Chung (2006)

that no implicit argument or even existential entailment is present? Building on the above

intuition that at least some cases of sprouting involve accommodation of some sort, An-

derBois (2014) proposes an account which is partially semantic and partially pragmatic.

On the semantic side, the proposal extends the inquisitivity that we have thus far asso-

ciated with overt indefinites and disjunctions to existential quantification quite generally,

including covert quantification over neo-Davidsonian eventuality arguments. The meaning

of a simple sentence like ‘John left’ not only includes the information that there is some

event or other which is a leaving event and of which John is the agent, but also the is-

sue of which event it is that satisfies these requirements. On the pragmatic side, then, the

account claims that sluicing is felicitous to the extent that the alternatives in the inter-

rogative E-clause covary with that of the A-clause, an accommodation process AnderBois

dubs ‘issue-bridging’. The rest of this section spells out both parts of this proposal a bit

more, though we refer the reader to AnderBois (2014) for further details.

A central notion in inquisitive semantics is the idea that the kind of indeterminacy

we find in indefinites and disjunctions is intimately related (and in some cases composi-

tionally related) to the inquisitivity we find in questions. In both cases, a set of alterna-

tives is made salient, leaving the issue of which alternative(s) in fact hold as at least a

safe potential topic for future conversation. For example, the sentence in (39a) introduces

a set of alternatives in (39c), makes salient the issue of which of these in fact hold, and

conveys the information that at least one does.

22
 Ellipsis in inquisitive semantics

(39) a. Someone left.

b. 9x.leave0 (x)
8 9
>
> John left >
>
>
> >
>
>
> >
>
>
> Maribel left >
>
>
> >
>
< =
c. Alexis left >
>
> >
>
> >
>
>
> Ignacio left >
>
>
> >
>
>
> >
>
: ;
...

Here, we extend this idea beyond overt indefinites to existential quantification over

covert arguments, in this case, the neo-Davidsonian event(uality) argument. As is clear in

(40), the proposed semantics for the covert existential is formally entirely parallel to what

we have seen in (39). A sentence like (40a) introduces a set of alternatives, (40c), makes

salient the issue of which of these alternatives in fact hold, and conveys the information

that at least one does.

(40) a. Seth arrived.

b. 9e.arrive0 (e) ^ Agent(S, e)

8 9
>
> e is an event of Seth arriving>
>
>
> 1 >
>
>
> >
>
>
> e is an event of Seth arriving>
>
>
> 2 >
>
< =
c. e 3 is an event of Seth arriving
>
> >
>
>
> >
>
>
> e4 is an event of Seth arriving> >
>
> >
>
>
> >
>
: ;
...

The issue it makes salient is, however, a somewhat odd one, paraphraseable as “Which

event is an event of Seth leaving?”. The apparent oddity of this issue, however, is due not

to anything about the inquisitive quantification itself, but rather the ontological status

of events in the first place (as discussed, for example, by Parsons (1990)). Although it

is standard to take events to be things in the actual world in more or less the same way

that individuals are, it is far less intuitive to do so for events.

With this semantics in place, we turn now to the pragmatic part of the story. For

concreteness, we will work with the example in (36b), repeated as (41). The semantics

23
 Ellipsis in inquisitive semantics

above holds that the A-clause makes salient a quite fine-grained issue of the form “which

event?”. The E-clause, on the other hand, makes salient a more coarse-grained issue about

some aspect of the event in question, in this case its manner. The inquisitive entailment

condition is therefore not met given the di↵erence in the granularity of these two issues.12

(41) [Seth arrived]A , but I don’t know [on which bus Seth arrived]E .

While these two issues are not identical, the claim is that they are sufficiently simi-

lar that the E-clause can be accommodated. AnderBois (2014) calls this accommodation

process ‘issue-bridging’, on analogy with bridging definite descriptions like that in (42).

The existence of a driver is not simply accommodated directly, but rather by virtue of

a salient relationship with something whose existence and discourse salience are already

established, a bus.

(42) A bus went by. The driver had on sunglasses.

Rather than bridging to an individual, however, indirect sprouting involves bridging

to an issue introduced in the A-clause. Just as a driver is typically an aspect of a bus,

times, locations, manners, etc. are typically aspects of events (see also Barros (2014) for

a related approach which works directly with these categories rather than events). Con-

cretely, then, the prediction is that sprouting (and indeed sluicing more generally) should

be subject to the condition in (43). One important feature of the account to note is that

it relies crucially on the presence of inquisitive material in the A-clause, and thus avoids

overgenerating and allowing examples like (37) since the event quantification in question

does not take wide scope.13

(43) Covariation condition: Sprouting is felicitous to the extent that the context

allows for the inference that the alternatives in the A-clause covary with the alter-

natives in the E-clause.

This section has shown two ways to extend the inquisitive semantic account of sluic-

ing to sprouting – one for cases where an indefinite implicit argument is present, and one

24
 Ellipsis in inquisitive semantics

for cases where no such argument is found. Both accounts rely on independently known

scopal properties of implicit existential quantification to help constrain the account, thus

deriving the asymmetries between merger and sprouting first discussed by Chung et al.

(1995).

4 Conclusions: structure, recoverability, and licensing

In this chapter, we have informally introduced inquisitive semantics, reviewed the most

fleshed out inquisitive semantic account of ellipsis to date – AnderBois (2014)’s work on

sluicing – and explored various kinds of data consistent with this view. We conclude here

by considering how three major questions in the theory of ellipsis – Structure, Recov-

erability, and Licensing – are answered under this account as well as briefly consider-

ing other potential ways of incorporating inquisitive semantics into the theory of ellipsis.

By Structure, we mean the question of what syntactic structure, if any, is found within

the ellipsis site itself. Recoverability refers to the way in which the ellipsis site’s in-

terpretation is arrived at. Licensing covers any additional constraints or conditions on

ellipsis that are not clearly part of the latter two categories.

Since inquisitive semantics is a theory of semantic content rather than of the inter-

faces between semantics and syntax or phonology, it in principle need not impose any

requirements on the theory of ellipsis. This said, for inquisitive semantics to play a role

in accounting for a given ellipsis process, the condition on Recoverability must be at

least partially semantic in nature. Inquisitive semantics locates alternatives in the inter-

pretation itself, rather than in the LF, and so even an LF-syntactic approach to recov-

erability such as Chung et al. (1995) will not suffice. A complete theory of ellipsis must

of course address the other major questions as well, which we do presently both for the

main account described here, AnderBois (2014), as well as briefly discussing how these

answers might change under other potential ways of incorporating inquisitive semantics

into a more comprehensive theory of ellipsis.

25
 Ellipsis in inquisitive semantics

Since the account of sluicing in §3 builds o↵ of Merchant (2001), it addresses these

major issues in largely similar ways. On the question of Structure, both accounts posit

silent linguistic material with ellipsis consisting of PF-deletion. Typically, we assume that

the deleted material is full interrogative clauses, though nothing in the analysis rules out

other underlying structures such as clefts provided that they satisfy the relevant iden-

tity conditions (see, e.g. Barros (2014) for a closely related, but non-inquisitive, approach

making use of this option).

On the question of Recoverability, the primary condition is a semantic one: sym-

metric entailment between A- and E-clauses with entailment crucially being defined over

inquisitive semantic representations rather than just truth-conditions. Beyond this, we

have departed slightly from Merchant (2001), along with Chung (2006) and others in sup-

plementing this semantic condition with a minimal lexico-syntactic one to handle certain

issues that arise in sprouting. Ultimately, then the approach to Recoverability in §3 is

a hybrid one, in line with recent works in a variety of otherwise quite di↵erent approaches

(e.g. Ginzburg & Sag (2001), Chung (2006), Jacobson (2013)).

Finally, for Licensing, it does not seem that inquisitive semantics imposes any par-

ticular constraints on possible accounts. Given the close parallels with Merchant (2001),

we refer the reader to Merchant (2001), Ch. 2 (and Lobeck (1995)’s work cited therein)

for discussion of this issue. One place where inquisitive semantics does help shape the

range of answers to the Licensing question (or perhaps Recoverability) is that it pro-

vides a semantic account to certain kinds of cases that one might have thought were due

to syntactic or other form-based constraints (e.g. the case of double negation above).

However, we again stress that the way the account in §3 answers the question of

Structure is not intrinsic to inquisitive semantics per se, and that there are in princi-

ple many di↵erent frameworks for understanding ellipsis in which inquisitive semantics

could be incorporated. Given the tight connection between inquisitive semantic issues

and Questions Under Discussion (QUDs), one obvious candidate would be to build on

Ginzburg & Sag (2001)’s QUD-based approach. In a nutshell, their approach is a structure-

26
 Ellipsis in inquisitive semantics

free one which, beyond a minimal condition on form referring to a salient utterance (SAL-

UTT), fills in the interpretation of the wh-phrase anaphorically from the maximal QUD

(MAX-QUD).

On a classical semantics for indefinites and disjunctions, however, this approach of-

fers no clear way to explain the privileged role that these elements play in sluicing (in

fact, Ginzburg & Sag (2001) briefly argue against this claim on p. 321). Adding inquisi-

tive semantics to this picture, however, these elements conventionally make salient a pos-

sible QUD, thus explaining their privileged role. While such an account is in many ways

an attractive one, further work on the various linguistic and non-linguistic ways in which

QUDs arise in discourse is needed to make it viable. For example, (44) is a case where a

clear QUD is established contextually, it would seem, and yet sluicing appears to be quite

bad.

(44) Scenario: I see the silhouette of someone is knocking on my office door, but can’t

see the person’s face.

#Who? // #I wonder who. // #Do you know who? // Who is it? // Who’s there?

Another alternative would be to seize upon the deep parallels between inquisitive

semantics and dynamic semantics and draw upon work that treats ellipsis as discourse

reference of a special sort such as Hardt (1993) (possibly supplemented with a lexico-

syntactic condition of some sort). One challenge for such an approach, however, is that

individual discourse reference does not exhibit the same interactions with double nega-

tion and appositives as does sluicing, as discussed by AnderBois (2014). Nonetheless, such

an approach is in principle possible and, again, would give quite di↵erent answers to the

question of Structure at least and possibly Recoverability and Licensing as well.

To summarize, inquisitive semantics proposes that the context change potential of

sentences containing disjunctions, indefinites, and other existential quantification is ‘alternative-

rich’. We expect therefore that we might find ramifications of this richer notion of seman-

tic content in various areas of the grammar. We hope to have shown that this is so for el-

27
 Ellipsis in inquisitive semantics

lipsis and to have shown one way of modeling such e↵ects by focusing primarily on a par-

ticular ellipsis process: sluicing. Given the central role of interrogatives, indefinites, and

disjunction in sluicing, these issues are naturally most salient here. However, it should

be clear that once we adopt alternative-rich sentence meanings, the question arises of

whether other ellipsis processes may similarly require reference to inquisitive semantic

representations in some form.

Notes
1
It should be noted that while this was a driving motivation in many early works in inquisitive semantics
(e.g. Groenendijk (2007), Groenendijk & Roelofsen (2009), Ciardelli (2009)), some more recent works have used
the moniker ‘inquisitive semantics’, yet lack this second property (e.g. Farkas & Roelofsen (t.a.) propose that
declarative clauses include a closure operator eliminating this possibility).
2
Formally, it has been argued that this extension is in fact somewhat more fraught in the case of models
with non-finite domains. See Ciardelli (2009) for detailed discussion.
3
While in principle they are equally applicable, pictorial representations like (2b) become unhelpful for larger
sets of alternatives.
4
While other scope-taking elements such as conjunction and universal quantification are defined by most
authors in ways that allows their alternatives to be ‘passed up’ the composition, one could alternatively de-
fine them in ways that do not have this e↵ect. The choice ultimately depends on how one wishes to handle
scope-taking more generally and we therefore set aside this concern here.
5
We would also note that it is not entirely clear how to interpret felicity judgments for these data. Read-
ing these sentences, it is relatively easy to figure out what such examples were supposed to have meant after
the fact. However, it still seems somewhat unlikely in our opinion that speakers in fact produce such sentences
frequently and possible that they have difficulty processing them when encountering them in natural speech.
Therefore, it seems quite possible to imagine an analysis of this gradient pattern of judgments in which the
grammar of appositives and sluicing does not generate such sentences, but rich context and other ‘repair mech-
anisms’ allow speakers to figure out what they are to have meant when encountered in experimental settings.
In any case, further corpus and experimental work is needed in this area.
6
Barros (2013) also notes a further interaction with the presence or absence of exactly in the E-clause which
we will set aside here.
7
See Barros (2013) and references therein for further discussion and independent linguistic diagnostics of
these categories.

28
 Ellipsis in inquisitive semantics

8
This condition itself builds on the one proposed by Schwarzschild (1999) for deaccenting. It is worth not-
ing, therefore that, as observed by AnderBois (2014), deaccenting appear to be sensitive only to truth-conditions
rather than inquisitive content as well. Consider, for example the contrast between (12) and the following (un-
derlining indicates deaccenting):

(i) It’s not the case that Bill didn’t donate a book to the library, but I don’t know which book he
donated.

9
The focus-closure (F-clo) part of the definition is needed primarily to handle two kinds of sluices which
we will not discuss here: ones where the wh-phrase contains else, as in (i), and so-called ‘contrast’ sluices like
(ii), both examples from Merchant (2001).

(i) Abby called BenF an idiot, but I don’t know who else.

(ii) She has five catsF , but I don’t know how many dogsF .

We set aside these cases here, while not denying their importance.
10
The idea that such a notion of entailment is relevant for natural language has been independently proposed
in the literature on NPI-licensing (von Fintel (1999) et seq.), where it has been dubbed ‘Strawson entailment’.
11
Barker (2013), it should be noted, does briefly address Barros (2013)’s claim that the answer ban follows
from the inquisitive entailment condition, but expresses skepticism for two reasons. First, Barker apparently
was familiar only with AnderBois (2010), which only analyzed cases of merger, whereas AnderBois (2011) and
AnderBois (2014) also address sprouting (see §3.3). Second, Barker notes that Ciardelli et al. (2009) propose
an inquisitive semantic-based account of epistemic possibility modals like might which do not license sluicing:

(i) #John might leave, but I don’t know which.

However, while it is true that Ciardelli et al. (2009)’s account of might is couched in a version of inquis-
itive semantics, their approach does not actually claim that ‘John might leave’ has the same semantics as ‘John
will leave or John won’t leave’, and in fact does not claim that ‘John might leave’ is inquisitive at all (but in-
stead define a new category: ‘attentive’). Therefore, even taking this analysis for might at face value, we do
not predict (i) to be well-formed.
12
In this particular example, there is a second apparent lack of symmetric entailment: the fact that the E-
clause in this example has a presupposition not found in the A-clause. However, this concern is already taken
care of given that we have adopted a notion of Strawson entailment.
13
The account raises a number of issues related to scope-taking which we cannot address here for lack of space.
See Charlow (2014) for recent discussion of scope-taking and sluicing in a closely related semantics.

29
 Ellipsis in inquisitive semantics

References

Alonso-Ovalle, Luis (2006) Disjunction in Alternative Semantics. Ph.D. thesis, University

of Massachusetts.

Amaral, Patricia, Craige Roberts, & E. Allyn Smith (2007) Review of The Logic of Con-

ventional Implicatures. Linguistics and Philosophy 30: 707–749.

AnderBois, Scott (2010) Sluicing as Anaphora to Issues. In Proceedings of SALT 20.

AnderBois, Scott (2011) Issues and Alternatives. Ph.D. thesis, UC Santa Cruz.

AnderBois, Scott (2012a) Focus and Uninformativity in Yucatec Maya Questions. Natural

Language Semantics 20(4): 349–390.

AnderBois, Scott (2012b) Indefiniteness and the typology of implicit arguments. In Pro-

ceedings of West Coast Conference on Formal Linguistics (WCCFL) 30, Cascadilla

Press.

AnderBois, Scott (2014) The semantics of sluicing: beyond truth-conditions. Language

90(4): 887–926.

AnderBois, Scott, Adrian Brasoveanu, & Robert Henderson (2015) At-issue proposals and

appositive impositions in discourse. Journal of Semantics 32(1): 93–138.

Barker, Chris (2013) Scopability and sluicing. Linguistics and Philosophy 36: 187–223.

Barros, Matthew (2013) Harmonic sluicing: which remnant/correlate pairs work and why.

In Proceedings of SALT 23, 295–315.

Barros, Matthew (2014) Sluicing and identity in ellipsis. Ph.D. thesis, Rutgers.

Bhat, D.N.S. (2000) The Indefinite-interrogative puzzle. Linguistic Typology 4: 365–400.

Brown, Roger (1958) How shall a thing be called? Psychological review 65(1): 14–21.

Charlow, Simon (2014) On the semantics of exceptional scope. Ph.D. thesis, NYU.

30
 Ellipsis in inquisitive semantics

Chung, Sandra, William Ladusaw, & James McCloskey (1995) Sluicing and Logical Form.

Natural Language Semantics 3: 239–82.

Chung, Sandra, William Ladusaw, & James McCloskey (2011) Sluicing(:) Between Struc-

ture and Inference. In Representing Language: Essays in Honor of Judith Aissen.

Chung, Sandy (2006) Sluicing and the Lexicon: The Point of No Return. In BLS 31:

General Session and Parasession on Prosodic Variation and Change, 73–91.

Ciardelli, Ivano (2009) Inquisitive Semantics and Intermediate Logics, mSc Thesis, Uni-

versity of Amsterdam.

Ciardelli, Ivano, Jeroen Groenendijk, & Floris Roelofsen (2009) Might and Free Choice in

Inquisitive Semantics. In Proceedings of SALT 19.

Ciardelli, Ivano, Jeroen Groenendijk, & Floris Roelofsen (2013) Inquisitive semantics: a

new notion of meaning. Language and Linguistics Compass 7(9): 459–476.

Collins, James (2013) The discourse potential of narrow scope indefinites in Samoan.

Collins, James, Daria Popova, Ivan Sag, & Thomas Wasow (t.a.) Sluicing and the Inquisi-

tive Potential of Appositives.

Cruse, Alan (1977) The pragmatics of lexical specificity. Journal of Linguistics 13: 153–

164.

Culicover, Peter & Ray Jackendo↵ (2005) Simpler Syntax. Oxford University Press.

Dayal, Veneeta & Roger Schwarzschild (2010) Definite inner antecedents and wh-

correlates in sluices. Rutgers Working Papers in Linguistics 3: 92–114.

Farkas, Donka & Floris Roelofsen (t.a.) Polar initiatives and polar-

ity particle responses in an inquisitive discourse model. Language

Accepted subject to minor revisions, downloaded 7/1/2014 from

https://sites.google.com/site/inquisitivesemantics/papers.

31
 Ellipsis in inquisitive semantics

Fillmore, Charles (1969) Types of lexical information. In Studies in syntax and semantics,

F. Kiefer, ed., Reidel, 109–137.

von Fintel, Kai (1999) NPI licensing, Strawson entailment and context dependency. Jour-

nal of Semantics : 97–148.

Fortin, Catherine (2007) Indonesian sluicing and Verb Phrase Ellipsis: Description and

explanation in the Minimalist framework. Ph.D. thesis, University of Michigan.

Fortin, Catherine (2011) We need LF-copying: a few good reasons why. In Proceedings of

the 28th West Coast Conference on Formal Linguistics, Mary Byram Washburn, Sarah

Ouwayda, Chuoying Ouyang, Bin Yin, Canan Ipek, Lisa Marston, & Aaron Walker,

eds., Cascadilla Press, 87–95.

Ginzburg, Jonathan & Ivan Sag (2001) Interrogative Investigations. CSLI Publications.

Groenendijk, Jeroen (2007) Inquisitive Semantics: Two Possibilities for Disjunction. In

Tbilisi 2007: Seventh International Tbilisi Symposium on Language, Logic, and Com-

munication.

Groenendijk, Jeroen (2011) Erotetic languages and the inquisitive hierarchy. In This is

not a Festschrift: a Festschrift for Martin Stokhof.

Groenendijk, Jeroen & Floris Roelofsen (2009) Inquisitive Semantics and Pragmatics. In

Proceedings of SPR 09.

Haida, Andreas (2008) The Indefiniteness and Focusing of Wh-Words. Ph.D. thesis, Hum-

boldt Universit at, Berlin.

Hamblin, C. L. (1973) Questions in Montague English. Foundations of Language 10: 41–

53.

Hankamer, Jorge & Ivan Sag (1976) Deep and Surface Anaphora. Linguistic Inquiry 7(3):

391–426.

32
 Ellipsis in inquisitive semantics

Hardt, Daniel (1993) Verb phrase ellipsis: Form, meaning and processing. Ph.D. thesis,

University of Pennsylvania.

Haspelmath, Martin (1997) Indefinite Pronouns. Oxford University Press.

Heim, Irene (1982) The Semantics of Definite and Indefinite Noun Phrases. Ph.D. thesis,

University of Massachusetts.

Jacobson, Pauline (2013) The short answer: implications for direct compositionality and

vice versa.

Krahmer, Emiel & Reinhard Muskens (1995) Negation and Disjunction in Discourse Rep-

resentation Theory. Journal of Semantics 12: 357–376.

Kratzer, Angelika & Junko Shimoyama (2002) Indeterminate Pronouns: The View From

Japanese. In The Proceedings of the Third Tokyo Conference on Psycholinguistics, 1–

25.

Lobeck, Anne (1995) Ellipsis: Functional heads, licensing and identification. Oxford Uni-

versity Press.

Merchant, Jason (2001) The Syntax of Silence. Oxford University Press.

Parsons, Terence (1990) Events in the Semantics of English: A Study of Subatomic Se-

mantics. MIT Press.

Potts, Christopher (2005) The Logic of Conventional Implicatures. Oxford University

Press.

Roelofsen, Floris (2013) Algebraic foundations for the semantic treatment of inquisitive

content. Synthese 190(1): 79–102.

Romero, Maribel (1998) Focus and Reconstruction E↵ects in Wh-Phrases. Ph.D. thesis,

University of Massachusetts, Amherst.

Sag, Ivan (1976) Deletion and logical form. Ph.D. thesis, MIT.

33
 Ellipsis in inquisitive semantics

Sag, Ivan & Jorge Hankamer (1984) Toward a theory of anaphoric processing. Linguistics

and Philosophy 7: 325–345.

Schwarzschild, Roger (1999) GIVENness, Avoid F and other Constraints on the Place-

ment of Focus. Natural Language Semantics 7(2): 141–177.

Simons, Mandy (2005) Semantics and Pragmatics in the Interpretation of or. In Proceed-

ings of SALT 15.

Simons, Mandy, Judith Tonhauser, David Beaver, & Craige Roberts (2011) What projects

and why. In Proceedings of Semantics And Linguistic Theory (SALT) 20.

34