I S

K  F I H

MIT S  E

A note about the lecture notes:

The notes for this course have been evolving for years now, starting with some
old notes from the early s by Angelika Kratzer, Irene Heim, and myself,
which have since been modified and expanded every year by Irene or myself.
Because this version of the notes has not been seen by my co-author, I alone am
responsible for any defects.

– Kai von Fintel, Spring 

This is a work in progress. We may eventually publish these materials as a follow-

up volume to Heim & Kratzer’s Semantics in Generative Grammar, Blackwell
. In the meantime, we encourage the use of these notes in courses at other
institutions. Of course, you need to give full credit to the authors and you may
not use the notes for any commercial purposes. If you use the notes, we would
like to be notified and we would very much appreciate any comments, criticism,
and advice on these materials. We have already profited from feedback sent in
by several people who have used the notes and the more the better.

Direct your communication to:

Kai von Fintel

Department of Linguistics & Philosophy
Room ·
Massachusetts Institute of Technology
 Massachusetts Avenue
Cambridge,  -
U S  A

[email protected]
http://kaivonfintel.org

Here is the homepage for the course that these notes are designed for:
http://stellar.mit.edu/S/course//sp/.
Advice about using these notes
. These notes presuppose familiarity with the material, concepts, and nota-
tion of the Heim & Kratzer textbook.
. There are numerous exercises throughout the notes. It is highly recom-
mended to do all of them and it is certainly necessary to do so if you at all
anticipate doing semantics-related work in the future.
. At the moment, the notes are designed to go along with explanatory lectures.
You should ask questions and make comments as you work through the
notes.
. Students with semantic ambitions should also at an early point start reading
supplementary material (as for example listed at the end of each chapter of
these notes).
. Lastly, prospective semanticists may start thinking about how they would
teach this material.
— T     —
 C

 Beginnings 
. Displacement 
. An Intensional Semantics in  Easy Steps 
. Comments and Complications 
. Supplemental Readings 

 Propositional Attitudes 
. Hintikka’s Idea 
. Accessibility Relations 
. Supplemental Readings 

 Modality 
. The Quantificational Theory of Modality 
. Flavors of Modality 
. *Kratzer’s Conversational Backgrounds 
. Supplementary Readings 

 Conditionals 
. The Material Implication Analysis 
. The Strict Implication Analysis 
. If -Clauses as Restrictors 
Supplemental Readings 

 Ordering 
. The Driveway 
. Kratzer’s Solution: Doubly Relative Modality 
. The Paradox of the Good Samaritan 
. Non-Monotonicity of Conditionals 
Supplemental Readings 

 Basics of Tense and Aspect 

. A First Proposal for Tense 
. Are Tenses Referential? 
. The Need for Intervals 
. Aktionsarten 
. The Progressive 
 . Tense in Embedded Clauses 
Supplemental Readings 

 DPs and Scope in Modal Contexts 

. De re vs. De dicto as a Scope Ambiguity 
. Raised subjects 

 Beyond de re — de dicto : The Third Reading 

. A Problem: Additional Readings and Scope Paradoxes 
. The Standard Solution: Overt World Variables 
. Alternatives to Overt World Variables 
. Scope, Restrictors, and the Syntax of Movement 
. A Recurring Theme: Historical Overview 

Bibliography 
 C O
B

Language is the main instrument

of man’s refusal to accept the world
as it is.
George Steiner, After Babel, p. 

We introduce the idea of extension vs. intension and its main use: taking
us from the actual here and now to past, future, possible, counterfac-
tual situations. We develop a compositional framework for intensional
semantics.

. Displacement 
. An Intensional Semantics in  Easy Steps 
.. Laying the Foundations 
.. Intensional Operators 
. Comments and Complications 
.. Intensions All the Way? 
.. Why Talk about Other Worlds? 
.. The Worlds of Sherlock Holmes 
. Supplemental Readings 

. Displacement
Hockett () in a famous article (and a follow-up, Hockett & Altmann ()) Hockett, Charles F. .
presented a list of     . This list continues to The origin of speech. Sci-
play a role in current discussions of animal communication. One of the design entific American . –
 B C 



features is . Human language is not restricted to discourse about

the actual here and now.
How does natural language untie us from the actual here and now? One
degree of freedom is given by the ability to name entities and refer to them even
if they are not where we are when we speak:
() Thomas is in Hamburg.
This kind of displacement is not something we will explore here. We’ll take it
for granted.
Consider a sentence with no names of absent entities in it:
() It is snowing (in Cambridge).
On its own, () makes a claim about what is happening right now here in
Cambridge. But there are devices at our disposal that can be added to (),
resulting in claims about snow in displaced situations. Displacement can occur in
*The terms  and - the  dimension and/or in what might be called the * dimension.
 descend from the Latin
modus, “way”, and are ancient
Here’s an example of temporal displacement:
terms pertaining to the way a
proposition holds, necessarily, () At noon yesterday, it was snowing in Cambridge.
contingently, etc.
This sentence makes a claim not about snow now but about snow at noon
yesterday, a different time from now.
Here’s an example of modal displacement:
() If the storm system hadn’t been deflected by the jet stream, it would
have been snowing in Cambridge.
This sentence makes a claim not about snow in the actual world but about snow
in the world as it would have been if the storm system hadn’t been deflected by
the jet stream, a world distinct from the actual one (where the system did not hit
us), a merely  .
Natural language abounds in modal constructions. () is a so-called -
See Kratzer (, ) for  . Here are some other examples:
more examples of modal
constructions.
() M A
It may be snowing in Cambridge.
() M A
Possibly, it will snow in Cambridge tomorrow.
() P A
Jens believes that it is snowing in Cambridge.

 Steiner (: ) writes: “Hypotheticals, ‘imaginaries’, conditionals, the syntax of counter-
factuality and contingency may well be the generative centres of human speech”.
§. A I S   E S



() H
Jane smokes.
() G
Bears like honey.
The plan for this course is as follows. In Part , we explore modality and
associated topics. In Part , we explore temporal matters.
In this chapter, we will put in place the basic framework of 
, the kind of semantics that models displacement of the point of
evaluation in temporal and modal dimensions. To do this, we will start with one
rather special example of modal displacement:
() In the world of Sherlock Holmes, a detective lives at B Baker Street.
() doesn’t claim that a detective lives at B Baker Street in the actual world
(presumably a false claim), but that in the world as it is described in the Sherlock Check out http://
bakerstreet.org/.
Holmes stories of Sir Arthur Conan Doyle, a detective lives at B Baker Street
(a true claim, of course). We choose this example rather than one of the more run-
of-the-mill displacement constructions because we want to focus on conceptual
and technical matters before we do serious empirical work.
The questions we want to answer are: How does natural language achieve
this feat of modal displacement? How do we manage to make claims about other
possible worlds? And why would we want to?
The basic idea of the account we’ll develop is this:
• expressions are assigned their semantic values relative to a possible world;
• in particular, sentences have truth-values in possible worlds;
• in the absence of modal displacement, we evaluate sentences with respect
to the “actual” world, the world in which we are speaking;
• modal displacement changes the world of evaluation;
• displacement is effected by special operators, whose semantics is our pri-
mary concern here.
A terminological note: we will call the sister of the intensional operator its
, a useful term introduced by our medieval colleagues.

. An Intensional Semantics in  Easy Steps

.. Laying the Foundations
S : P W. Our first step is to introduce possible worlds. This is
not the place to discuss the metaphysics of possible worlds in any depth. Instead,
we will just start working with them and see what they can do for us. Basically, a
possible world is a way that things might have been. In the actual world, there
 B C 



are two coffee mugs on my desk, but there could have been more or less. So,
there is a possible world — albeit a rather bizarre one — where there are  coffee
mugs on my desk. We join Heim & Kratzer in adducing this quote from Lewis
(: f.):

David Lewis
The world we live in is a very inclusive thing. Every stick and every
stone you have ever seen is part of it. And so are you and I. And
so are the planet Earth, the solar system, the entire Milky Way, the
remote galaxies we see through telescopes, and (if there are such
things) all the bits of empty space between the stars and galaxies.
There is nothing so far away from us as not to be part of our world.
Anything at any distance at all is to be included. Likewise the world
is inclusive in time. No long-gone ancient Romans, no long-gone
pterodactyls, no long-gone primordial clouds of plasma are too far
in the past, nor are the dead dark stars too far in the future, to be
part of the same world. . . .
The way things are, at its most inclusive, means the way the en-
tire world is. But things might have been different, in ever so many
ways. This book of mine might have been finished on schedule. Or,
had I not been such a commonsensical chap, I might be defending
not only a plurality of possible worlds, but also a plurality of impos-
sible worlds, whereof you speak truly by contradicting yourself. Or I
might not have existed at all — neither myself, nor any counterparts
of me. Or there might never have been any people. Or the physical
constants might have had somewhat different values, incompatible
with the emergence of life. Or there might have been altogether
different laws of nature; and instead of electrons and quarks, there
might have been alien particles, without charge or mass or spin
but with alien physical properties that nothing in this world shares.
There are ever so many ways that a world might be: and one of these
many ways is the way that this world is.
Previously, our “metaphysical inventory” included a domain of entities and a
set of two truth-values and increasingly complex functions between entities,
truth-values, and functions thereof. Now, we will add possible worlds to the
inventory. Let’s assume we are given a set W , the set of all possible worlds, which
is a vast space since there are so many ways that things might have been different
from the way they are. Each world has as among its parts entities like you and me
and these coffee mugs. Some of them may not exist in other possible worlds. So,
strictly speaking each possible worlds has its own, possibly distinctive, domain
of entities. What we will use in our system, however, will be the grand union of
all these world-specific domains of entities. We will use D to stand for the set of
all possible individuals.
§. A I S   E S



Among the many possible worlds that there are — according to Lewis, there
is a veritable plenitude of them — is the world as it is described in the Sherlock
Holmes stories by Sir Arthur Conan Doyle. In that world, there is a famous
detective Sherlock Holmes, who lives at B Baker Street in London and has a
trusted sidekick named Dr. Watson. Our sentence In the world of Sherlock Holmes,
a detective lives at B Baker Street displaces the claim that a famous detective
lives at B Baker Street from the actual world to the world as described in the
Sherlock Holmes stories. In other words, the following holds:
() The sentence In the world of Sherlock Holmes, a detective lives at B
Baker Street is true in a world w iff the sentence a detective lives at B
Baker Street is true in the world as it is described in the Sherlock Holmes
stories.
What this suggests is that we need to make space in our system for having devices
that control in what world a claim is evaluated. This is what we will do now.

S : T E W P. Recall from H& K that we

were working with a semantic interpretation function that was relativized to
an assignment function g, which was needed to take care of pronouns, traces,
variables, etc. From now on, we will relativize the semantic values in our
system to possible worlds as well. What this means is that from now on, our
interpretation function will have two superscripts: a world w and an assignment
g: J·Kw,g .
So, the prejacent embedded in () will have its truth-conditions described
as follows:
() Ja famous detective lives at B Baker StreetKw,g = 
iff a famous detective lives at B Baker Street in world w.
It is customary to refer to the world for which we are calculating the extension
of a given expression as the  . In the absence of any shifting
devices, we would normally evaluate a sentence in the actual world. But then
there are shifting devices such as our in the world of Sherlock Holmes. We will
soon see how they work. But first some more pedestrian steps: adding lexical
entries and composition principles that are formulated relative to a possible
world. This will allow us to derive the truth-conditions as stated in () in a
compositional manner.

 We will see in Section .. that this is not quite right. It’ll do for now.
 Recall from H& K, pp.f, that what’s inside the interpretation brackets is a mention of an
object language expression. They make this clear by bold-facing all object language expressions
inside interpretation brackets. In these notes, we will follow common practice in the field and
not use a special typographic distinction, but let it be understood that what is interpreted are
object language expressions.
 B C 



S : L E. Among our lexical items, we can distinguish between
items which have a - semantic value and those that are world-
independent. Predicates are typically world-dependent. Here are some sample
entries.

() For any w ∈ W and any assignment function g:

a. JfamousKw,g = λx ∈ D. x is famous in w.,
b. JdetectiveKw,g = λx ∈ D. x is a detective in w.
c. Jlives-atKw,g = λx ∈ D. λy ∈ D. y lives-at x in w.

The set of detectives will obviously differ from world to world, and so will the
set of famous individuals and the set of pairs where the first element lives at the
second element.
Other items have semantic values which do not differ from world to world.
The most important such items are certain “logical” expressions, such as truth-
Note the ruthless condensation functional connectives and determiners:
of the notation in (c) and (d).
() a. JandKw,g = λu ∈ Dt . λv ∈ Dt . u = v = .
b. JtheKw,g = λf ∈ Dhe,ti : ∃!x. f(x) = . the y such that f(y) = .
c. JeveryKw,g = λfhe,ti . λghe,ti . ∀xe : f(x) =  → g(x) = .
d. Ja/someKw,g = λfhe,ti . λghe,ti . ∃xe : f(x) =  & g(x) = .

Note that there is no occurrence of w on the right-hand side of the entries in

(). That’s the tell-tale sign of the world-independence of the semantics of these
items.
We will also assume that proper names have world-independent semantic
values, that is, they refer to the same individual in any possible world.

() a. JNoam ChomskyKw,g = Noam Chomsky.

b. JSherlock HolmesKw,g = Sherlock Holmes.
c. JB Baker StreetKw,g = B Baker Street.

S : C P. The old rules of Functional Application,

Predicate Modification, and λ-Abstraction can be retained almost intact. We
just need to modify them by adding world-superscripts to the interpretation
function. For example:

 Of course, “λx ∈ D. . . . ” is short for “λx : x ∈ D. . . . ”. Get used to semanticists condensing

their notation whenever convenient!
 Always make sure that you actually understand what the notation means. Here, for example,
we are saying that the semantic value of the word famous with respect to a given possible world
w and a variable assignment g is that function that is defined for an argument x only if x is a
member of the domain of individuals and that, if it is defined, yields the truth-value  if and
only if x is famous in w.
§. A I S   E S



() F A (FA)

If α is a branching node and {β, γ} the set of its daughters, then, for
any world w and assignment g: if JβKw,g is a function whose domain
contains JγKw,g , then JαKw,g = JβKw,g (JγKw,g ).
The rule simply passes the world parameter down.
S : T. Lastly, we will want to connect our semantic system to the
notion of the    . We first adopt the “Appropriateness
Condition” from Heim & Kratzer (p.):
() A C
A context c is appropriate for an LF φ only if c determines a variable
assignment gc whose domain includes every index which has a free
occurrence in φ.
We then intensionalize Heim & Kratzer’s definition of truth and falsity of
utterances:
() T  F C  U
An utterance of a sentence φ in a context c in a possible world w is true
iff JφKw,gc =  and false if JφKw,gc = .
E .: Compute under what conditions an utterance in possible world
w (which may or may not be the one we are all living in) of the sentence a
famous detective lives at B Baker Street is true. [Since this is the first exercise of
the semester, please do this in excrutiating detail, not skipping any steps.] 

.. Intensional Operators

So far we have merely “redecorated” our old system inherited from last semester.
We have introduced possible worlds into our inventory, our lexical entries and
our old composition principles. But with the tools we have now, all we can do
so far is to keep track of the world in which we evaluate the semantic value of
an expression, complex or lexical. We will get real mileage once we introduce
  which are capable of shifting the world parameter.
We mentioned that there are a number of devices for modal displacement. As
advertised, for now, we will just focus on a very particular one: the expression
in the world of Sherlock Holmes. We will assume, as seems reasonable, that this
expression is a sentence-modifier both syntactically and semantically.
S : A S E. We begin with a heuristic step. We want
to derive something like the following truth-conditions for our sentence:
() Jin the world of Sherlock Holmes,
a famous detective lives at B Baker StreetKw,g = 
 B C 



iff the world w 0 as it is described in the Sherlock Holmes stories is such

that there exists a famous detective in w 0 who lives at B Baker Street
in w 0 .
We would get this if in general we had this rule for in the world of Sherlock
Holmes:
() For any sentence φ, any world w, and any assignment g:
Jin the world of Sherlock Holmes φKw,g = 
iff the world w 0 as it is described in the Sherlock Holmes stories is such
that JφKw ,g = .
0

This is a so-called  treatment of the meaning of this expres-

sion. Instead of giving an explicit semantic value to the expression, we specify
what effect it has on the meaning of a complex expression that contains it. In (),
we do not compute the meaning for in the world of Sherlock Holmes, φ from the
combination of the meanings of its parts, since in the world of Sherlock Holmes
is not given a separate meaning, but in effect triggers a special composition
The diamond ♦ symbol for principle. This format is very common in modal logic systems, which usually
possibility is due to C.I. Lewis,
first introduced in Lewis &
give a syncategorematic semantics for the two modal operators (the necessity
Langford (), but he made operator  and the possibility operator ♦). When one only has a few closed
no use of a symbol for the dual
combination ¬♦¬. The dual class expressions to deal with that may shift the world parameter, employing
symbol  was later devised by syncategorematic entries is a reasonable strategy. But we are facing a multitude
F.B. Fitch and first appeared in
print in  in a paper by his of displacement devices. So, we will need to make our system more modular.
doctoral student Barcan (). So, we want to give in the world of Sherlock Holmes its own meaning and com-
See footnote  of Hughes
& Cresswell (). Another bine that meaning with that of its prejacent by a general composition principle.
notation one finds is L for The Fregean slogan we adopted says that all composition is function application
necessity and M for possibility,
the latter from the German (modulo the need for λ-abstraction and the possible need for predicate modifica-
möglich ‘possible’. tion). So, what we will want to do is to make () be the result of functional
application. But we can immediately see that it cannot be the result of our usual
rule of functional application, since that would feed to in the world of Sherlock
Holmes the semantic value of a famous detective lives in B Baker Street in w,
which would be a particular truth-value,  if a famous detective lives at B
Baker Street in w and  if there doesn’t. And whatever the semantics of in the
world of Sherlock Holmes is, it is certainly not a truth-functional operator.
So, we need to feed something else to in the world of Sherlock Holmes. At the
same time, we want the operator to be able to shift the evaluation world of its
prejacent. Can we do this?

S : I. We will define a richer notion of semantic value, the

 of an expression. This will be a function from possible worlds to

 See Heim & Kratzer, Section ., pp. – for a reminder about the status of predicate
modification.
§. A I S   E S



the extension of the expression in that world. The intension of a sentence can
be applied to any world and give the truth-value of the sentence in that world.
Intensional operators take the intension of their prejacent as their argument, that
is we will feed the intension of the embedded sentence to the shifting operator.
The operator will use that intension and apply it to the world it wants the
evaluation to happen in. Voilà.
Now let’s spell that account out. Our system actually provides us with two
kinds of meanings. For any expression α, we have JαKw,g , the semantic value
of α in w, also known as the  of α in w. But we can also calculate
λw.JαKw,g , the function that assigns to any world w the extension of α in that
world. This is usually called the  of α. We will sometimes use an
abbreviatory notation for the intension of α:
g
() JαK¢ := λw.JαKw,g .

It should be immediately obvious that since the definition of intension abstracts

over the evaluation world, intensions are not world-dependent.,
Note that strictly speaking, it now makes no sense anymore to speak of
“the semantic value” of an expression α. What we have is a semantic system
that allows us to calculate extensions (for a given possible world w) as well as
intensions for all (interpretable) expressions. We will see that when α occurs in a
particular bigger tree, it will always be determinate which of the two “semantic
values” of α is the one that enters into the compositional semantics. So, that
one — whichever one it is, the extension or the intension of α — might then be
called “the semantic value of α in the tree β”.
It should be noted that the terminology of  vs.  is The Port-Royal logicians dis-
tinguished  from
time-honored but that the possible worlds interpretation thereof is more recent. . Leibniz
The technical notion we are using is certainly less rich a notion of meaning than preferred the term 
rather than .
tradition assumed. The notion probably goes
back even further. See Spencer
() for some notes on this.
The possible worlds interpre-
tation is due to Carnap ().
 The notation with the subscripted cent-sign comes from Montague Grammar. See e.g. Dowty
et al. (: ).
 Since intensions are by definition not dependent on the choice of a particular world, it makes
,g
no sense to put a world-superscript on the intension-brackets. So don’t ever write “J. . .Kw¢ ”;
we’ll treat that as undefined nonsense.
 The definition here is simplified, in that it glosses over the fact that some expressions, in
particular those that contain  , may fail to have an extension in certain
worlds. In such a case, the intension has no extension to map such a world to. Therefore, the
intension will have to be a partial function. So, the official, more “pedantic”, definition will have
to be as follows: JαKg¢ := λw : α ∈ dom(JKw,g ).JαKw,g .
 For example, Frege’s “modes of presentation” are not obviously captured by this possible
worlds implementation of extension/intension.
 B C 



S : S T  S D. If we want to be able to

feed the intensions to lexical items like in the world of Sherlock Holmes, we need
to have the appropriate types in our system.
Recall that W is the set of all possible worlds. And recall that D is the set of
all   and thus contains all individuals existing in the actual
world plus all individuals existing in any of the merely possible worlds.
We now expand the set of semantic types, to add intensions. Intensions are
functions from possible worlds to all kinds of extensions. So, basically we want
to add for any kind of extension we have in our system, a corresponding kind of
intension, a function from possible worlds to that kind of extension.
We add a new clause, (c), to the definition of semantic types:
() S T
a. e and t are semantic types.
b. If σ and τ are semantic types, then hσ, τi is a semantic type.
c. If σ is a semantic type, then hs, σi is a semantic type.
d. Nothing else is a semantic type.
We also add a fourth clause to the previous definition of semantic domains:
() S D
a. De = D, the set of all possible individuals
b. Dt = {, }, the set of truth-values
c. If σ and τ are semantic types, then Dhσ,τi is the set of all functions
from Dσ to Dτ .
d. I: If σ is a type, then Dhs,σi is the set of all functions
from W to Dσ .
Clause (d) is the addition to our previous system of types. The functions of the
schematic type hs, . . .i are intensions. Here are some examples of intensions:
• The intensions of sentences are of type hs, ti, functions from possible
worlds to truth values. These are usually called . Note that
if the function is total, then we can see the sentence as picking out a
set of possible worlds, those in which the sentence is true. More often
than not, however, propositions will be  functions from worlds
to truth-values, that is functions that fail to map certain possible worlds
into either truth-value. This will be the case when the sentence contains
a presupposition trigger, such as the. The famous sentence The King of

 Note a curious feature of this set-up: there is no type s and no associated domain. This
corresponds to the assumption that there are no expressions of English that take as their extension
a possible world, that is, there are no pronouns or names referring to possible worlds. We will
actually question this assumption in a later chapter. For now, we will stay with this more
conventional set-up.
§. A I S   E S



France is bald has an intension that (at least in the analysis sketched in
Heim & Kratzer) is undefined for any world where there fails to be a
unique King of France.
• The intensions of one-place predicates are of type hs, he, tii, functions
from worlds to set of individuals. These are usually called .
• The intensions of expressions of type e are of type hs, ei, functions from
worlds to individuals. These are usually called  .

S : A L E   S. We are ready to formulate the lexical
entry for in the world of Sherlock Holmes:
() Jin the world of Sherlock HolmesKw,g =
λphs,ti . the world w 0 as it is described in the Sherlock Holmes stories
is such that p(w 0 ) = .
That is, in the world of Sherlock Holmes expects as its argument a function of type
hs, ti, a proposition. It yields the truth-value  iff the proposition is true in the
world as it is described in the Sherlock Holmes stories.
All that’s left to do now is to provide in the world of Sherlock Holmes with a
proposition as its argument. This is the job of a new composition principle.

S : I F A. We add the new rule of

Intensional Functional Application.
() I F A (IFA)
If α is a branching node and {β, γ} the set of its daughters, then, for
any world w and assignment g: if JβKw,g is a function whose domain
contains JγKg¢ , then JαKw,g = JβKw,g (JγKg¢ ).
This is the crucial move. It makes space for expressions that want to take the
intension of their sister as their argument and do stuff to it. Now, everything is
in place. Given (), the semantic argument of in the world of Sherlock Holmes
will not be a truth-value but a proposition. And thus, in the world of Sherlock
Holmes will be able to check the truth-value of its prejacent in various possible
worlds. To see in practice that we have all we need, please do the following
exercise.
E .: Calculate the conditions under which an utterance in a given
possible world w of the sentence in the world of the Sherlock Holmes stories, a
famous detective lives at B Baker Street is true. 

 This is not yet the final semantics, see Section . for complications. One complication we
will not even start to discuss is that obviously it is not a necessity that there are Sherlock Holmes
stories in the first place and that the use of this operator presupposes that they exist; so a more
fully explicit semantics would need to build in that presuppositional component. Also, note
again the condensed notation: “λphs,ti . . . . ” stands for the fully official “λp : p ∈ Dhs,ti . . . . ”.
 B C 



E .: What in our system prevents us from computing the extension of
Watson is slow, for example, by applying the intension of slow to the extension
of Watson? What in our system prevents us from computing the extension of
Watson is slow by applying the intension of slow to the intension of Watson? 

E .: What is wrong with the following equation:

() (λx. x is slow in w) (Watson) = Watson is slow in w.
[ Hint: there is nothing wrong with the following:
() (λx. x is slow in w) (Watson) =  iff Watson is slow in w. ] 

. Comments and Complications

.. Intensions All the Way?
We have seen that to adequately deal with expressions like in the world of
Sherlock Holmes, we need an intensional semantics, one that gives us access to the
extensions of expressions across the multitude of possible worlds. At the same
time, we have kept the semantics for items like and, every, and a unchanged and
extensional. This is not the only way one can set up an intensional semantics.
The following exercise demonstrates this.

E .: Consider the following “intensional” meaning for and:

() JandKw,g = λphs,ti . λqhs,ti . p(w) = q(w) = .

With this semantics, and would operate on the intensions of the two conjoined
sentences. In any possible world w, the complex sentence will be true iff the
component propositions are both true of that world.
Compute the truth-conditions of the sentence In the world of Sherlock Holmes,
Holmes is quick and Watson is slow both with the extensional meaning for and
given earlier and the intensional meaning given here. Is there any difference in
the results? 

There are then at least two ways one could develop an intensional system.
(i) We could “generalize to the worst case” and make the semantics deliver
intensions as the semantic value of an expression. Such systems are common
in the literature (see Cresswell ; Lewis b).
(ii) We could maintain much of the extensional semantics we have developed
so far and extend it conservatively so as to account for non-extensional
contexts.
§. C  C



We have chosen to pursue (ii) over (i), because it allows us to keep the semantics
of extensional expressions simpler. The philosophy we follow is that we will only
move to the intensional sub-machinery when triggered by an expression that
creates a non-extensional context. As the exercise just showed, this is more a
matter of taste than a deep scientific decision.

.. Why Talk about Other Worlds?

Why would natural language bother having such elaborate mechanisms to talk
about other possible worlds? While having devices for spatial and temporal
displacement (talking about Hamburg or what happened yesterday) seems emi-
nently reasonable, talking about worlds other than the actual world seems only
suitable for poets and the like. So, why?
The solution to this puzzle lies in a fact that our current semantics of the
shifter in the world of Sherlock Holmes does not yet accurately capture: modal
sentences have empirical content, they make  claims, claims that
are true or false depending on the circumstances in the actual world.
Our example sentence In the world of Sherlock Holmes, a famous detective
lives at  Baker Street is true in this world but it could easily have been false.
There is no reason why Sir Arthur Conan Doyle could not have decided to locate
Holmes’ abode on Abbey Road.
To see that our semantics does not yet capture this fact, notice that in the
semantics we gave for in the world of Sherlock Holmes:
() Jin the world of Sherlock HolmesKw,g =
λphs,ti . the world w 0 as it is described in the Sherlock Holmes stories
is such that p(w 0 ) = .
there is no occurrence of w on the right hand side. This means that the truth-
conditions for sentences with this shifter are world-independent. In other words,
they are predicted to make non-contingent claims that are either true no-matter-
what or false no-matter-what. This needs to be fixed.
The fix is obvious: what matters to the truth of our sentence is the content
of the Sherlock Holmes stories as they are in the evaluation world. So, we need
the following semantics for our shifter:
() Jin the world of Sherlock HolmesKw,g =
λphs,ti . the world w 0 as it is described in the Sherlock Holmes stories
in w is such that p(w 0 ) = .
We see now that sentences with this shifter do make a claim about the evaluation
world: namely, that the Sherlock Holmes stories as they are in the evaluation
world describe a world in which such-and-such is true. So, what is happening
is that although it appears at first as if modal statements concern other possible
 B C 



worlds and thus couldn’t really be very informative, they actually only talk about
certain possible worlds, those that stand in some relation to what is going on at
the ground level in the actual world. As a crude analogy, consider:
() My grandmother is sick.
At one level this is a claim about my grandmother. But it is also a claim about me:
namely that I have a grandmother who is sick. Thus it is with modal statements.
They talk about possible worlds that stand in a certain relation to the actual
world and thus they make claims about the actual world, albeit slightly indirectly.

.. The Worlds of Sherlock Holmes

So far, we have played along with colloquial usage in talking of the world of
Sherlock Holmes. But it is important to realize that this is sloppy talk. Lewis
() writes:
[I]t will not do to follow ordinary language to the extent of supposing
that we can somehow single out a single one of the worlds [as the
one described by the stories]. Is the world of Sherlock Holmes a
world where Holmes has an even or an odd number of hairs on
his head at the moment when he first meets Watson? What is
Inspector Lestrade’s blood type? It is absurd to suppose that these
questions about the world of Sherlock Holmes have answers. The
best explanation of that is that the worlds of Sherlock Holmes are
plural, and the questions have different answers at different ones.
The usual move at this point is to talk about the set of worlds “ 
the (content of ) Sherlock Holmes stories in w”. We imagine that we ask of each
possible world whether what is going on in it is compatible with the stories as
they were written in our world. Worlds where Holmes lives on Abbey Road are
not compatible. Some worlds where he lives at B Baker Street are compatible
(again not all, because in some such worlds he is not a famous detective but
an obscure violinist). Among the worlds compatible with the stories are ones
where he has an even number of hairs on his head at the moment when he first
meets Watson and there are others where he has an odd number of hairs at that
moment.
What the operator in the world of Sherlock Holmes expresses is that its comple-
ment is true throughout the worlds compatible with the stories. In other words,
the operator universally quantifies over the compatible worlds. Our next iteration
of the semantics for the operator is therefore this:
() Jin the world of Sherlock HolmesKw,g =
λphs,ti . ∀w 0 compatible with the Sherlock Holmes stories in w :
p(w 0 ) = .
§. S R



At a very abstract level, the way we parse sentences of the form in the world of
Sherlock Holmes, φ is that both components, the in-phrase and the prejacent,
determine sets of possible worlds and that the set of possible worlds representing
the content of the fiction mentioned in the in-phrase is a subset of the set
of possible worlds determined by the prejacent. We will see the same rough
structure of relating sets of possible worlds in other intensional constructions.
This is where we will leave things. There is more to be said about fiction
operators like in the world of Sherlock Holmes, but we will just refer to you to the
relevant literature. In particular, one might want to make sense of Lewis’ idea
that a special treatment is needed for cases where the sentence makes a claim
about things that are left open by the fiction (no truth-value, perhaps?). One
also needs to figure out how to deal with cases where the fiction is internally
inconsistent. In any case, for our purposes we’re done with this kind of operator.

. Supplemental Readings

There is considerable overlap between this chapter and Chapter  of Heim &
Kratzer’s textbook:
Heim, Irene & Angelika Kratzer. . Semantics in generative grammar. Oxford:
Blackwell.
Here, we approach intensional semantics from a different angle. It would
probably be beneficial if you read H& K’s Chapter  in addition to this chapter
and if you did the exercises in there.
Come to think of it, some other ancillary reading is also recommended. You may
want to look at relevant chapters in other textbooks:
Dowty, David, Robert Wall & Stanley Peters. . Introduction to Montague
semantics. Kluwer. [Chapters & ].
Gamut, L. T. F. . Logic, language, and meaning. Chicago University Press.
[Volume II: Intensional Logic and Logical Grammar].
Chierchia, Gennaro & Sally McConnell-Ginet. . Meaning and grammar: An
introduction to semantics (nd edition). MIT Press. [Chapter : Intensionality].
An encyclopedia article by Perry on possible worlds semantics:
Perry, John R. . Semantics, possible worlds. In E. Craig (ed.), Routledge
encyclopedia of philosophy, London: Routledge. Preprint http://www-csli.
stanford.edu/~john/PHILPAPERS/posswld.pdf.
A couple of influential philosophical works on the metaphysics and uses of
possible worlds:
 B C 



Kripke, Saul. . Naming and necessity. Oxford: Blackwell.

Lewis, David. . On the plurality of worlds. Oxford: Blackwell.
An interesting paper on the origins of the modern possible worlds semantics for
modal logic:
Copeland, B. Jack. . The genesis of possible worlds semantics. Journal of
Philosophical Logic (). –. doi:./A:.
A personal history of formal semantics:
Partee, Barbara H. . Reflections of a formal semanticist as of Feb .
Ms. (longer version of introductory essay in  book). URL http://people.
umass.edu/partee/docs/BHP_Essay_Feb.pdf.
A must read for students who plan to go on to becoming specialists in semantics,
together with a handbook article putting it in perspective:
Montague, Richard. . The proper treatment of quantification in ordinary
English. In Jaako Hintikka, Julius Moravcsik & Patrick Suppes (eds.), Ap-
proaches to natural language, –. Dordrecht: Reidel. URL http://www.
blackwellpublishing.com/content/BPL_Images/Content_store/Sample_chapter/
/Portner.pdf. Reprinted in Portner & Partee (), pp. –.
Partee, Barbara H. & Herman L.W. Hendriks. . Montague grammar. In
Johan van Benthem & Alice ter Meulen (eds.), Handbook of logic and language,
–. Elsevier.
To learn more about discourse about fiction, read Lewis:
Lewis, David. . Truth in fiction. American Philosophical Quarterly (). –
. URL http://www.jstor.org/stable/. Reprinted with postscripts in
Lewis (), pp. –.
Recent reconsiderations:
Bonomi, Andrea & Sandro Zucchi. . A pragmatic framework for truth
in fiction. Dialectica (). –. doi:./j.-..tb.x.
Preprint http://filosofia.dipafilo.unimi.it/~bonomi/Pragmatic.pdf.
Hanley, Richard. . As good as it gets: Lewis on truth in fiction. Australasian
Journal of Philosophy (). –. doi:./.
Inconsistencies in fictions and elsewhere are discussed in:
Varzi, Achille. . Inconsistency without contradiction. Notre Dame Journal of
Formal Logic (). –. doi:./ndjfl/.
Lewis, David. . Logic for equivocators. Noûs (). –. doi:./.
Reprinted in Lewis (: pp. –).
§. S R



Some other interesting work on stories and pictures and their content:
Ross, Jeff. . The semantics of media (Studies in Linguistics and Philosophy
(SLAP) ). Dordrecht: Kluwer.
Zucchi, Sandro. . Tense in fiction. In Carlo Cecchetto, Gennaro Chierchia
& Maria Teresa Guasti (eds.), Semantic interfaces: Reference, anaphora and
aspect, –. CSLI Publications. URL http://tinyurl.com/ulwxwg.
Blumson, Ben. . Pictures, perspective and possibility. Philosophical Studies
doi:./s---.
Astonishingly, Lewis’ doctrine of the reality of the plurality of possible worlds is
being paralleled (pun absolutely intended) by theoretical physicists in a number
of ways. There is a controversial “many worlds” interpretation of quantum
mechanics, for example. Other terms found are the “multiverse” and “parallel
universes”. See for starters, Kai’s blog entry on a popular book on the issue, http://
kaivonfintel.org////many-worlds/, MIT physics professor Max Tegmark’s
page on the topic, http://space.mit.edu/home/tegmark/crazy.html, and a Fresh
Air interview with physicist Brian Greene, who just wrote a book called The Hid-
den Reality: Parallel Universes and the Deep Laws of the Cosmos: http://www.npr.
org/////a-physicist-explains-why-parallel-universes-may-exist.
— T     —
 C T
P A
With the basic framework in place, we now proceed to analyze a number
of intensional constructions. We start with the basic possible worlds
semantics for propositional attitude ascriptions. We talk briefly about
the formal properties of accessibility relations.

. Hintikka’s Idea 

. Accessibility Relations 
.. Reflexivity 
.. *Transitivity 
.. *Symmetry 
. Supplemental Readings 

. Hintikka’s Idea

Expressions like believe, know, doubt, expect, regret, and so on are usually said According to Hintikka
(), the term -
to describe  , expressing relations between individuals   goes
(the attitude holder) and propositions (intensions of sentences). back to Russell ().
The simple idea is that George believes that Henry is a spy claims that George
believes of the proposition that Henry is a spy that it is true. Note that for the Of course, the possible worlds
semantics for propositional
attitude ascription to be true it does not have to hold that Henry is actually a attitudes was in place long
spy. But where — in which world(s) — does Henry have to be a spy for it be true before the extension to fic-
tion contexts was proposed.
that George believes that Henry is a spy? We might want to be inspired by the Our discussion here has in-
colloquial phrase “in the world according to George” and say that George believes verted the historical sequence
for pedagogical purposes.
that Henry is a spy is true iff in the world according to George’s beliefs, Henry is
a spy. We immediately recall from the previous chapter that we need to fix this
idea up by making space for multiple worlds compatible with George’s beliefs
and by tying the truth-conditions to contingent facts about the evaluation world.
That is, what George believes is different in different possible worlds.
The following lexical entry thus offers itself:
 P A C 



() JbelieveKw,g =
λphs,ti . λx. ∀w 0 compatible with x 0 s beliefs in w : p(w 0 ) = .

It is important to realize the What is going on in this semantics? We conceive of George’s beliefs as a state of
modesty of this semantics: we
are not trying to figure out
his mind about whose internal structure we will remain agnostic, a matter left
what belief systems are and to other cognitive scientists. What we require of it is that it embody opinions
particularly not what their
internal workings are like. about what the world he is located in looks like. In other words, if his beliefs
That is the job of psychologists are confronted with a particular possible world w 0 , they will determine whether
(and philosophers of mind,
perhaps). For our semantics, that world may or may not be the world as they think it is. What we are asking
we treat the belief system as a of George’s mental state is whether any state of affairs, any event, anything in
black box that determines for
each possible world whether it w 0 is in contradiction with anything that George believes. If not, then w 0 is
considers it possible that it is compatible with George’s beliefs. For all George believes, w 0 may well be the
the world it is located in.
world where he lives. Many worlds will pass this criterion, just consider as one
factor that George is unlikely to have any precise opinions about the number of
leaves on the tree in front of my house. George’s belief system determines a set
of worlds compatible with his beliefs: those worlds that are viable candidates for
being the actual world, as far as his belief system is concerned.
Now, George believes a proposition iff that proposition is true in all of the
worlds compatible with his beliefs. If there is just one world compatible with his
beliefs where the proposition is not true, that means that he considers it possible
that the proposition is not true. In such a case, we can’t say that he believes the
proposition. Here is the same story in the words of Hintikka (), the source
for this semantics for propositional attitudes:

My basic assumption (slightly simplified) is that an attribution of any

Jaakko Hintikka propositional attitude to the person in question involves a division
of all the possible worlds (. . . ) into two classes: into those possible
worlds which are in accordance with the attitude in question and
into those which are incompatible with it. The meaning of the
division in the case of such attitudes as knowledge, belief, memory,
perception, hope, wish, striving, desire, etc. is clear enough. For
instance, if what we are speaking of are (say) a’s memories, then
these possible worlds are all the possible worlds compatible with
everything he remembers. [. . . ]
How are these informal observations to be incorporated into
a more explicit semantical theory? According to what I have said,
understanding attributions of the propositional attitude in question
(. . . ) means being able to make a distinction between two kinds
of possible worlds, according to whether they are compatible with
the relevant attitudes of the person in question. The semantical
counterpart to this is of course a function which to a given individual
person assigns a set of possible worlds.
However, a minor complication is in order here. Of course,
§. H’ I



the person in question may himself have different attitudes in the

different worlds we are considering. Hence this function in effect
becomes a relation which to a given individual and to a given possible
world µ associates a number of possible worlds which we shall call
the  to µ. The relation will be called the alternativeness
relation. (For different propositional attitudes, we have to consider
different alternativeness relations.)
E .: Let’s adopt Hintikka’s idea that we can use a function that maps x
and w into the set of worlds w 0 compatible with what x believes in w. Call this
function B. That is,
() B = λx. λw. {w 0 : w 0 is compatible with what x believes in w}.

Using this notation, our lexical entry for believe would look as follows:
() JbelieveKw,g = λphs,ti . λx. B(x)(w) ⊆ p.

We are here indulging in the usual sloppiness in treating p both as a function

from worlds to truth-values and as the set characterized by that function.
Here now are two “alternatives” for the semantics of believe:
() A  ( )
JbelieveKw,g = λp ∈ Dhs,ti . λx ∈ D. p = B(x)(w) .
 

() A  (  )

JbelieveKw,g = λp ∈ Dhs,ti . λx ∈ D. p ∩ B(x)(w) 6= ∅ .
 

Explain why these do not adequately capture the meaning of believe. 

E .: Follow-up: The semantics in () would have made believe into
an existential quantifier of sorts: it would say that some of the worlds compatible
with what the subject believes are such-and-such. You have argued (successfully,
of course) that such an analysis is wrong for believe. But are there attitude
predicates with such an “existential” meaning? Discuss some candidates. If you
can’t find any candidates that survive scrutiny, can you speculate why there might
be no existential attitude predicates? [Warning: this is unexplored territory!]

We can also think of belief states as being represented by a function BS, which BS is meant to stand for
‘belief state’, not for what
maps an individual and a world into a set of propositions: those that the you might have thought!
individual believes. From there, we could calculate the set of worlds compatible
with an individual x’s beliefs in world w by retrieving the set of those possible
worlds in which all of the propositions in BS(x)(w) are true: {w 0 : ∀p ∈
BS(x)(w) : p(w 0 ) = }, which in set talk is simply the big intersection of all the
propositions in the set: ∩BS(x)(w). Our lexical entry then would be:
 P A C 



() JbelieveKw,g = λphs,ti . λx. ∩ BS(x)(w) ⊆ p.

E .: Imagine that our individual x forms a new opinion. Imagine
that we model this by adding a new proposition p to the pool of opinions. So,
BS(x)(w) now contains one further element. There are now more opinions.
What happens to the set of worlds compatible with x’s beliefs? Does it get bigger
or smaller? Is the new set a subset or superset of the previous set of compatible
worlds? 

. Accessibility Relations

Another way of reformulating Hintikka’s semantics for propositional attitudes
is via the notion of an  . We talk of a world w 0 being
accessible from w. Each attitude can be associated with such an accessibility
relation. For example, we can introduce the relation wRB a w which holds iff w
0 0

is compatible with a’s belief state in w. We have then yet another equivalent way
of specifying the lexical entry for believe:
() JbelieveKw,g = λphs,ti . λx. ∀w 0 : wRB 0 0
x w → p(w ) = .

It is profitable to think of different attitudes (belief, knowledge, hope, regret,

Kirill Shklovsky (in class) asked memory, . . . ) as corresponding to different accessibility relations. Recall now that
why we call reflexivity, transi-
tivity, and symmetry “formal”
the linguistic study of determiners benefitted quite a bit from an investigation of
properties of relations. The the formal properties of the relations between sets of individuals that determiners
idea is that certain properties
are “formal” or “logical”, while express. We can do the same thing here and ask about the formal properties
others are more substantial. So, of the accessibility relation associated with belief versus the one associated with
the fact that the relation “have
the same birthday as” is sym- knowledge, etc. The obvious properties to think about are reflexivity, transitivity,
metric seems a more formal and symmetry.
fact about it than the fact that
the relation holds between my
daughter and my brother-in-
law. Nevertheless, one of the
most common ways of charac-
.. Reflexivity
terizing formal/logical notions
(permutation-invariance, if A relation is reflexive iff for any object in the domain of the relation we know that
you’re curious) does not in the relation holds between that object and itself. Which accessibility relations
fact make symmetry etc. a
formal/logical notion. So, are reflexive? Take knowledge:
while intuitively these do seem
to be formal/logical proper- () wRK
x w iff w is compatible with what x knows in w.
0 0
ties, we do not know how to
substantiate that intuition.
See MacFarlane () for We are asking whether for any given possible world w, we know that RK x holds
discussion.
between w and w itself. It will hold if w is a world that is compatible with what
We talk here about knowledge we know in w. And clearly that must be so. Take our body of knowledge in w.
entailing (or even presup-
posing) truth but we do not
The concept of knowledge crucially contains the concept of truth: what we know
mean to say that knowledge must be true. So if in w we know that something is the case then it must be the
simply equals true belief. Pro-
fessors Socrates and Gettier
case in w. So, w must be compatible with all we know in w. RK x is reflexive.
and their exegetes have further
considerations.
§. A R



Now, if an attitude X corresponds to a reflexive accessibility relation, then we

can conclude from a Xs that p being true in w that p is true in w. This property In modal logic notation:
p → p. This pattern
of an attitude predicate is often called . It is to be distinguished is sometimes called T or
from , which is a property of attitudes which presuppose – rather than M, as is the correspond-
ing system of modal logic.
(merely) entail – the truth of their complement.
If we consider the relation RB x pairing with a world w those worlds w which
0

are compatible with what x believes in w, we no longer have reflexivity: belief is

not a veridical attitude. It is easy to have false beliefs, which means that the actual The difference between believe
and know in natural discourse
world is not in fact compatible with one’s beliefs, which contradicts reflexivity. is quite delicate, especially
And many other attitudes as well do not involve veridicality/reflexivity: what we when one considers first person
uses (I believe the earth is flat
hope may not come true, what we remember may not be what actually happened, vs. I know the earth is flat).
etc.
In modal logic, the correspondence between formal properties of the acces-
sibility relation and the validity of inference patterns is well-studied. What we
have just seen is that reflexivity of the accessibility relation corresponds to the
validity of p → p. Other properties correspond to other characteristic patterns.
Let’s see this for transitivity and symmetry.

.. *Transitivity
Transitivity of the accessibility relation corresponds to the inference p → p. Starred sections are optional.
The pattern seems not obviously wrong for knowledge: if one knows that p, In the literature on epistemic
modal logic, the pattern is
doesn’t one thereby know that one knows that p? But before we comment known as the KK T or
on that, let’s establish the formal correspondence between transitivity and that P I.
In general modal logic, it is
inference pattern. This needs to go in both directions. the characteristic axiom 
of the modal logic system
S, which is a system that
not p adds  to the previous axiom
M/T. Thus, S is the logic of
p accessibility relations that are
w3
both reflexive and transitive.
w2

w1

Figure .: Transitivity

What does it take for the pattern to be valid? Assume that p holds for an
arbitrary world w, i.e. that p is true in all worlds w 0 accessible from w. Now,
the inference is to the fact that p again holds in any world w 00 accessible from
any of those worlds w 0 accessible from w. But what would prevent p from being
false in some w 00 accessible from some w 0 accessible from w? That could only
be prevented from happening if we knew that w 00 itself is accessible from w as
 P A C 



well, because then we would know from the premiss that p is true in it (since p
is true in all worlds accessible from w). Ah, but w 00 (some world accessible from
a world w 0 accessible from w) is only guaranteed to be accessible from w if the
accessibility relation is transitive (if w 0 is accessible from w and w 00 is accessible
from w 0 , then transitivity ensures that w 00 is accessible from w). This reasoning
has shown that validity of the pattern requires transitivity. The other half of
proving the correspondence is to show that transitivity entails that the pattern is
valid.
The proof proceeds by reductio. Assume that the accessibility relation is
transitive. Assume that (i) p holds for some world w but that (ii) p doesn’t
hold in w. We will show that this situation cannot obtain. By (i), p is true in all
worlds w 0 accessible from w. By (ii), there is some non-p world w 00 accessible
from some world w 0 accessible from w. But by transitivity of the accessibility
relation, that non-p world w 00 must be accessible from w. And since all worlds
accessible from w are p worlds, w 00 must be a p world, in contradiction to (ii).
So, as soon as we assume transitivity, there is no way for the inference not to go
through.
Now, do any of the attitudes have the transitivity property? It seems rather
obvious that as soon as you believe something, you thereby believe that you
believe it (and so it seems that belief involves a transitive accessibility relation).
And in fact, as soon as you believe something, you believe that you know it. But
one might shy away from saying that knowing something automatically amounts
to knowing that you know it. For example, many are attracted to the idea that to
know something requires that (i) that it is true, (ii) that you believe it, and (iii)
that you are justified in believing it: the justified true belief analysis of knowledge.
So, now couldn’t it be that you know something, and thus (?) that you believe
you know it, and thus that you believe that you are justified in believing it,
but that you are not justified in believing that you are justified in believing it?
After all, one’s source of knowledge, one’s reliable means of acquiring knowledge,
might be a mechanism that one has no insight into. So, while one can implicitly
trust (believe) in its reliability, and while it is in fact reliable, one might not have
any means to have trustworthy beliefs about it. [Further worries about the KK
Thesis are discussed by Williamson ().]

.. *Symmetry
What would the consequences be if the accessibility relation were symmetric?
Symmetry of the accessibility relation R corresponds to the validity of the
following principle:
§. A R



() Brouwer’s Axiom: h In modal logic notation:

i p → ♦p, known simply as
∀p∀w : w ∈ p → ∀w wRw → ∃w [w Rw & w ∈ p]
0 0 00 0 00 00

B in modal logic. The system
that combines T/M with B is
often called Brouwer’s System
(B), after the mathematician
not p L.E.J. Brouwer, not because
he proposed it but because it
was thought that it had some
w2 w3 connections to his doctrines.

p
w1

Figure .: Symmetry

Here’s the reasoning: Assume that R is in fact symmetric. Pick a world w in

which p is true. Now, could it be that the right hand side of the inference fails L.E.J. Brouwer
to hold in w? Assume that it does fail. Then, there must be some world w 0
accessible from w in which ♦p is false. In other words, from that world w 0
there is no accessible world w 00 in which p is true. But since R is assumed to be
symmetric, one of the worlds accessible from w 0 is w and in w, p is true, which
contradicts the assumption that the inference doesn’t go through. So, symmetry
ensures the validity of the inference.
The other way (validity of the inference requires symmetry): the inference
says that from any p world we only have worlds accessible from which there is
at least one accessible p world. But imagine that p is true in w but not true in
any other world. So, the only way for the conclusion of the inference to hold
automatically is to have a guarantee that w (the only p world) is accessible from
any world accessible from it. That is, we need to have symmetry. QED.
To see whether a particular kind of attitude is based on a symmetric accessibil-
ity relation, we can ask whether Brouwer’s Axiom is intuitively valid with respect
to this attitude. If it is not valid, this shows that the accessibility relation can’t
be symmetric. In the case of a knowledge-based accessibility relation (epistemic
accessibility), one can argue that symmetry does not hold:
The symmetry condition would imply that if something happens
to be true in the actual world, then you know that it is compatible
with your knowledge (Brouwer’s Axiom). This will be violated by
any case in which your beliefs are consistent, but mistaken. Suppose
that while p is in fact true, you feel certain that it is false, and so
think that you know that it is false. Since you think you know this,
it is compatible with your knowledge that you know it. (Since we

 Thanks to Bob Stalnaker (pc to Kai von Fintel) for help with the following reasoning.
 P A C 



are assuming you are consistent, you can’t both believe that you
know it, and know that you do not). So it is compatible with your
knowledge that you know that not p. Equivalently : you don’t know
that you don’t know that not p. Equivalently: you don’t know that
it’s compatible with your knowledge that p. But by Brouwer’s Axiom,
since p is true, you would have to know that it’s compatible with
your knowledge that p. So if Brouwer’s Axiom held, there would
be a contradiction. So Brouwer’s Axiom doesn’t hold here, which
shows that epistemic accessibility is not symmetric.
Game theorists and theoretical computer scientists who traffic in logics of knowl-
edge often assume that the accessibility relation for knowledge is an equivalence
relation (reflexive, symmetric, and transitive). But this is appropriate only if
one abstracts away from any error, in effect assuming that belief and knowledge
All one really needs to make coincide. One striking consequence of working with an equivalence relation
NI valid is to have a E-
 accessibility relation:
as the accessibility relation for knowledge is that one predicts the principle of
any two worlds accessible from N I to hold:
the same world are accessible
from each other. It is a nice
little exercise to prove this, if () N I ()
you have become interested in If one doesn’t know that p, then one knows that one doesn’t know that
this sort of thing. Note that
all reflexive and Euclidean p. (¬p → ¬p).
accessibility relations are
transitive and symmetric as
well — another nice little thing
This surely seems rather dubious: imagine that one strongly believes that p but
to prove. that nevertheless p is false, then one doesn’t know that p, but one doesn’t seem
to believe that one doesn’t know that p, in fact one believes that one does know
that p.

. Supplemental Readings

A recent survey on attitudes:
Swanson, Eric. . Propositional attitudes. Ms, prepared for Semantics: An
international handbook of meaning, edited by Klaus von Heusinger, Claudia
Maienbon, and Paul Portner. URL http://tinyurl.com/swanson-attitudes.
Further connections between mathematical properties of accessibility relations
and logical properties of various notions of necessity and possibility are studied
extensively in modal logic:
Hughes, G.E. & M.J. Cresswell. . A new introduction to modal logic. London:
Routledge.

 This and the following step rely on the duality of necessity and possibility: q is compatible
with your knowledge iff you don’t know that not q.
§. S R



Garson, James. . Modal logic. In Edward N. Zalta (ed.), The Stanford
encyclopedia of philosophy, URL http://plato.stanford.edu/entries/logic-modal/,
especially section  and , “Modal Axioms and Conditions on Frames”, “Map
of the Relationships between Modal Logics”.
A thorough discussion of the possible worlds theory of attitudes, and some of its
potential shortcomings, can be found in Bob Stalnaker’s work:
Stalnaker, Robert. . Inquiry. MIT Press.
Stalnaker, Robert. . Context and content. Oxford: Oxford University Press.
A quick and informative surveys about the notion of knowledge:
Steup, Matthias. . The analysis of knowledge. In Edward N. Zalta (ed.), The
Stanford encyclopedia of philosophy, Fall  edn. URL http://plato.stanford.
edu/archives/fall/entries/knowledge-analysis/.
Linguistic work on attitudes has often been concerned with various co-occurrence
patterns, particularly which moods (indicative or subjunctive or infinitive) occur
in the complement and whether negative polarity items are licensed in the
complement.
Mood licensing:
Portner, Paul. . The semantics of mood, complementation, and conversa-
tional force. Natural Language Semantics (). –. doi:./A:.
NPI-Licensing:
Kadmon, Nirit & Fred Landman. . Any. Linguistics and Philosophy ().
–. doi:./BF.
von Fintel, Kai. . NPI licensing, Strawson entailment, and context depen-
dency. Journal of Semantics (). –. doi:./jos/...
Giannakidou, Anastasia. . Affective dependencies. Linguistics and Philosophy
(). –. doi:./A:.
There is some interesting work out of Amherst rethinking the way attitude
predicates take their complements:
Kratzer, Angelika. . Decomposing attitude verbs. Handout from a talk
honoring Anita Mittwoch on her th birthday at the Hebrew Univer-
sity of Jerusalem July , . URL http://semanticsarchive.net/Archive/
DcwYJkM/attitude-verbs.pdf.
Moulton, Keir. . Clausal complementation and the Wager-class. Proceed-
ings of the North East Linguistics Society . URL http://sites.google.com/
site/keirmoulton/Moultonnelswager.pdf. http://people.umass.edu/keir/
Wager.pdf.
 P A C 



Moulton, Keir. . Natural selection and the syntax of clausal complementation:
University of Massachusetts at Amherst dissertation. URL http://scholarworks.
umass.edu/open_access_dissertations//.
Tamina Stephenson in her MIT dissertation and related work explores the way
attitude predicates interact with epistemic modals and taste predicates in their
complements:
Stephenson, Tamina. a. Judge dependence, epistemic modals, and predicates
of personal taste. Linguistics and Philosophy (). –. doi:./s-
--.
Stephenson, Tamina. b. Towards a theory of subjective meaning: Mas-
sachusetts Institute of Technology dissertation. URL http://semanticsarchive.
net/Archive/QxMjkO/Stephenson--thesis.pdf.
Jon Gajewski in his MIT dissertation and subsequent work explores the distribu-
tion of the - property among attitude predicates and traces it back to
presuppositional components of the meaning of the predicates:
Gajewski, Jon. . Neg-raising: Polarity and presupposition: Massachusetts
Institute of Technology dissertation. doi:./.
Gajewski, Jon. . Neg-raising and polarity. Linguistics and Philosophy
doi:./s---z.
Interesting work has also been done on presupposition projection in attitude
contexts:
Asher, Nicholas. . A typology for attitude verbs and their anaphoric proper-
ties. Linguistics and Philosophy (). –. doi:./BF.
Heim, Irene. . Presupposition projection and the semantics of attitude verbs.
Journal of Semantics (). –. doi:./jos/...
Geurts, Bart. . Presuppositions and anaphors in attitude contexts. Linguistics
and Philosophy (). –. doi:./A:.
 C T
M

We turn to modal auxiliaries and related constructions. The main

difference from attitude constructions is that their semantics is more
context-dependent. Otherwise, we are still quantifying over possible
worlds.

. The Quantificational Theory of Modality 

.. Syntactic Assumptions 
.. Quantification over Possible Worlds 
. Flavors of Modality 
.. Contingency 
.. Epistemic vs. Circumstantial Modality 
.. Contingency Again 
.. Iteration 
.. A technical variant of the analysis 
. *Kratzer’s Conversational Backgrounds 
. Supplementary Readings 

. The Quantificational Theory of Modality

We will now be looking at modal auxiliaries like may, must, can, have to, etc.
Most of what we say here should carry over straightforwardly to modal adverbs
like maybe, possibly, certainly, etc. We will make certain syntactic assumptions,
which make our work easier but which leave aside many questions that at some
point deserve to be addressed.
 M C 



.. Syntactic Assumptions

We will assume, at least for the time being, that a modal like may is a 
predicate (rather than a  predicate), i.e., its subject is not its own argu-
The issue of raising vs. control ment, but has been moved from the subject-position of its infinitival complement.
will be taken up later. If you
are eager to get started on it
So, we are dealing with the following kind of structure:
and other questions of the
morphosyntax of modals, read () a. Ann may be smart.
the handout from an LSA class
Sabine and Kai taught last b. [ Ann [ λ [ may [ t be smart ]]]]
summer: http://web.mit.edu/
fintel/lsa-class--handout. Actually, we will be working here with the even simpler structure below, in
pdf.
which the subject has been reconstructed to its lowest trace position. (E.g.,
these could be generated by deleting all but the lowest copy in the movement
We will talk about reconstruc- chain.) We will be able to prove that movement of a name or pronoun never
tion in more detail later.
affects truth-conditions, so at any rate the interpretation of the structure in (b)
would be the same as of (). As a matter of convenience, then, we will take
the reconstructed structures, which allow us to abstract away from the (here
irrelevant) mechanics of variable binding.
() may [ Ann be smart ]
So, for now at least, we are assuming that modals are expressions that take a full
sentence as their semantic argument. Now then, what do modals mean?

.. Quantification over Possible Worlds

This idea goes back a long The basic idea of the possible worlds semantics for modal expressions is that they
time. It was famously held
by Leibniz, but there are
are quantifiers over possible worlds. Toy lexical entries for must and may, for
precedents in the medieval example, would look like this:
literature, see Knuuttila ().
See Copeland () for the
modern history of the possible () JmustKw,g = λphs,ti . ∀w 0 : p(w 0 ) = .
worlds analysis of modal
expressions. () JmayKw,g = λphs,ti . ∃w 0 : p(w 0 ) = .

This analysis is too crude (in particular, notice that it would make modal sen-
tences non-contingent — there is no occurrence of the evaluation world on the
right hand side!). But it does already have some desirable consequences that we
will seek to preserve through all subsequent refinements. It correctly predicts a
number of intuitive judgments about the logical relations between must and may
and among various combinations of these items and negations. To start with
some elementary facts, we feel that must φ entails may φ, but not vice versa:

 We will assume that even though Ann be smart is a non-finite sentence, this will not have any
effect on its semantic type, which is that of a sentence, which in turn means that its semantic
value is a truth-value. This is hopefully independent of the (interesting) fact that Ann be smart
on its own cannot be used to make a truth-evaluable assertion.
§. T Q T  M



() You must stay.

Therefore, you may stay. 
() You may stay.
Therefore, you must stay. 
() a. You may stay, but it is not the case that you must stay.
b. You may stay, but you don’t have to stay. 

We judge must φ incompatible with its “inner negation” must [not φ ], but find
may φ and may [not φ ] entirely compatible:

() You must stay, and/but also, you must leave. (leave = not stay).

() You may stay, but also, you may leave.


We also judge that in each pair below, the (a)-sentence and the (b)-sentences say
the same thing.

() a. You must stay.

b. It is not the case that you may leave.
You aren’t allowed to leave.
(You may not leave.)
(You can’t leave.)
() a. You may stay.
b. It is not the case that you must leave.
You don’t have to leave.
You don’t need to leave.
(You needn’t leave.)

Given that stay and leave are each other’s negations (i.e. JleaveKw,g = Jnot stayKw,g ,
and JstayKw,g = Jnot leaveKw,g ), the LF-structures of these equivalent pairs of

 The somewhat stilted it is not the case-construction is used in to make certain that negation
takes scope over must. When modal auxiliaries and negation are together in the auxiliary complex
of the same clause, their relative scope seems not to be transparently encoded in the surface
order; specifically, the scope order is not reliably negation  modal. (Think about examples
with mustn’t, can’t, shouldn’t, may not etc. What’s going on here? This is an interesting topic
which we must set aside for now. See the references at the end of the chapter for relevant work.)
With modal main verbs (such as have to), this complication doesn’t arise; they are consistently
inside the scope of clause-mate auxiliary negation. Therefore we can use (b) to (unambiguously)
express the same scope order as (a), without having to resort to a biclausal structure.
 The parenthesized variants of the (b)-sentences are pertinent here only to the extent that we
can be certain that negation scopes over the modal. In these examples, apparently it does, but as
we remarked above, this cannot be taken for granted in all structures of this form.
 M C 



sentences can be seen to instantiate the following schemata:

() a. must φ ≡ not [may [not φ]]
b. must [not ψ] ≡ not [may ψ]
() a. may φ ≡ not [must [not φ]]
b. may [not ψ] ≡ not [must ψ]
Our present analysis of must, have-to, . . . as universal quantifiers and of may, can,
. . . as existential quantifiers straightforwardly predicts all of the above judgments,
More linguistic data regarding as you can easily prove.
the “parallel logic” of modals
and quantifiers can be found
in Larry Horn’s dissertation () a. ∀xφ ≡ ¬∃¬φ
(Horn ). b. ∀x¬φ ≡ ¬∃xφ
() a. ∃xφ ≡ ¬∀x¬φ
b. ∃x¬φ ≡ ¬∀xφ

. Flavors of Modality

.. Contingency
We already said that the semantics we started with is too simple-minded. In
particular, we have no dependency on the evaluation world, which would make
modal statements non-contingent. This is not correct.
If one says It may be snowing in Cambridge, that may well be part of useful,
practical advice about what to wear on your upcoming trip to Cambridge. It may
be true or it may be false. The sentence seems true if said in the dead of winter
when we have already heard about a Nor’Easter that is sweeping across New
England. The sentence seems false if said by a clueless Australian acquaintance
of ours in July.
The contingency of modal claims is not captured by our current semantics.
All the may-sentence would claim under that semantics is that there is some
possible world where it is snowing in Cambridge. And surely, once you have
read Lewis’ quote in Chapter , where he asserts the existence of possible worlds
with different physical constants than we enjoy here, you must admit that there
have to be such worlds even if it is July. The problem is that in our semantics,
repeated here
() JmayKw,g = λphs,ti . ∃w 0 : p(w 0 ) = .

there is no occurrence of w on the right hand side. This means that the truth-
conditions for may-sentences are world-independent. In other words, they make

 In logicians’ jargon, must and may behave as  of each other. For definitions of “dual”,
see Barwise & Cooper (: ) or Gamut (: vol.,).
§. F  M



non-contingent claims that are either true whatever or false whatever, and because
of the plenitude of possible worlds they are more likely to be true than false.
This needs to be fixed. But how?
Well, what makes it may be snowing in Cambridge seem true when we know
about a Nor’Easter over New England? What makes it seem false when we
know that it is summer in New England? The idea is that we only consider
possible worlds       . And since
what evidence is available to us differs from world to world, so will the truth of a
may-statement.
() JmayKw,g = λp. ∃w 0 compatible with the evidence in w : p(w 0 ) = .
() JmustKw,g = λp. ∀w 0 compatible with the evidence in w : p(w 0 ) = .

Let us consider a different example:

() You have to be quiet.
Imagine this sentence being said based on the house rules of the particular
dormitory you live in. Again, this is a sentence that could be true or could
be false. Why do we feel that this is a contingent assertion? Well, the house
rules can be different from one world to the next, and so we might be unsure
or mistaken about what they are. In one possible world, they say that all noise
must stop at pm, in another world they say that all noise must stop at pm.
Suppose we know that it is : now, and that the dorm we are in has either
one or the other of these two rules, but we have forgotten which. Then, for all
we know, you have to be quiet may be true or it may be false. This suggests a
lexical entry along these lines:
() Jhave-toKw,g = λp. ∀w 0 compatible with the rules in w : p(w 0 ) = .

Again, we are tying the modal statement about other worlds down to certain
worlds that stand in a certain relation to actual world: those worlds where the
rules as they are here are obeyed.
A note of caution: it is very important to realize that the worlds compatible
with the rules as they are in w are those worlds where nothing happens that
violates any of the w-rules. This is not at all the same as saying that the worlds
compatible with the rules in w are those worlds where the same rules are in
force. Usually, the rules do not care what the rules are, unless the rules contain
some kind of meta-statement to the effect that the rules have to be the way they
are, i.e. that the rules cannot be changed. So, in fact, a world w 0 in which
nothing happens that violates the rules as they are in w but where the rules are
quite different and in fact what happens violates the rules as they are in w 0 is

 From now on, we will leave off type-specifications such as that p has to be of type hs, ti,
whenever it is obvious what they should be and when saving space is aesthetically called for.
 M C 



nevertheless a world compatible with the rules in w. For example, imagine that
the only relevant rule in w is that students go to bed before midnight. Take a
world w 0 where a particular student goes to bed at : pm but where the rules
are different and say that students have to go to bed before  pm. Such a world
w 0 is compatible with the rules in w (but of course not with the rules in w 0 ).
Apparently, there are different flavors of modality, varying in what kind of
facts in the evaluation world they are sensitive to. The semantics we gave for
must and may above makes them talk about evidence, while the semantics we
gave for have-to made it talk about rules. But that was just because the examples
were hand-picked. In fact, in the dorm scenario we could just as well have said
You must be quiet. And, vice versa, there is nothing wrong with using it has to
be snowing in Cambridge based on the evidence we have. In fact, many modal
expressions seem to be multiply ambiguous.
Traditional descriptions of modals often distinguish a number of “readings”:
, , , , , . . . . (Beyond “epis-
temic” and “deontic,” there is a great deal of terminological variety. Sometimes
all non-epistemic readings are grouped together under the term  .)
Here are some initial illustrations.

() E M

A: Where is John?
B: I don’t know. He may be at home.
() D M
A: Am I allowed to stay over at Janet’s house?
B: No, but you may bring her here for dinner.
() C/D M
A: I will plant the rhododendron here.
B: That’s not a good idea. It can grow very tall.

How are may and can interpreted in each of these examples? What do the
interpretations have in common, and where do they differ?
In all three examples, the modal makes an existentially quantified claim about
possible worlds. This is usually called the   of the claim. What
differs is what worlds are quantified over. In  modal sentences, we
quantify over worlds compatible with the available evidence. In  modal
sentences, we quantify over worlds compatible with the rules and/or regulations.
And in the  modal sentence, we quantify over the set of worlds
which conform to the laws of nature (in particular, plant biology). What speaker
B in () is saying, then, is that there are some worlds conforming to the laws of
nature in which this rhododendron grows very tall. (Or is this another instance
of an epistemic reading? See below for discussion of the distinction between
§. F  M



circumstantial readings and epistemic ones.)

How can we account for this variety of readings? One way would be to
write a host of lexical entries, basically treating this as a kind of (more or less
principled) ambiguity. Another way, which is preferred by many people, is to
treat this as a case of context-dependency, as argued in seminal work by Kratzer
(, , , ). Angelika Kratzer
According to Kratzer, what a modal brings with it intrinsically is just a It is well-known that natural
modal force, that is, whether it is an existential (possibility) modal or a universal language quantification is in
general subject to contextual re-
(necessity) modal. What worlds it quantifies over is determined by context. In striction. See Stanley & Szabó
essence, the context has to supply a restriction to the quantifier. How can we () for a recent discussion.

implement this idea?

We encountered context-dependency before when we talked about pronouns
and their referential (and E-Type) readings (H& K, chapters –). We treated
referential pronouns as free variables, appealing to a general principle that free
variables in an LF need to be supplied with values from the utterance context. If
we want to describe the context-dependency of modals in a technically analogous
fashion, we can think of their LF-representations as incorporating or subcate-
gorizing for a kind of invisible pronoun, a free variable that stands for a set of
possible worlds. So we posit LF-structures like this:
() [ I 0 [ I must phs,ti ] [ VP you quiet]]
phs,ti here is a variable over (characteristic functions of ) sets of worlds, which — like
all free variables — needs to receive a value from the utterance context. Possible
values include: the set of worlds compatible with the speaker’s current knowl-
edge; the set of worlds in which everyone obeys all the house rules of a certain
dormitory; and many others. The denotation of the modal itself now has to be
of type hst, hst, tii rather than hst, ti, thus it will be more like a quantificational
determiner rather than a complete generalized quantifier. Only after the modal
has been combined with its covert restrictor do we obtain a value of type hst, ti.
() a. JmustKw,g = Jhave-toKw,g = Jneed-toKw,g = . . . =
λp ∈ Dhs,ti . λq ∈ Dhs,ti . ∀w ∈ W [p(w) =  → q(w) = ] in set talk: p ⊆ q
b. JmayKw,g = JcanKw,g = Jbe-allowed-toKw,g = . . . =
λp ∈ Dhs,ti . λq ∈ Dhs,ti . ∃w ∈ W [p(w) =  & q(w) = ] in set talk: p ∩ q 6= ∅

On this approach, the epistemic, deontic, etc. “readings” of individual occur-

rences of modal verbs come about by a combination of two separate things.
The lexical semantics of the modal itself encodes just a quantificational force,
a relation between sets of worlds. This is either the subset-relation (universal
quantification; necessity) or the relation of non-disjointness (existential quantifi-
cation; possibility). The covert variable next to the modal picks up a contextually
salient set of worlds, and this functions as the quantifier’s restrictor. The labels
“epistemic”, “deontic”, “circumstantial” etc. group together certain conceptually
 M C 



natural classes of possible values for this covert restrictor.

Notice that, strictly speaking, there is not just one deontic reading (for
example), but many. A speaker who utters
() You have to be quiet.
might mean: ‘I want you to be quiet,’ (i.e., you are quiet in all those worlds
that conform to my preferences). Or she might mean: ‘unless you are quiet, you
won’t succeed in what you are trying to do,’ (i.e., you are quiet in all those worlds
in which you succeed at your current task). Or she might mean: ‘the house rules
of this dormitory here demand that you be quiet,’ (i.e., you are quiet in all those
worlds in which the house rules aren’t violated). And so on. So the label “deontic”
appears to cover a whole open-ended set of imaginable “readings”, and which
one is intended and understood on a particular utterance occasion may depend
on all sorts of things in the interlocutors’ previous conversation and tacit shared
assumptions. (And the same goes for the other traditional labels.)

.. Epistemic vs. Circumstantial Modality

Is it all context-dependency? Or do flavors of modality correspond to some sorts
of signals in the structure of sentences? Read the following famous passage from
Kratzer and think about how the two sentences with their very different modal
meanings differ in structure:
Quoted from Kratzer (). In Consider sentences () and ():
Kratzer (), the hydrangeas
were Zwetschgenbäume ‘plum
trees’. The German word
() Hydrangeas can grow here.
Zwetschge, by the way, is
etymologically derived from
() There might be hydrangeas growing here.
the name of the city Damascus
(Syria), the center of the The two sentences differ in meaning in a way which is illustrated by
ancient plum trade.
the following scenario.
“Hydrangeas”
Suppose I acquire a piece of land in a far away country and
discover that soil and climate are very much like at home, where
hydrangeas prosper everywhere. Since hydrangeas are my favorite
plants, I wonder whether they would grow in this place and inquire
about it. The answer is (). In such a situation, the proposition
expressed by () is true. It is true regardless of whether it is or
isn’t likely that there are already hydrangeas in the country we are
considering. All that matters is climate, soil, the special properties
of hydrangeas, and the like. Suppose now that the country we are in
has never had any contacts whatsoever with Asia or America, and
the vegetation is altogether different from ours. Given this evidence,
my utterance of () would express a false proposition. What counts
§. F  M



here is the complete evidence available. And this evidence is not

compatible with the existence of hydrangeas.
() together with our scenario illustrates the pure -
 reading of the modal can. [. . . ]. () together with our scenario
illustrates the epistemic reading of modals. [. . . ] circumstantial
and epistemic conversational backgrounds involve different kinds
of facts. In using an epistemic modal, we are interested in what else
may or must be the case in our world given all the evidence available.
Using a circumstantial modal, we are interested in the necessities
implied by or the possibilities opened up by certain sorts of facts.
Epistemic modality is the modality of curious people like historians,
detectives, and futurologists. Circumstantial modality is the modal-
ity of rational agents like gardeners, architects, and engineers. A
historian asks what might have been the case, given all the available
facts. An engineer asks what can be done given certain relevant facts.
Consider also the very different prominent meanings of the following two
sentences, taken from Kratzer as well:
() a. Cathy can make a pound of cheese out of this can of milk.
b. Cathy might make a pound of cheese out of this can of milk.
E .: Come up with examples of epistemic, deontic, and circumstantial
uses of the necessity verb have to. Describe the set of worlds that constitutes the
understood restrictor in each of your examples. 

.. Contingency Again

We messed up. If you inspect the context-dependent meanings we have on the
table now for our modals, you will see that the right hand sides again do not
mention the evaluation world w. Therefore, we will again have the problem of
not making contingent claims, indirectly about the actual world. This needs to
be fixed. We need a semantics that is both context-dependent and contingent.
The problem, it turns out, is with the idea that the utterance context supplies
a determinate set of worlds as the restrictor. When I understand that you meant
your use of must, in you must be quiet, to quantify over the set of worlds in which
the house rules of our dorm are obeyed, this does not imply that you and I have
to know or agree on which set exactly this is. That depends on what the house
rules in our world actually happen to say, and this may be an open question at
the current stage of our conversation. What we do agree on, if I have understood
your use of must in the way that you intended it, is just that it quantifies over
whatever set of worlds it may be that the house rules pick out.
 M C 



You will of course recognize The technical implementation of this insight requires that we think of the
that functions of type hs, sti
are simply a schönfinkeled
context’s contribution not as a set of worlds, but rather as a function which for
version of the  each world it applies to picks out such a set. For example, it may be the function
 we introduced in
the previous chapter. which, for any world w, yields the set {w 0 : the house rules that are in force in w
are obeyed in w 0 }. If we apply this function to a world w , in which the house
rules read “no noise after  pm”, it will yield a set of worlds in which nobody
makes noise after  pm. If we apply the same function to a world w , in which
the house rules read “no noise after  pm”, it will yield a set of worlds in which
nobody makes noise after  pm.
Suppose, then, that the covert restrictor of a modal predicate denotes such a
function, i.e., its value is of type hs, sti.
() [ I’ [ I must Rhs,sti ] [ VP you quiet]]
And the new lexical entries for must and may that will fit this new structure are
these:
() a. JmustKw,g = Jhave-toKw,g = Jneed-toKw,g = . . . =
in set talk: (R(w) ⊆ q λR ∈ Dhs,sti . λq ∈ Dhs,ti . ∀w 0 ∈ W [R(w)(w 0 ) =  → q(w 0 ) = ]
b. JmayKw,g = JcanKw,g = Jbe-allowed-toKw,g = . . . =
in set talk: (R(w) ∩ q 6= ∅ λR ∈ Dhs,sti . λq ∈ Dhs,ti . ∃w 0 ∈ W [R(w)(w 0 ) =  & q(w 0 ) = ]

Let us see now how this solves the contingency problem.

() Let w be a world, and assume that the context supplies an assignment g
such that g(R) = λw. λw 0 . the house rules in force in w are obeyed in
w0

Jmust R you quietKw,g = (IFA)

Jmust RKw,g (λw 0 Jyou quietKw ) =
0
(FA)
JmustKw,g (JRKw,g ) (λw 0 Jyou quietKw ) =
0
(lex. entries you, quiet)
JmustKw,g (JRKw,g ) (λw 0 . you are quiet in w 0 ) = (lex. entry must)
∀w 0 ∈ W : JRKw,g (w)(w 0 ) =  → you are quiet in w 0 = (pronoun rule)
∀w ∈ W : g(R)(w)(w ) =  → you are quiet in w =
0 0 0
(def. of g)
∀w 0 ∈ W [the house rules in force in w are obeyed in w 0
→ you are quiet in w 0 ]

As we see in the last line of (), the truth-value of () depends on the evaluation
world w.
E .: Describe two worlds w and w so that
Jmust R you quietKw ,g =  and Jmust R you quietKw ,g = . 

E .: In analogy to the deontic relation g(R) defined in (), define an
appropriate relation that yields an epistemic reading for a sentence like You may
be quiet. 
§. F  M



.. Iteration
Consider the following example:
() You might have to leave.
What does this mean? Under one natural interpretation, we learn that the speaker
considers it possible that the addressee is under the obligation to leave. This
seems to involve one modal embedded under a higher modal. It appears that
this sentence should be true in a world w iff some world w 0 compatible with
what the speaker knows in w is such that every world w 00 in which the rules as
they are in w 0 are followed is such that you leave in w 00 .
Assume the following LF: There is more to be said about
which modals can embed
under which other modals.
() [ I 0 [ might R ] [ VP [ have-to R ] [ IP you leave]]] See for some discussion the
handout mentioned earlier:
Suppose w is the world for which we calculate the truth-value of the whole http://web.mit.edu/fintel/
lsa-class--handout.pdf.
sentence, and the context maps R to the function which maps w to the set of all
those worlds compatible with what is known in w. might says that some of those
worlds are worlds w 0 that make the tree below might true. Now assume further
that the context maps R to the function which assigns to any such world w 0 the
set of all those worlds in which the rules as they are in w 0 are followed. have to
says that all of those worlds are worlds w 00 in which you leave.
In other words, while it is not known to be the case that you have to leave,
for all the speaker knows it might be the case.
E .: Describe values for the covert hs, sti-variable that are intuitively
suitable for the interpretation of the modals in the following sentences:
() As far as John’s preferences are concerned, you may stay with us.
() According to the guidelines of the graduate school, every PhD candidate
must take  credit hours outside his/her department.
() John can run a mile in  minutes.
() This has to be the White House.
() This elevator can carry up to  pounds.
For some of the sentences, different interpretations are conceivable depending
on the circumstances in which they are uttered. You may therefore have to
sketch the utterance context you have in mind before describing the accessibility
relation. 
E .: Collect two naturally occurring examples of modalized sentences
(e.g., sentences that you overhear in conversation, or read in a newspaper or novel
– not ones that are being used as examples in a linguistics or philosophy paper!),
and give definitions of values for the covert hs, sti-variable which account for the
 M C 



way in which you actually understood these sentences when you encountered
them. (If the appropriate interpretation is not salient for the sentence out of
context, include information about the relevant preceding text or non-linguistic
background.) 

.. A technical variant of the analysis

In our account of the contingency of modalized sentences, we adopted lexi-
cal entries for the modals that gave them world-dependent extensions of type
hhs, sti, hst, tii:

() (repeated from earlier):

For any w ∈ W : JmustKw,g
λR ∈ Dhs,sti . λq ∈ Dhs,ti . ∀w 0 ∈ W [R(w)(w 0 ) =  → q(w 0 ) = ]
(in set talk: λRhs,sti . λqhs,ti . (R(w) ⊆ q)).
Unfortunately, this treatment somewhat obscures the parallel between the modals
and the quantificational determiners, which have world-independent extensions
of type het, het, tii.
Let’s explore an alternative solution to the contingency problem, which will
allow us to stick with the world-independent type-hst, hst, tii-extensions that
we assumed for the modals at first:
() (repeated from even earlier):
JmustKw,g = λp ∈ Dhs,ti . λq ∈ Dhs,ti . ∀w ∈ W [p(w) =  → q(w) = ]
(in set talk: λp ∈ Dhs,ti . λq ∈ Dhs,ti . p ⊆ q).
We posit the following LF-representation:
() [ I 0 [ I must [ Rh,hs,stii w*]] [ VP you quiet]]
What is new here is that the covert restrictor is complex. The first part, Rh,hs,stii ,
is (as before) a free variable of type hs, sti, which gets assigned an accessibility
relation by the context of utterance. The second part is a special terminal symbol
which is interpreted as picking out the evaluation world:
() For any w ∈ W : Jw∗Kw,g = w.
When Rh,hs,stii and w* combine (by Functional Application), we obtain a
constituent whose extension is of type hs, ti (a proposition or set of worlds). This
is the same type as the extension of the free variable p in the previous proposal,
hence suitable to combine with the old entry for must (by FA). However, while
the extension of p was completely fixed by the variable assignment, and did not

 Dowty () introduced an analogous symbol to pick out the evaluation time. We have
chosen the star-notation to allude to this precedent.
§. *K’ C B



vary with the evaluation world, the new complex constituent’s extension depends
on both the assignment and the world:
() For any w ∈ W and any assignment g:
JRh,hs,stii (w*)Kw,g = g(h, hs, stii)(w).

As a consequence of this, the extensions of the higher nodes I and I 0 will also
vary with the evaluation world, and this is how we capture the fact that () is
contingent.
Maybe this variant is more appealing. But for the rest of this chapter, we
continue to assume the original analysis as presented earlier. In the next chapter
on conditionals, we will however make crucial use of this way of formulating the
semantics for modals. So, make sure you understand what we just proposed.

. *Kratzer’s Conversational Backgrounds

Angelika Kratzer has some interesting ideas on how accessibility relations are
supplied by the context. She argues that what is really floating around in a
discourse is a  . Accessibility relations can be
computed from conversational backgrounds (as we shall do here), or one can
state the semantics of modals directly in terms of conversational backgrounds (as
Kratzer does).
A conversational background is the sort of thing that is identified by phrases
like what the law provides, what we know, etc. Take the phrase what the law
provides. What the law provides is different from one possible world to another.
And what the law provides in a particular world is a set of propositions. Likewise,
what we know differs from world to world. And what we know in a particular
world is a set of propositions. The intension of what the law provides is then that
function which assigns to every possible world the set of propositions p such that
the law provides in that world that p. Of course, that doesn’t mean that p holds in
that world itself: the law can be broken. And the intension of what we know will
be that function which assigns to every possible world the set of propositions we
know in that world. Quite generally, conversational backgrounds are functions
of type hs, hst, tii, functions from worlds to (characteristic functions of ) sets of
propositions.
Now, consider:
() (In view of what we know,) Brown must have murdered Smith.
The in view of -phrase may explicitly signal the intended conversational back-
ground. Or, if the phrase is omitted, we can just infer from other clues in the
discourse that such an epistemic conversational background is intended. We will
focus on the case of pure context-dependency.
 M C 



How do we get from a conversational background to an accessibility relation?

Take the conversational background at work in (). It will be the following:
() λw. λp. p is one of the propositions that we know in w.

This conversational background will assign to any world w the set of propositions
p that in w are known by us. So we have a set of propositions. From that we can
get the set of worlds in which all of the propositions in this set are true. These
are the worlds that are compatible with everything we know. So, this is how we
get an accessibility relation:
() For any conversational background f of type hs, hst, tii, we define the
corresponding accessibility relation Rf of type hs, sti as follows:
Rf := λw. λw 0 . ∀p [f(w)(p) =  → p(w 0 ) = ].

In words, w 0 is f-accessible from w iff all propositions p that are assigned by f to

w are true in w 0 .
Kratzer uses the term   for the conversational background that
determines the set of accessible worlds. We can be sloppy and use this term for a
number of interrelated concepts:

(i) the conversational background (type hs, hst, tii),

(ii) the set of propositions assigned by the conversational background to a
particular world (type hst, ti),
(iii) the accessibility relation (type hs, sti) determined by (i),
(iv) the set of worlds accessible from a particular world (type hs, ti).
Kratzer calls a conversational background (modal base)  iff it assigns to
any world a set of propositions that are all true in that world. The modal base
what we know is realistic, the modal bases what we believe and what we want are
not.
What follows are some (increasingly technical exercises) on conversational back-
grounds.

E .: Show that a conversational background f is realistic iff the corre-
sponding accessibility relation Rf (defined as in ()) is reflexive. 

E .: Let us call an accessibility relation  if it makes every world
accessible from every world. R is  iff ∀w∀w 0 : w 0 ∈ R(w). What would
the conversational background f have to be like for the accessibility relation Rf
to be trivial in this sense? 

E .: The definition in () specifies, in effect, a function from Dhs,hst,tii
to Dhs,sti . It maps each function f of type hs, hst, tii to a unique function Rf of
§. *K’ C B



type hs, sti. This mapping is not one-to-one, however. Different elements of
Dhs,hst,tii may be mapped to the same value in Dhs,sti .
• Prove this claim. I.e., give an example of two functions f and f ’ in Dhs,hst,tii
for which () determines Rf = Rf 0 .
• As you have just proved, if every function of type hs, hst, tii qualifies as
a ‘conversational background’, then two different conversational back-
grounds can collapse into the same accessibility relation. Conceivably,
however, if we imposed further restrictions on conversational backgrounds
(i.e., conditions by which only a proper subset of the functions in Dhs,hst,tii
would qualify as conversational backgrounds), then the mapping between
conversational backgrounds and accessibility relations might become one-
to-one after all. In this light, consider the following potential restriction:
() Every conversational background f must be “closed under entail-
ment”; i.e., it must meet this condition:
∀w.∀p [∩f(w) ⊆ p → p ∈ f(w)].

(In words: if the propositions in f(w) taken together entail p, then p

must itself be in f(w).) Show that this restriction would ensure that the
mapping defined in () will be one-to-one. 

 In this exercise, we systematically substitute sets for their characteristic functions. I.e., we
pretend that Dhs,ti is the power set of W (i.e., elements of Dhs,ti are sets of worlds), and Dhst,ti is
the power set of Dhs,ti (i.e., elements of Dhst,ti are sets of sets of worlds). On these assumptions,
the definition in () can take the following form:

(i) For any conversational background f of type hs, hst, tii,

we define the corresponding accessibility relation Rf of type hs, sti as follows:
Rf := λw. {w 0 : ∀p [p ∈ f(w) → w 0 ∈ p]}.

The last line of this can be further abbreviated to:

(ii) Rf := λw. ∩ f(w)

This formulation exploits a set-theoretic notation which we have also used in condition () of
the second part of the exercise. It is defined as follows:

(iii) If S is a set of sets, then ∩S := {x : ∀Y [Y ∈ S → x ∈ Y]}.

 M C 



. Supplementary Readings

The most important background readings for this chapter are the following two
papers by Kratzer:
Kratzer, Angelika. . The notional category of modality. In Hans-Jürgen
Eikmeyer & Hannes Rieser (eds.), Words, worlds, and contexts: New approaches
in word semantics (Research in Text Theory ), –. Berlin: de Gruyter.
Kratzer, Angelika. . Modality. In Arnim von Stechow & Dieter Wunderlich
(eds.), Semantics: An international handbook of contemporary research, –.
Berlin: de Gruyter.
Kratzer has been updating her classic papers for a volume of her collected
work on modality and conditionals. These are very much worth studying:
http://semanticsarchive.net/Archive/TcNjAM/.
A major new resource on modality is Paul Portner’s book:
Portner, Paul. . Modality. Oxford University Press.
You might also profit from other survey-ish type papers:
von Fintel, Kai. . Modality and language. In Donald M. Borchert (ed.),
Encyclopedia of philosophy – second edition, MacMillan. URL http://mit.edu/
fintel/fintel--modality.pdf.
von Fintel, Kai & Anthony S. Gillies. . An opinionated guide to epistemic
modality. In Tamar Szabó Gendler & John Hawthorne (eds.), Oxford studies
in epistemology: Volume , –. Oxford University Press. URL http://mit.
edu/fintel/fintel-gillies--ose.pdf.
Swanson, Eric. . Modality in language. Philosophy Compass (). –.
doi:./j.-...x.
Hacquard, Valentine. . Modality. Ms, prepared for Semantics: An in-
ternational handbook of meaning, edited by Klaus von Heusinger, Claudia
Maienbon, and Paul Portner. URL http://ling.umd.edu/~hacquard/papers/
HoS_Modality_Hacquard.pdf.
On the syntax of modals, there are only a few papers of uneven quality. Some
of the more recent work is listed here. Follow up on older references from the
bibliographies in these papers.
Bhatt, Rajesh. . Obligation and possession. In Heidi Harley (ed.), Papers
from the upenn/mit roundtable on argument structure and aspect, vol.  MIT
Working Papers in Linguistics, –. URL http://people.umass.edu/bhatt/
papers/bhatt-haveto.pdf.
§. S R



Wurmbrand, Susi. . Modal verbs must be raising verbs. West Coast Conference
on Formal Linguistics . –. URL http://citeseerx.ist.psu.edu/viewdoc/
download?doi=....&rep=rep&type=pdf.
Cormack, Annabel & Neil Smith. . Modals and negation in English. In
Sjef Barbiers, Frits Beukema & Wim van der Wurff (eds.), Modality and its
interaction with the verbal system, –. Benjamins.
Butler, Jonny. . A minimalist treatment of modality. Lingua (). –.
doi:./S-()-.
The following paper explores some issues in the LF-syntax of epistemic modals:
von Fintel, Kai & Sabine Iatridou. . Epistemic containment. Linguistic
Inquiry (). –. doi:./.
Valentine Hacquard’s MIT dissertation is a rich source of cross-linguistic issues
in modality, as is Fabrice Nauze’s Amsterdam dissertation:
Hacquard, Valentine. . Aspects of modality: Massachusetts Institute of
Technology dissertation. URL http://people.umass.edu/hacquard/hacquard_
thesis.pdf.
Nauze, Fabrice. . Modality in typological perspective: Universiteit van Ams-
terdam dissertation. URL http://www.illc.uva.nl/Publications/Dissertations/
DS--.text.pdf.
The semantics of epistemic modals has become a hot topic recently. Here are
some of the main references:
Hacking, Ian. . Possibility. The Philosophical Review (). –.
doi:./. URL http://www.jstor.org/stable/.
Teller, Paul. . Epistemic possibility. Philosophia (). –. doi:./BF.
DeRose, Keith. . Epistemic possibilities. The Philosophical Review ().
–. doi:./.
Egan, Andy, John Hawthorne & Brian Weatherson. . Epistemic modals in
context. In Gerhard Preyer & Georg Peter (eds.), Contextualism in philosophy:
Knowledge, meaning, and truth, –. Oxford: Oxford University Press.
Egan, Andy. . Epistemic modals, relativism, and assertion. Philosophical
Studies (). –. doi:./s---x.
MacFarlane, John. . Epistemic modals are assessment-sensitive. Ms, Uni-
versity of California, Berkeley, forthcoming in an OUP volume on epis-
temic modals, edited by Brian Weatherson and Andy Egan. URL http:
//sophos.berkeley.edu/macfarlane/epistmod.pdf.
 M C 



Stephenson, Tamina. a. Judge dependence, epistemic modals, and predicates

of personal taste. Linguistics and Philosophy (). –. doi:./s-
--.
Hawthorne, John. . Eavesdroppers and epistemic modals. Ms, Rutgers
University, to appear in the proceedings of the  Sofia Conference in
Mexico, in a supplement to Noûs.
von Fintel, Kai & Anthony S. Gillies. a. CIA leaks. The Philosophical Review
(). –. doi:./--.
von Fintel, Kai & Anthony S. Gillies. b. Might made right. To appear in a
volume on epistemic modality, edited by Andy Egan and Brian Weatherson,
Oxford University Press. URL http://mit.edu/fintel/fintel-gillies--mmr.
pdf.
A recent SALT paper by Pranav Anand and Valentine Hacquard tackles what
happens to epistemic modals under attitude predicates:
Anand, Pranav & Valentine Hacquard. . Epistemics with attitude. Proceed-
ings of Semantics and Linguistic Theory . doi:/.
Evidentiality is a topic closely related to epistemic modality. Some references:
Willett, Thomas. . A cross-linguistic survey of the grammaticalization of
evidentiality. Studies in Language (). –.
Aikhenvald, Alexandra Y. . Evidentiality. Oxford: Oxford University Press.
Drubig, Hans Bernhard. . On the syntactic form of epistemic modality. Ms,
Universität Tübingen. URL http://www.sfb.uni-tuebingen.de/b/papers/
DrubigModality.pdf.
Blain, Eleanor M. & Rose-Marie Déchaine. . Evidential types: Evidence
from Cree dialects. International Journal of American Linguistics (). –.
doi:./.
McCready, Eric & Norry Ogata. . Evidentiality, modality and probability.
Linguistics and Philosophy (). –. doi:./s---.
Speas, Peggy. . On the syntax and semantics of evidentials. Language and
Linguistics Compass (). –. doi:./j.-X...x.
von Fintel, Kai & Anthony S. Gillies. . Must . . . stay . . . strong! Natural
Language Semantics (). –. doi:./s---.
Modals interact with disjunction and indefinites to generate so-called 
-readings, which are a perennial puzzle. Here is just a very small set of
initial references:
§. S R



Kamp, Hans. . Free choice permission. Proceedings of the Aristotelian Society,
New Series . –.
Zimmermann, Thomas Ede. . Free choice disjunction and epistemic possi-
bility. Natural Language Semantics (). –. doi:./A:.
Schulz, Katrin. . A pragmatic solution for the paradox of free choice
permission. Synthese (). –. doi:./s---y.
Aloni, Maria. . Free choice, modals, and imperatives. Natural Language
Semantics (). –. doi:./s---.
Alonso-Ovalle, Luis. . Disjunction in alternative semantics: University of
Massachusetts at Amherst dissertation. URL http://alonso-ovalle.net/index.
php?page_id=.
Fox, Danny. . Free choice and the theory of scalar implicatures. In Uli
Sauerland & Penka Stateva (eds.), Presupposition and implicature in com-
positional semantics, –. New York: Palgrave Macmillan. URL http:
//web.mit.edu/linguistics/people/faculty/fox/free_choice.pdf.
van Rooij, Robert. . Free choice counterfactual donkeys. Journal of Semantics
(). –. doi:./jos/ffl.
— T     —
 C F
C

We integrate conditionals into the semantics of modal expressions that

we are developing. We show that the material implication analysis and
the strict implication analysis are inferior to the restrictor analysis. Our
discussion will remain focussed on some simple questions and we refer
you to the rich literature on conditionals for further topics.

. The Material Implication Analysis 

. The Strict Implication Analysis 
. If -Clauses as Restrictors 
Supplemental Readings 

. The Material Implication Analysis

Consider the following example:

() If I am healthy, I will come to class.

The simplest analysis of such conditional constructions is the so-called 

 analysis, which treats if as contributing a truth-function operating
on the truth-values of the two component sentences (which are called the -
 and  — from Latin — or  and  — from
Greek). The lexical entry for if would look as follows:
Note that as a truth-functional
() JifK = λu ∈ Dt . λv ∈ Dt . u =  or v = . connective, this if does not
vary its denotation depending
on the evaluation world.
 Quoth the Stoic philosopher Philo of Megara: “a true conditional is one which does not have It’s its arguments that vary
a true antecedent and a false consequent” (according to Sextus Empiricus (c. : II, –)). with the evaluation world.
 C C 



Applied to example in (), this semantics would predict that the example is false
just in case the antecedent is true, I am healthy, but the consequent false, I do
not come to class. Otherwise, the sentence is true. We will see that there is much
to complain about here. But one should realize that under the assumption that
if denotes a truth-function, this one is the most plausible candidate.
Suber () does a good job of persuading (or at least trying to persuade)
recalcitrant logic students:

After saying all this, it is important to note that material implication

does conform to some of our ordinary intuitions about implication.
For example, take the conditional statement, If I am healthy, I will
come to class. We can symbolize it: H ⊃ C.
The question is: when is this statement false? When will I have
broken my promise? There are only four possibilities:
H C H⊃ C
T T ?
T F ?
F T ?
F F ?
• In case #, I am healthy and I come to class. I have clearly kept
my promise; the conditional is true.
• In case #, I am healthy, but I have decided to stay home and
read magazines. I have broken my promise; the conditional is
false.
• In case #, I am not healthy, but I have come to class anyway. I
am sneezing all over you, and you’re not happy about it, but I
did not violate my promise; the conditional is true.
• In case #, I am not healthy, and I did not come to class. I did
not violate my promise; the conditional is true.
But this is exactly the outcome required by the material implication.
The compound is only false when the antecedent is true and the
consequence is false (case #); it is true every other time.

Despite the initial plausibility of the analysis, it cannot be maintained. Consider

this example:
 The symbol ⊃ which Suber uses here is called the “horseshoe”. We have been using the right
arrow → as the symbol for implication. We think that this is much preferable to the confusing
horseshoe symbol. There is an intimate connection between universal quantification, material
implication, and the subset relation, usually symbolized as ⊂, which is the other way round from
the horseshoe. The horseshoe can be traced back to the notation introduced by Peano (), a
capital C standing for ‘conseguenza’ facing backwards. The C facing in the other (more “logical”)
direction was actually introduced first by Gergonne (), but didn’t catch on.
§. T M I A



() If there is a major earthquake in Cambridge tomorrow, my house will

collapse.

If we adopt the material implication analysis, we predict that () will be false
just in case there is indeed a major earthquake in Cambridge tomorrow but my
house fails to collapse. This makes a direct prediction about when the negation
of () should be true. A false prediction, if ever there was one:

() a. It’s not true that if there is a major earthquake in Cambridge

tomorrow, my house will collapse.
b. 6≡ There will be a major earthquake in Cambridge tomorrow, and
my house will fail to collapse.

Clearly, one might think that (a) is true without at all being committed to
what the material implication analysis predicts to be the equivalent statement in
(b). This is one of the inadequacies of the material implication analysis.
These inadequacies are sometimes referred to as the “paradoxes of material
implication”. But that is misleading. As far as logic is concerned, there is nothing
wrong with the truth-function of material implication. It is well-behaved and
quite useful in logical systems. What is arguable is that it is not to be used as a
reconstruction of what conditionals mean in natural language.
E .: Under the assumption that if has the meaning in (), calculate
the truth-conditions predicted for ():

() a. No student will succeed if he goofs off.

b. No student λx (if x goofs off, x will succeed)

State the predicted truth-conditions in words and evaluate whether they corre- S
spond to the actual meaning of (). 
Conditional S

if S
Modal S

A problem that is not often raised for the material implication analysis is how we are on Route  might R we be in Lockhart

badly it interacts with the analysis of modal expressions, once we look at sentences
Figure .: LF A for
involving both a conditional clause and a modal. Consider:
()
() If we are on Route , we might be in Lockhart now. S

Modal S

() If you keep this fern dry, it cannot grow. might R

Conditional S

if S we be in Lockhart

We need to consider two possible LFs for these sentences, depending on whether we are on Route 

wider scope is given to the modal or to the conditional clause. For example, in
the margin you see LFs A and B for (). Figure .: LF B for
()
 C C 



The reading for () we have in mind is an epistemic

Austin
one; imagine for instance that () is uttered in a car
by Mary to Susan, while Susan is driving and Mary
is looking at a map. The information provided by
the map, together with other background knowledge,
Martindale
Route 142
Lockhart
constitutes the relevant context for the modal might
here. The accessibility relation is roughly this:
Route 183

() λw. λw 0 . w 0 is compatible with what the map

Route 80

Luling
says in w and what Mary knows about the
geography of the relevant area in w.

Let’s suppose () is uttered in the actual world w

Figure .: A schematic
and we are interested in its truth-value at this world.
map of the relevant
We now proceed to show that neither of the LFs A and
area in Texas
B represent the intuitively natural meaning of () if
we assume the material implication analysis of if.
Consider first LF A. There are two respects in which
the predicted truth-conditions for this LF deviate from
intuitive judgment. First, suppose that Susan and Mary are not on Route 
in w . Then () is predicted to be true in w , regardless of the geographical
facts, e.g. even if Lockhart is nowhere near Route . This is counterintuitive.
Imagine the following quite sensible dialogue:

() Mary: If we are on Route , we might be in Lockhart now.

Susan (stops the car and looks at the map): You are wrong. Look here,
Route  doesn’t run anywhere near Lockhart.

If Mary concedes Susan’s claim that Route  doesn’t go through Lockhart, she
has to also concede that her original assertion was false. It wouldn’t do for her to
respond: “I know that  runs about  miles east of Lockhart, but maybe we
are not on Route , so I may still be right.” Yet we predict that this should be a
reasonable way for her to defend ().
A second inadequacy is this: we predict that the truth of the consequent of ()
is a sufficient condition for the truth of () as a whole. If this were right, it
would take very little for () to be true. As long as the map and the rest of
Mary’s knowledge in w don’t rule out the possibility that they are in Lockhart,
we might be in Lockhart will be true in w — regardless, once again, of whether
Lockhart is anywhere near . It should therefore be reasonable for Mary to
continue the dialogue in () with the rejoinder: “But how can you be so sure we
are not in Lockhart?” According to intuitive judgment, however, this would not
§. T S I A



be a pertinent remark and certainly would not help Mary defend () against
Susan’s objection.
Now let’s look at LF B, where the modal has widest scope. Given the material
implication analysis of if, this is predicted to mean, in effect: “It might be the
case that we are either in Lockhart or not on Route ”. This truth-condition
is also far too easy to satisfy: All it takes is that the map and the rest of Mary’s
knowledge in w are compatible with Mary and Susan not being on Route ,
or that they are compatible with their being in Lockhart. So as long as it isn’t
certain that they are on Route , Mary should be justified in asserting (),
regardless, once again, of her information about the relative location of Lockhart
and Route .

E .: Show that similar difficulties arise for the analysis of (). 

. The Strict Implication Analysis

Some of the problems we encountered would go away if we treated if as intro-
ducing a modal meaning. The simplest way to do that would be to treat it as
a universal quantifier over possible worlds. If p, q would simply mean that the
set of p-worlds is a subset of the q-worlds. This kind of analysis is usually called
 . The difference between if and must would be that if takes
an overt restrictive argument. Here is what the lexical entry for if might look
like:

() JifKw,g = λp ∈ Dhs,ti . λq ∈ Dhs,ti . ∀w 0 : p(w 0 ) =  → q(w 0 ) = .

(in set talk: p ⊆ q)

Applied to (), we would derive the truth-conditions that () is true iff all of
the worlds where there is a major earthquake in Cambridge tomorrow are worlds
where my house collapses.
We immediately note that this analysis has the same problem of non-contingency
that we faced with one of our early attempts at a quantificational semantics for
modals like must and may. The obvious way to fix this here is to assume that
if takes a covert accessibility function as one of its arguments. The antecedent
clause then serves as an additional restrictive device. Here is the proposal:

() JifKw,g = λR ∈ Dhs,hs,tii . λp ∈ Dhs,ti . λq ∈ Dhs,ti .

∀w 0 : (R(w)(w 0 ) =  & p(w 0 ) = ) → q(w 0 ) = .
(in set talk: R(w) ∩ p ⊆ q)
 C C 



If we understand () as involving an epistemic accessibility relation, it would

claim that among the worlds epistemically accessible from the actual world
(i.e. the worlds compatible with what we know), those where there is a major
earthquake in Cambridge tomorrow are worlds where my house collapses. This
would appear to be quite adequate — although potentially traumatic to me.

E .: Can you come up with examples where a conditional is interpreted
relative to a non-epistemic accessibility relation? 

E .: What prediction does the strict implication analysis make about
the negated conditional in (a)? 

What happens when we let this analysis loose on ()? We again need to assess
two LFs depending on the relative scope of if and might. Both LFs would have
two covert variables over accessibility relations, one for if and one for might.
Before we can assess the adequacy of the two candidate analyses, we need to
decide what the contextually salient values for the accessibility relations might be.
S
One would think that the epistemic accessibility relation that we have already
Conditional S encountered is the most likely value, and in fact for both variables.
S Modal S
if R
Next, we need to consider the particular epistemic state that Mary is in. By
we are on Route  might R we be in Lockhart

assumption, Mary does not know where they are. Nothing in her visual environ-
Figure .: LF A 0 for
ment helps her figure out where they are. She does see from the map that if they
()
are on Route , one of the towns they might be in is Lockhart. But she doesn’t
know whether they are on Route . Even if they are on , she doesn’t know
S

Modal
that they are and her epistemic state would still be what it is: one of being lost.
S

might R

Consider then LF A 0 , with the modal in the scope of the conditional. Here,
Conditional S

we be in Lockhart

we derive the claim that all worlds w 0 compatible with what Mary knows in w
S
if R
we are on Route 

and where they are on  are such that some world w 00 compatible with what
Figure .: LF B 0 for Mary knows in w 0 is such that they are in Lockhart. Is that adequate? Not really.
() We have just convinced ourselves that whether they are on  or not has no
relevant influence on Mary’s epistemic state, since she wouldn’t know it either
way. But that means that our analysis would predict that () is true as long as it
is possible as far as Mary knows that they are in Lockhart. Whether they are on
 or not doesn’t change that. So, we would expect () to not be distinct in
truth-value from something like:

() If we are on the Route , we might be in Lockhart.

But that is not right — Mary knows quite well that if they are on the Route ,
they cannot be in Lockhart.
§. If -C  R



Turning to LF B 0 , with the modal having widest scope, doesn’t help us either.
Here, we would derive the claim that it is compatible with what Mary knows
that from being on  it follows (according to what she knows) that they are in
Lockhart. Clearly, that is not what () means. Mary doesn’t consider it possible
that if they are on , she knows that they are in Lockhart. After all, she’s well
aware that she doesn’t know where they are.

. If -Clauses as Restrictors

The problem we have encountered here with the interaction of an if -clause
and the modal operator might is similar to others that have been noted in the
literature. Most influentially, David Lewis in his paper “Adverbs of Quantifica-
tion” showed how hard it is to find an adequate analysis of the interaction of
if -clauses and    like never, rarely, sometimes, often,
usually, always. Lewis proposed that in the cases he was considering, the adverb
is the only operator at work and that the if -clause serves to restrict the adverb.
Thus, it has much the same function that a common noun phrase has in a
determiner-quantification.

The if of our restrictive if -clauses should not be regarded as a

sentential connective. It has no meaning apart from the adverb
it restricts. The if in always if . . . , . . . , sometimes if . . . , . . . , and
the rest is on a par with the non-connective and in between . . . and
. . . , with the non-connective or in whether . . . or . . . , or with the
non-connective if in the probability that . . . if . . . . It serves merely to
mark an argument-place in a polyadic construction. (Lewis : )

Building on Lewis’ insight, Kratzer argued for a uniform treatment of if -clauses

as restrictors. She claimed that

the history of the conditional is the story of a syntactic mistake.

There is no two-place if . . . then connective in the logical forms of
natural languages. If -clauses are devices for restricting the domains
of various operators. (Kratzer )

Let us repeat this:

() K’ T

If -clauses are devices for restricting the domains of various operators.

Kratzer’s Thesis gives a unified picture of the semantics of conditional clauses.

Note that it is not meant to supplant previous accounts of the meaning of
 C C 



conditionals. It just says that what those accounts are analyzing is not the
meaning of if itself but the meaning of the operators that if -clauses restrict.
Let us see how this idea helps us with our Lockhart-sentence. The idea is to
deny that there are two quantifiers over worlds in (). Instead, the if -clause
merely contributes a further restriction to the modal might. In effect, the modal
is not quantifying over all the worlds compatible with Mary’s knowledge but
only over those where they are on Route . It then claims that at least some of
S
those worlds are worlds where they are in Lockhart. We cannot anymore derive
the problematic conclusion that it should also be true that if they are on the
Modal S
Route , they might be in Lockhart. In all, we have a good analysis of what
() means.
we be in Lockhart
might

R w* (if ) S

What we don’t yet have is a compositional calculation. What does it mean in

we are on Route 

structural terms for the if -clause to be restricting the domain of the modal? We
Figure .: LF C for
will assume a structure as in LF C. Here, the if -clause is the sister to what used
()
to be the covert set-of-worlds argument of the modal. As you can see, we have
chosen the variant of the semantics for modals that was discussed in Section ...
The idea now is that the two restrictive devices work together: we just feed to
the modal the intersection of (i) the set of worlds that are R-accessible from the
actual world, and (ii) the set of worlds where they are on Route .
E .: To make the composition work, we need to be able to intersect
the set of accessible worlds with the antecedent proposition. This could be
done in two ways: (i) a new composition principle, which would be a slight
modification of the P M rule, (ii) give if a functional
meaning that accomplishes the intersection. Formulate such a meaning for if.
Alternatively, we could do without the w∗ device and instead give if a meaning
that takes a proposition p and then modifies an accessibility relation to give
a new accessibility relation, which is restricted to p-worlds. Formulate such a
meaning for if. 

What about cases like (), now? Here there is no modal operator for the if -
clause to restrict. Should we revert to treating if as an operator on its own?
Kratzer proposes that we should not and that such cases simply involve covert
modal operators.
§. If -C  R



Supplementary Readings
A short handbook article on conditionals:
von Fintel, Kai. . Conditionals. Ms, prepared for Semantics: An international
handbook of meaning, edited by Klaus von Heusinger, Claudia Maienborn, and
Paul Portner. URL http://mit.edu/fintel/fintel--hsk-conditionals.pdf.
Overviews of the philosophical work on conditionals:
Edgington, Dorothy. . On conditionals. Mind (). –. URL
./mind/...
Bennett, Jonathan. . A philosophical guide to conditionals. Oxford University
Press.
A handbook article on the logic of conditionals:
Nute, Donald. . Conditional logic. In Dov Gabbay & Franz Guenthner
(eds.), Handbook of philosophical logic. volume ii, –. Dordrecht: Reidel.
Three indispensable classics:
Lewis, David. . Counterfactuals. Oxford: Blackwell.
Stalnaker, Robert. . A theory of conditionals. In Nicholas Rescher (ed.), Stud-
ies in logical theory (American Philosophical Quarterly Monograph Series ),
–. Oxford: Blackwell.
Stalnaker, Robert. . Indicative conditionals. Philosophia (). –.
doi:./BF.
The Restrictor Analysis:
Lewis, David. . Adverbs of quantification. In Edward Keenan (ed.), Formal
semantics of natural language, –. Cambridge University Press.
Kratzer, Angelika. . Conditionals. Chicago Linguistics Society (). –.
The application of the restrictor analysis to the interaction of nominal quantifiers
and conditionals:
von Fintel, Kai. . Quantifiers and ‘if ’-clauses. The Philosophical Quarterly
(). –. doi:./-.. URL http://mit.edu/fintel/
www/qandif.pdf.
von Fintel, Kai & Sabine Iatridou. . If and when If -clauses can restrict
quantifiers. Ms, MIT. URL http://mit.edu/fintel/fintel-iatridou--ifwhen.
pdf.
Higginbotham, James. . Conditionals and compositionality. Philosophical
Perspectives (). –. doi:./j.-...x.
 C C 



Leslie, Sarah-Jane. . If, unless, and quantification. In Robert J. Stainton &
Christopher Viger (eds.), Compositionality, context and semantic values: Essays
in honour of Ernie Lepore, –. Springer. doi:./----_.
Huitink, Janneke. b. Quantified conditionals and compositionality. Ms, to
appear in Language and Linguistics Compass. URL http://user.uni-frankfurt.
de/~huitink/compass-conditionals-final.pdf.
Syntax of conditionals:
von Fintel, Kai. . Restrictions on quantifier domains: University of Mas-
sachusetts at Amherst dissertation. URL http://semanticsarchive.net/Archive/
jANIwN/fintel--thesis.pdf, Chapter : “Conditional Restrictors”
Iatridou, Sabine. . On the contribution of conditional Then. Natural
Language Semantics (). –. doi:./BF.
Bhatt, Rajesh & Roumyana Pancheva. . Conditionals. In The Blackwell
companion to syntax, vol. , –. Blackwell. URL http://www-rcf.usc.edu/
~pancheva/bhatt-pancheva_syncom.pdf.
A shifty alternative to the restrictor analysis:
Gillies, Anthony S. . On truth-conditions for if (but not quite only if ).
The Philosophical Review (). –. doi:./--.
Gillies, Anthony S. . Iffiness. Semantics and Pragmatics (). –. doi:./sp...
The Belnap alternative:
Belnap, Jr., Nuel D. . Conditional assertion and restricted quantification.
Noûs (). –. doi:./.
Belnap, Jr., Nuel D. . Restricted quantification and conditional assertion.
In Hugues Leblanc (ed.), Truth, syntax and modality: Proceedings of the Temple
University conference on alternative semantics, vol.  Studies in Logic and the
Foundations of Mathematics, –. Amsterdam: North-Holland.
von Fintel, Kai. . If : The biggest little word. Slides from a plenary
address given at the Georgetown University Roundtable, March , . URL
http://mit.edu/fintel/gurt-slides.pdf.
Huitink, Janneke. . Modals, conditionals and compositionality: Radboud Uni-
versiteit Nijmegen dissertation. URL http://user.uni-frankfurt.de/~huitink/
Huitink-dissertation.pdf, Chapters  and  give a nice summary of what we’re
covering in this class, while Chapter  is about the Belnap-method.
Huitink, Janneke. a. Domain restriction by conditional connectives. Ms,
Goethe-University Frankfurt. URL http://semanticsarchive.net/Archive/zgMDMM/
Huitink-domainrestriction.pdf.
 C F
O

We have stressed throughout the previous two chapters that there are
numerous parallels between quantification over ordinary individuals via
determiner quantifiers and quantification over possible worlds via modal
operators (including conditionals). Now, we turn to a phenomenon that
(at least at first glance) appears to show that there are non-parallels
as well: a sensitivity to an  of the elements in the domain
of quantification. We first look at this in the context of simple modal
sentences and then we look at conditionals.

. The Driveway 

. Kratzer’s Solution: Doubly Relative Modality 
. The Paradox of the Good Samaritan 
. Non-Monotonicity of Conditionals 
Supplemental Readings 

. The Driveway

Consider a typical use of a sentence like ().

() John must pay a fine.

This is naturally understood in such a way that its truth depends both on facts
about the law and facts about what John has done. For instance, it will be
judged true if (i) the law states that driveway obstructors are fined, and (ii) John
has obstructed a driveway. It may be false either because the law is different or
because John’s behavior was different.
 O C 



What accessibility relation provides the implicit restriction of the quantifier must
on this reading of ()? A naïve attempt might go like this:

() λw. λw 0 . [what happened in w 0 up to now is the same as what happened

in w and w 0 conforms to what the law in w demands].

The problem with () is that, unless there were no infractions of the law at
all in w up to now, no world w 0 will be accessible from w. Therefore, () is
predicted to follow logically from the premise that John broke some law. This
does not represent our intuition about its truth conditions.
A better definition of the appropriate accessibility relation has to be more com-
plicated:

() λw. λw 0 . [what happened in w 0 up to now is the same as what happened

in w and w 0 conforms at least as well to what the law in w demands as
does any other world in which what happened up to now is the same as
in w].

() makes explicit that there is an important difference between the ways in
which facts about John’s behavior on the one hand, and facts about the law on
the other, enter into the truth conditions of sentences like (). Worlds in
which John didn’t do what he did are simply excluded from the domain of must
here. Worlds in which the law isn’t obeyed are not absolutely excluded. Rather,
we restrict the domain to those worlds in which the law is obeyed as well as it
can be, considering what has happened. We exclude only those worlds in which
there are infractions above and beyond those that are shared by all the worlds
in which John has done what he has done. The analysis of () thus crucially
involves the notion of an ordering of worlds: here they are ordered according to
how well they conform to what the law in w demands.

. Kratzer’s Solution: Doubly Relative Modality

Kratzer proposes that modal operators are sensitive to two context-dependent
parameters: a set of accessible worlds (provided by an accessibility function com-
puted from a conversational background, the  ), and a partial ordering
of the accessible worlds (computed from another conversational background,
called the  ).
Let’s see how the analysis applies to the previous example.

• The modal base will be a function that assigns to any evaluation world
§. K’ S: D R M



a set of propositions describing the relevant circumstances, for example,

what John did. Since in our stipulated evaluation world John obstructed a
driveway, the modal base will assign the proposition that John obstructed
a driveway to this world. The set of worlds accessible from the evaluation
world will thus only contain worlds where John obstructed a driveway.
• The ordering source will be a function that assigns to any evaluation world
a set of propositions P whose truth is demanded by the law. Imagine that
for our evaluation world this set of propositions contains (among others)
the following two propositions: (i) nobody obstructs any driveways, (ii)
anybody who obstructs a driveway pays a fine.
• The idea is now that such a set P of propositions can be used to order the
worlds in the modal base. For any pair of worlds w and w , we say that
w comes closer than w to the ideal set up by P (in symbols: w <P w ),
iff the set of propositions from P that are true in w is a proper subset of
the set of propositions from P that are true in w .
• For our simple example then, any world in modal base where John pays a
fine will count as better than an otherwise similar world where he doesn’t.
• Modals then make quantificational claims about the best worlds in the
modal base (those for which there isn’t a world that is better than them).
• In our case, () claims that in the best worlds (among those where John
obstructed a driveway), he pays a fine.

More technically:

() Given a set of worlds X and a set of propositions P, define the 
  <P as follows:
∀w , w ∈ X : w <P w iff {p ∈ P : p(w ) = } ⊂ {p ∈ P : p(w ) = }.

() For a given strict partial order <P on worlds, define the selection function
maxP that selects the set of <P -best worlds from any set X of worlds:
∀X ⊆ W : maxP (X) = {w ∈ X : ¬∃w 0 ∈ X : w 0 <P w}.

() JmustKw,g = λfhs,sti . λghs,sti . λqhs,ti .

∀w 0 ∈ maxg(w) (∩f(w)) : q(w 0 ) = .

T N: This only works if we can in general assume that the <P
relation has minimal elements, that there always are accessible worlds that
come closest to the P-ideal, worlds that are better than any world they can
be compared with via <P . It is possible, with some imagination, to cook up
scenarios where this assumption fails. This problem has been discussed primarily
in the area of the semantics of conditionals. There, Lewis presents relevant
scenarios and argues that one shouldn’t make this assumption, which he calls the
Limit Assumption. Stalnaker, on the one other hand, defends the assumption
 O C 



against Lewis’ arguments by saying that in actual practice, in actual natural

language semantics and in actual modal/conditional reasoning, the assumption
is eminently reasonable. Kratzer is persuaded by Lewis’ evidence and does not
make the Limit Assumption; hence her semantics for modals is more convoluted
than what we have in () and (). I will side with Stalnaker, not the least
because it makes life easier. For further discussion, see Lewis (: –) and
Stalnaker (: Chapter , esp. pp. –); Pollock (), Herzberger (),
and Warmbrod () argue for the Limit Assumption as well.
E .: In her handbook article Kratzer (), Kratzer presents a number
of examples of modal statements and sketches an analyses in terms of doubly
relative modality. You should study her examples carefully. 

. The Paradox of the Good Samaritan

Prior () introduced the following “Paradox of the Good Samaritan”. Imagine
that someone has been robbed and John is walking by. It is easy to conceive of a
code of ethics that would make the following sentence true:

() John ought to help the person who was robbed.

In our previous one-factor semantics for modals, we would have said that ()
says that in all of the deontically accessible worlds (those compatible with the
code of ethics) John helps the person who was robbed. Prior’s point was that
under such a semantics, something rather unfortunate holds. Notice that in all
of the worlds where John helps the person who was robbed, someone was robbed
in the first place. Therefore, it will be true that in all of the deontically accessible
worlds, someone was robbed. Thus, () will entail:

() It ought to be the case that someone was robbed.

It clearly would be good not make such a prediction.

The doubly-relative analysis of modality can successfully avoid this unfortunate
prediction. We conceive of () as being uttered with respect to a circumstantial
modal base that includes the fact that someone was robbed. Among those already
somewhat ethically deficient worlds, the relatively best ones are all worlds where
John helps the victim.
Note that we still have the problematic fact that among the worlds in the modal
base, all are worlds where someone was robbed, and we would thus appear to
still make the unfortunate prediction that () should be true. But this can now
be fixed. For example, we could say that ought p is semantically defective if p is
§. N-M  C



true throughout the worlds in the modal base. This could be a presupposition
or some other ingredient of meaning. So, with respect to a modal base which
pre-determines that someone was robbed, one couldn’t felicitously say ().
Consequently, saying () would only be felicitous if a different modal base
is intended, one that contains both p and non-p worlds. And given a choice
between worlds where someone was robbed and worlds where nobody was
robbed, most deontic ordering sources would probably choose the no-robbery
worlds, which would make () false, as desired.

K’    S P Kratzer () argues that

the restrictor approach to deontic conditionals is the crucial ingredient in the
solution to a conditional version of the Samaritan Paradox:

() If a murder occurs, the jurors must convene.

Kratzer points out that if one tried to analyze () as a material implication
embedded under deontic necessity, then one quickly runs into a problem. Surely,
one wants the following to be a true statement about the law:

() There must be no murder.

But this means that in the deontically accessible worlds, all of them have no
murders occurring. Now, this means that in all of the deontically accessible
worlds, any material implication of the form “if a murder occurs, q” will be true
no matter what the consequent is since the antecedent will be false. Since that
is an absurd prediction, () cannot be analyzed as material implication under
deontic necessity. The combination of the restrictor approach to if -clauses and
the doubly-relative theory of modals can rescue us from this problem. () is
analyzed as the deontic necessity modal being restricted by the if -clause. The set
of accessible worlds is narrowed down by the if -clause to only include worlds
in which a murder occurs. The deontic ordering then identifies the best among
those worlds and those are plausibly all worlds where the jurors convene.

. Non-Monotonicity of Conditionals

The last case discussed takes us straight to the crucial role of the ordering of
worlds in the semantics of conditionals, as we would of course expect under
the analysis of if -clauses as restrictors of modal operators. In this arena, the
discussion usually revolves around the failure of certain inference patterns, which
 O C 



one would expect a universal quantifier to validate. Here are the most important
ones:

() L D M (“D E”)

Every A is a B. → Every A & C is a B.
() T
Every A is a B. Every B is a C. → Every A is a C.
() C
Every A is a B. → Every non-B is a non-A.

Conditionals were once thought to obey these patterns as well, known in condi-
tional logic as S  A, H S,
and C. But then spectacular counterexamples became known
through the work of Stalnaker and Lewis.

() F  S  A

a. If I strike this match, it will light.
If I dip this match into water and strike it, it will light.
b. If John stole the earrings, he must go to jail.
If John stole the earrings and then shot himself, he must go to jail.
c. If kangaroos had no tails, they would topple over. If kangaroos had
no tails but used crutches, they would topple over.
() F   H S (T)
a. If Brown wins the election, Smith will retire to private life.
If Smith dies before the election, Brown will win the election.
If Smith dies before the election,Smith will retire to private life.
b. If Hoover had been a Communist, he would have been a traitor.
If Hoover had been born in Russia, he would have been a Commu-
nist.
If Hoover had been born in Russia, he would have been a traitor.
() F  C
a. If it rained, it didn’t rain hard.
If it rained hard, it didn’t rain.
The Goethe example is due to b. (Even) if Goethe hadn’t died in , he would still be dead now.
Kratzer.
If Goethe were alive now, he would have died in .

Note that these cases involve examples of both “indicative” (epistemic) condi-
tionals and counterfactual conditionals. It is sometimes thought that indicative
§. N-M  C



conditionals are immune from these kinds of counterexamples, but it is clear that
they are not. Also note that in (b) we have a case of Failure of Strengthening
the Antecedent with a deontic conditional. Deontic counterexamples to the
other patterns seem harder to find.
The failure of these inference patterns again indicates that the semantics of
modal operators (restricted by if -clauses) is more complicated than the simple
universal quantification we had previously been assuming. The basic idea of
most approaches to this problem is this: the semantics of conditionals is more
complicated than simple universal quantification. The conditional does not make
a claim about simply every antecedent world, nor even about every contextually
relevant antecedent world. Instead, in each of the conditional statements, only a
particular subset of the antecedent worlds is quantified over. Informally, we can
call those the “most highly ranked antecedent worlds”. Consider:

() If I had struck this match, it would have lit.

If I had dipped this match into water and struck it, it would have lit.

According to the Stalnaker-Lewis account, this inference is semantically invalid.

The premise merely claims that the most highly ranked worlds in which I strike
this match are such that it lights. No claim is made about the most highly ranked
worlds in which I first dip this match into water and then strike it. Strengthening
the Antecedent will only be safe if it is additionally known that the strengthened
antecedent is instantiated among the worlds that verify the original antecedent.
The other fallacies receive similar treatments. Transitivity (Hypothetical Syllo-
gism) fails for the new non-monotonic quantifier because even if all the most
highly rated p-worlds are q-worlds and all the most highly rated q-worlds are
r-worlds, we are not necessarily speaking about the same q-worlds (the q-worlds
that p takes us to may be rather remote ones). So in the Hoover-example, we
get the following picture: The most highly ranked p-worlds in which Hoover
was born in Russia (but where he retains his level of civic involvement), are all
q-worlds in which he becomes a Communist. On the other hand, the most
highly ranked q-worlds in which he is a Communist (but retaining his having
been born in the United States and being a high level administrator) are all
r-worlds in which he is a traitor. However, the most highly ranked p-worlds do
not get us to the most highly ranked q-worlds, so the Transitive inference does
not go through.
Contraposition fails because the fact that the most highly rated p-worlds are
q-worlds does not preclude a situation where the most highly rated non q-worlds
are also p-worlds. The most highly rated p-worlds in which Goethe didn’t die in
 are all q-worlds where he dies nevertheless (well) before the present. But of
course, the most highly rated (in fact, all) non-q-worlds (where he is alive today)
 O C 



are also p-worlds where he didn’t die in .

In the conditionals literature, the ordering of worlds is usually given directly as
an evaluation parameter. The typical gloss is that the ordering ranks possible
worlds based on how similar they are to the evaluation world. Kratzer developed
an alternative where the ordering is computed from a set of propositions true in
the evaluation world. Lewis () showed that ordering semantics and premise
semantics are largely notational variants.

Supplementary Readings

The central readings for this chapter are two papers by Kratzer:
Kratzer, Angelika. . Modality. In Arnim von Stechow & Dieter Wunderlich
(eds.), Semantics: An international handbook of contemporary research, –.
Berlin: de Gruyter.
Kratzer, Angelika. . The notional category of modality. In Hans-Jürgen
Eikmeyer & Hannes Rieser (eds.), Words, worlds, and contexts: New approaches
in word semantics (Research in Text Theory ), –. Berlin: de Gruyter.
Some work that discusses and uses Kratzer’s two factor semantics for modals:
Frank, Anette. . Context dependence in modal constructions: Universität
Stuttgart dissertation. URL http://www.cl.uni-heidelberg.de/~frank/papers/
header.pdf.
von Fintel, Kai & Sabine Iatridou. . What to do if you want to go to
Harlem: Anankastic conditionals and related matters. Ms, MIT. URL http:
//mit.edu/fintel/fintel-iatridou--harlem.pdf.
von Fintel, Kai & Sabine Iatridou. . How to say ought in Foreign: The
composition of weak necessity modals. In Jacqueline Guéron & Jacqueline
Lecarme (eds.), Time and modality (Studies in Natural Language and Linguistic
Theory ), –. Springer. doi:./----.
Some work that discusses whether non-monotonicity could be or might have to
be relegated to a dynamic pragmatic component of meaning:
von Fintel, Kai. . Counterfactuals in a dynamic context. In Michael
Kenstowicz (ed.), Ken Hale: A life in language, –. MIT Press. URL
http://mit.edu/fintel/counterfactuals.pdf.
von Fintel, Kai. . NPI licensing, Strawson entailment, and context depen-
dency. Journal of Semantics (). –. doi:./jos/...
§. N-M  C



Gillies, Anthony S. . Counterfactual scorekeeping. Linguistics and Philosophy

(). –. doi:./s---.
Schlenker explored whether the apparent non-monotonicity in conditional is
paralleled in quantification over individuals:
Schlenker, Philippe. . Conditionals as definite descriptions (A referential
analysis). Research on Language and Computation (). –. doi:./s-
--.
— T     —
 C S
B  T  A

We explore an analysis of tense that treats tenses as intensional operators

manipulating a time parameter of evaluation. The treatment is formally
quite parallel to the treatment of modals in Chapter . We touch on
many basic questions about tense and aspect, without exploring them
fully.

. A First Proposal for Tense 

.. former 
.. Some Time Adverbials 
.. A Word of Caution 
. Are Tenses Referential? 
. The Need for Intervals 
. Aktionsarten 
. The Progressive 
. Tense in Embedded Clauses 
Supplemental Readings 

. A First Proposal for Tense

Tense logic, or temporal logic, is a branch of logic first developed by the aptly See the Stanford Encyclopedia
of Philosophy entry on tempo-
named Arthur Prior in a series of works, in which he proposed treating tense ral logic: http://plato.stanford.
in a way that is formally quite parallel to the treatment of modality discussed edu/entries/logic-temporal/
and the website for Prior
in Chapter . Since tense logic (and modal logic) typically is formulated at a studies: http://www.
high level of abstraction regarding the structure of sentences, it doesn’t concern prior.aau.dk/index.htm.
itself with the internal make-up of “atomic” sentences and thus treats tenses as
sentential operators (again, in parallel to the way modal operators are typically
 B  T  A C 



treated in modal logic). We will implement a version of Prior’s tense logic in our
framework.
The first step is to switch to a version of our intensional semantic system where
instead of a world parameter, the evaluation function is sensitive to a time
parameter (and a variable assignment). Eventually, we will want to deal with
the full complexity and relativize the evaluation function to both worlds and
times, but for now, we will just relativize to times. The composition principles
developed in Chapter  will be adopted mutatis mutandis. Predicates will now
have lexical entries that incorporate their sensitivity to time:

() JtiredKt,g = λx ∈ D. x is tired at t.

It is customary in the literature to introduce a new basic type for times; for
now, we will recycle the designation s as the type for times. Then, for example,
the intension of sentence will again be of type hs, ti, but now that would be a
temporal proposition, a function from times to truth-values.
In this framework, we can now formulate a very simple-minded first analysis
of the present and past tenses and the future auxiliary will. As for (LF) syntax
let’s assume that (complete matrix) sentences are TPs, headed by T (for “tense”).
There are two morphemes of the functional category T, namely PAST (past
tense) and PRES (present tense). The complement of T is an MP or a VP.
MP is headed by M (for “modal”). Morphemes of the category M include the
modal auxiliaries must, can, etc., which we talked about in previous chapters, the
semantically vacuous do (in so-called “do-support” structures), and the future
auxiliary will. Evidently, this is a semantically heterogeneous category, grouped
together solely because of their common syntax (they are all in complementary
distribution with each other). The complement of M is a VP. When the sentence
contains none of the items in the category M, we assume that MP isn’t projected
at all; the complement of T is just a VP in this case. We thus have LF-structures
like the following. (The corresponding surface sentences are given below, and
we won’t be explicit about the derivational relation between these and the LFs.
Assume your favorite theories of syntax and morphology here.)

() [TP Mary [ T’ PRES [ VP t [ V’ be tired ]]]]

= Mary is tired.
() [TP Mary [ T’ PAST [ VP t [ V’ be tired ]]]]
= Mary was tired.

 We remain vague for now about what we mean by “times” (points in time? time intervals?).
This will need clarification, We will also see the need to clarify what we mean by “at” in the
metalanguage in this entry and others.
§. A F P  T



() [ TP Mary [ T’ PRES [ MP t [ M’ woll [ VP t [ V’ be tired ]]]]]]

= Mary will be tired.

When we have proper name subjects, we will pretend for simplicity that they are
reconstructed somehow into their VP-internal base position. (We will talk more
about reconstruction later on.)
What are the meanings of PRES, PAST, and will? For PRES, the simplest
assumption is actually that it is semantically vacuous. This means that the
interpretation of the LF in () is identical to the interpretation of the bare VP
Mary be tired:

() For any time t:

JPRES (Mary be tired)Kt = JMary be tiredKt =  iff Mary is tired at t.

Does this adequately capture the intuitive truth-conditions of the sentence Mary
is tired ? It does if we make the following general assumption:

() An utterance of a sentence (= LF) φ that is made at a time t counts as

true iff JφKt =  (and as false if JφKt = ).

This assumption ensures that (unembedded) sentences are, in effect, interpreted

as claims about the time at which they are uttered (“utterance time” or “speech
time”). If we make this assumption and we stick to the lexical entries we have
adopted, then we are driven to conclude that the present tense has no semantic
job to do. A tenseless VP Mary be tired would in principle be just as good as
() to express the assertion that Mary is tired at the utterance time. Apparently
it is just not well-formed as an unembedded structure, but this fact must be
attributed to principles of syntax rather than semantics.
What about PAST ? When a sentence like () Mary was tired is uttered at a
time t, then what are the conditions under which this utterance is judged to be
true? A quick (and perhaps ultimately wrong) answer is: an utterance of () at
t is true iff there is some time before t at which Mary is tired. This suggests the
following entry:

() For any time t:

JPASTKt = λp ∈ Dhs,ti .∃t 0 before t : p(t 0 ) = 

So, the past tense seems to be an existential quantifier over times, restricted to
times before the utterance time.
For will, we can say something completely analogous:
 B  T  A C 



() For any time t:

JwillKt = λp ∈ Dhs,ti .∃t 0 after t : p(t 0 ) = 

Apparently, PAST and will are semantically alike, even mirror images of each
other, though they are of different syntactic categories. The fact that PAST
is the topmost head in its sentence, while will appears below PRES, is due to
the fact that syntax happens to require a T-node in every complete sentence.
Semantically, this has no effect, since PRES is vacuous.
Both () and () presuppose that the set T comes with an intrinsic order.
For concreteness, assume that the relation ‘precedes’ (in symbols: <) is a strict
linear order on T . The relation ‘follows’, of course, can be defined in terms of
‘precedes’ (t follows t 0 iff t 0 precedes t).
There are many things wrong with this simple analysis. We will not have time
here to diagnose most of the problems, much less correct them. But let’s see a
couple of things that work out OK and let’s keep problems and remedies for
later.

.. former
There is a brief discussion on p.  of H& K about the inadequacy of an
extensional semantics for the adjective former as in

() John is a former teacher.

We can now write a semantics for former. While there are a bunch of people who
are currently teachers, there are others that aren’t now teachers but were at some
previous time. The latter are the ones that the predicate former teacher should be
true of, In other words, former teacher is a predicate that is true of individuals just
in case the predicate teacher was true of them at some previous time (and is not
true of them now). So, former needs to be an intensional operator that “displaces”
the evaluation of time of its complement from “now” to some previous time.
To be able to do that, it needs to take the intension of its complement as its
argument. This suggests the following lexical entry:

 Definition: A relation R is a strict linear order on a set S iff it has the following four properties:
(i) ∀x∀y∀z((Rxy&Ryz) → Rxz) “Transitivity"
(ii) ∀x(¬Rxx) “Irreflexivity"
(iii) ∀x∀y(Rxy → ¬Ryx) “Asymmetry", and
(iv) ∀x∀y(x 6= y → (Rxy ∨ Ryx)) “Connectedness"
§. A F P  T



() JformerK = λf ∈ Dhs,he,tii .λx.[f(t)(x) =  & ∃t 0 before t : f(t 0 )(x) = ].

E .: H& K on p. mention the adjective alleged in one breath with
former. Formulate a lexical entry for alleged as used in John is an alleged murderer.
[This will use our original intensional system with a world parameter] 

.. Some Time Adverbials

At least to a certain extent, we can also provide a treatment of temporal adverbials
such as:

() Mary was tired on February , .

The basic idea would be that phrases like on February ,  are propositional
modifiers. Propositions are the intensions of sentences. At this point, proposi-
tions are functions from times to truth-values. Propositional modifiers take a
proposition and return a proposition with the addition of a further condition on
the time argument.

() Jon February , Kt

= λp ∈ Dhs,ti . [ p(t) =  & t is part of Feb ,  ]

() [T’ PAST [ VP [ VP Mary [ V’ be tired]] [ PP on February , ]]]

An alternative would be to treat on February ,  as a “sentence” by itself,

whose intension then would be a proposition.

() Jon February , Kt =  iff t is part of February , 

() JonKt = λx. t is part of x

To make this work, we would then have to devise a way of combining two
tenseless sentences (Mary be tired and on February , ) into one. We could
do this by positing a silent and or by introducing a new composition rule
(“Propositional Modification”?).
Let’s not spend time on such a project.

E .: Imagine that Mary was tired on February ,  is not given the
LF in () but this one:

() [ T’ [ T’ PAST [ VP Mary [ V’ be tired]]] [ PP on February , ]]

 B  T  A C 



What would the truth-conditions of this LF be? Does this result correspond at
all to a possible reading of this sentence (or any other analogous sentence)? If
not, how could we prevent such an LF from being produced?

E .: When a quantifier appears in a tensed sentence, we might expect

two scope construals. Consider a sentence like this:

() Every professor (in the department) was a teenager in the Sixties.

We can imagine two LFs:

() PAST [ [every professor be a teenager] [in the sixties] ]

() [every professor] λ [ PAST [ [t be a teenager] [in the sixties] ]

Describe the different truth-conditions which our system assigns to the two LFs.
Is the sentence ambiguous in this way?
If not this sentence, are there analogous sentences that do have the ambiguity?

E .: The following entry for every makes it a time-insensitive item:

() JeveryKt = λf ∈ Dhe,ti .λg ∈ Dhe,ti .∀x[f(x) =  → g(x) = ]

Consider now two possible variants (we have underlined the portion where they
differ):

() JeveryKt = λf ∈ Dhe,ti .λg ∈ Dhe,ti .∀x at t [f(x) =  → g(x) = ]

() JeveryKt = λf ∈ Dhe,ti .λg ∈ Dhe,ti .∀x[f(x) =  at t → g(x) =  at t]

Does either of these alternative entries make sense? If so, what does it say? Is it
equivalent to our official entry? Could it lead to different predictions about the
truth-conditions of English sentences?

.. A Word of Caution

Compare the semantics given for former and the one for PAST :

() JformerK = λf ∈ Dhs,he,tii .λx.[f(t)(x) =  & ∃t 0 before t : f(t 0 )(x) = ].

() JPASTKt = λp ∈ Dhs,ti . ∃t 0 before t : p(t 0 ) = 

§. A T R?



Notice that these entries have an interesting consequence:

() a. John is a former teacher.

b. John was a teacher.

The two sentences in () differ in their truth-conditions. The sentence in

(a) can only be true if John is not a teacher anymore while this is not part of
the truth-conditions of the sentence in (b). To see that this analysis is in fact
correct, consider this:

() Last night, John was reading a book about tense.

a. !! The authors are former Italians.
b. The authors were Italian.

Consider the past tense in (b). It is not (necessarily) interpreted as claiming

that the authors are not Italian anymore. But this is in fact required by (a).
There are some cases where it seems that the past tense does trigger inferences
that one would not expect from the lexical entry that we gave. Surely, if I tell you
My cousin John was a teacher you will infer that he isn’t a teacher anymore. In fact,
you may even infer that he is not alive anymore. One promising approach that
tries to reconcile a semantics like ours with the possibility of stronger inferences
in some contexts is based on pragmatic considerations, see Musan ().
Examples like the one in () are problematic for widely held naive conceptions
of what the past tense means. One often hears that PAST expresses the fact that
“the time of the reported situation precedes the speech time”. If this were to
mean that the time of the book’s authors being Italian precedes the speech time,
this would presumably wrongly predict that they would have to be not Italian
anymore for the sentence to be true (or usable).

. Are Tenses Referential?

Our semantics for the past tense treats it essentially as an existential quantifier
over times (albeit in the meta-language), the same way we treated possibility
modals as existential quantifiers over (accessible) worlds. This seems quite
adequate for examples like (), which seem to display the expected quantified
meaning:

() John went to a private school.

 B  T  A C 



All we learn from () is that at some point in the past, whenever it was that
John went to school, he went to a private school.
Partee in her famous paper “Some structural analogies between tenses and
pronouns in English” (Partee ) presented an example where tense appears to
act more “referentially”:

() I didn’t turn off the stove.

“When uttered, for instance, halfway down the turnpike, such a sentence clearly
does not mean either that there exists some time in the past at which I did not
turn off the stove or that there exists no time in the past at which I turned off
the stove. The sentence clearly refers to a particular time — not a particular
instant, most likely, but a definite interval whose identity is generally clear from
the extralinguistic context, just as the identity of the he in [He shouldn’t be in
here] is clear from the context.”
Partee here is arguing that neither of the two plausible LFs derivable in our
current system correctly captures the meaning of (). Given that the sentence
contains a past tense (which we have treated as an existential quantifier over past
times) and a negation, we need to consider two possible scopings of the two
operators:

() a. PAST NEG I turn off the stove.

b. NEG PAST I turn off the stove.

E .: Show that neither LF in () captures the meaning of ()
correctly.

At this point, we will not develop Partee’s analysis in formally explicit detail. If
tenses refer to times, it would be easiest to give up on the treatment of times as
evaluation parameters and move to a system where times are object language
arguments of time-sensitive expressions. We will see a system of that nature later
on.
In a commentary on Partee’s paper at the same conference it was presented at,
Stalnaker pointed out that the Priorean theory can in fact deal with (), if
one allows the existential quantifier over times to be contextually restricted to
times in the salient interval of Partee leaving her house — since natural language
quantifiers are typically subject to contextual restrictions, this is not a problematic
assumption. (Note that Partee formulated her observation in quite a circumspect
way: “The sentence refers to a particular time”; Stalnaker’s suggestion is that the
reference to a particular time is part of the restriction to the quantifier over times
expressed by tense, rather than tense itself being a referring item (of type s).)
§. T N  I



E .: Assuming a restricted existential quantification à la Stalnaker,

which of the LFs in () captures the meaning of () correctly?

Ogihara () argues that the restricted existential quantification view is in fact
superior to Partee’s analysis, since Partee’s analysis needs an existential quantifier
anyway. Note that it is clear that the time being referred to is a protracted interval Partee  adopts an exis-
tential quantifier analysis.
(the time during which Partee was leaving her house). But the sentence is not
interpreted as saying that this interval is not a time at which she turned off her
stove, which would have to be a fairly absurd turning-off-of-the-stove (turning
off the stove only takes a moment and doesn’t take up a significant interval).
Instead, the sentence says that in that salient interval there is no time at which
she turned off the stove. Clearly, we do need an existential quantifier in there
somewhere and the Priorean theory provides one. Ogihara makes the point
with the following example:

() John:Did you see Mary?

Bill: Yes, I saw her, but I don’t remember exactly when.

The question and answer in this dialogue concern the issue of whether Bill saw
Mary at some time in a contextually salient interval.

. The Need for Intervals

We have just seen a reason to recognize that natural language can talk not just
about moments of time but also about intervals (connected sets of moments),
which is a fairly trivial fact; after all, what does the year  refer to if not an
interval of time? We have to go even farther, though. It can be shown that we
need the time parameter of the evaluation function to be able to be an interval.
Consider the tenseless clause John build a house and consider a situation where
John starts building a house (the only house he has ever built) on April , 
and finishes building it on April , . Now, which times do we want to be
times at which “John build a house” is true? If we allow the clause to be true
at moments during the building, we would make it true at other times during
the building (the ones after the first times) that John built a house, but that is
wrong. So, the time(s) at which “John build a house” cannot be before April ,
. And clearly, times after April ,  cannot be times at which “John build
a house” is true. So, perhaps, the only time at which “John build a house” is true

 Clearly, the alternative is to say that the existential quantifier is not expressed by tense but
comes from somewhere else, perhaps aspect, perhaps in the lexical meaning of turn off. We will
not pursue those options here.
 B  T  A C 



is the moment on April ,  when he finishes building the house? But then
we would incorrectly predict that on the day before, when he has already been
building the house for almost a year, we can truthfully say that John will build
a house. So, no moment of time can be the time at which “John build a house”
is true. The solution is that the time at which “John build a house” is true is
exactly the interval that starts with the first moment of the building project and
ends with the last nail hammered into the wall. Then, we can say before April ,
 that John will build a house and after April , , that John built a house.
What can we say during the building of the house, though? The English present
tense is not correctly used in this circumstance:

() !!John builds a house.

Our analysis may be read as predicting this fact. Assume that for an unembedded
clause, the time parameter is set to be the speech time. But what is the speech
time? Perhaps, it is the exact interval it takes to utter the particular clause being
evaluated. If so, an example like () can only possibly be true if the speech
interval exactly coincides with the reported event, here the building of the house.
That is, the speaker of () would have to ensure that she starts speaking at the
very first moment of John’s building the house, continues speaking rather slowly,
and then finishes speaking with the very last nail. It is intriguing to note that
sentences like () become acceptable in situations where a sentence is conceived
of as exactly coinciding the event being reported, namely play-by-play sports
commentary (“He passes the ball to Messi”).
What English needs to do instead is to use the progressive:

() John is building a house.

() expresses that the speech time is included in an interval of John building
a house. Elements that connect the evaluation time to the time at which a
predicate holds are usually called aspectual operators or simply aspects. The
English progressive then is an aspectual expression. We will look closer at its
meaning in a little while.

 We cannot go into this fascinating topic further here, but there is much more to explore about
the peculiar nature of (). Bennett & Partee () assume that the speech time is a moment
and use that assumption to derive the nature of (). Ejerhed () calls the typical use of (),
the “voyeur present”; see also Cooper .
§. A



. Aktionsarten
We can distinguish predicates with respect to their temporal profile. The tradi-
tional classification has four categories:

• accomplishment predicates
• achievement predicates
• activity predicates
• stative predicates

Accomplishment predicates (build a house, cross the street) describe an event that
has a defined beginning and end (telos, ‘goal’) and takes some amount of time to
finish. Achievement predicates (reach the summit, notice the problem) also have a
telos but are conceived of as describing an instantaneous event. Accomplishment
predicates and achievement predicates constitute the class of telic predicates.
Activity predicates (run, dance) describe events that are not conceived of as
having a defined goal. Stative predicates (be in New York, know French) describe
states that are true of intervals. The difference between activity predicates and
stative predicates is often said to turn on whether there is an agent being active
in the described event.

— Read Rothstein : Chapter , pp. – —

. The Progressive

— Read Portner  —

. Tense in Embedded Clauses

What happens to the time-sensitivity of the verb in a tenseless clause? Consider
ECM complements to verbs of believing:

() John believed it to be raining.

Evidently, there is some kind of dependency of the time reference in the lower
clause and the higher clause. The simplest approach in our framework would be
to have believe pass down its evaluation time to the lower clause and to assume
that the lower clause doesn’t have a tense operator. Then, whatever time believe
is being interpreted at would be the same time that the lower verb would be
evaluated at.
 B  T  A C 



() JbelieveKw,t = λphs,ti .λx. p(w 0 , t), for all worlds w 0 compatible with
what x believes in w at t.

Together with the rest of the system, we predict that () will be true iff there
is a past time t such that it is raining at t in all worlds which conform to what
John believes at t, which seems adequate. Unfortunately, it only seems adequate.
Consider these four worlds:

w rain at am, John awake at am

w rain at am, John awake at am

w rain at am, John awake at am

w rain at am, John awake at am

Assume that in all four worlds, John wakes up, has no idea what time it is, hears a
dripping noise, and says to himself “it is raining (now)”. Which worlds conform
to what John believes at am in w ? In which worlds is it raining at am? Are
the former a subset of the latter? No!
Consider a variant of the story. Everything is the same as above, except that
John wakes up, thinks it is am and says to himself: “It was raining at am.”
Fact: Sentence () is not a true report of John’s beliefs in w in this story. Why
not? There is a description, viz. am, which in fact picks out the time of John’s
thinking, and under which he ascribes rain to that time.
Conclusion: Sentence () unambiguously means that there is a past time t such
that John at t ascribes rain to t under the description “now”. We need to capture
this but the proposal encapsulated in () doesn’t achieve this.
The solution: believe (and other attitude verbs, or perhaps the complementizer
they select) controls not just the world parameter of its prejacent but also the
time parameter.

() JbelieveKw,t = λphs,ti .λx. p(w 0 , t 0 ), for all worlds w 0 and t 0 such that
for all that x can tell in w at t, x might be located in w 0 at t 0 .

On this analysis, () means essentially that John located himself at a raining
time. This is intuitively correct.

— More on tense in tensed complement clauses —

§. T  E C



Supplementary Readings
A nice and gentle introduction to some of the issues discussed in this chapter
comes from Ogihara:
Ogihara, Toshiyuki. . Tense and aspect in truth-conditional semantics.
Lingua (). –. doi:./j.lingua....
Partee’s seminal paper is a must read:
Partee, Barbara H. . Some structural analogies between tenses and pronouns
in English. The Journal of Philosophy (). –. doi:./.
Musan’s work on the pragmatic effects of tense:
Musan, Renate. . Tense, predicates, and lifetime effects. Natural Language
Semantics (). –. doi:./A:.
The three essential works on the progressive:
Dowty, David R. . Toward a semantic analysis of verb aspect and the
english ‘imperfective’ progressive. Linguistics and Philosophy (). –.
doi:./BF.
Landman, Fred. . The progressive. Natural Language Semantics (). –.
doi:./BF.
Portner, Paul. . The progressive in modal semantics. Language ().
–. doi:./.
The first chapter of Susan Rothstein’s book on lexical aspect gives a nice overview
of Aktionsarten/aspectual classes:
Rothstein, Susan. . Structuring events: A study in the semantics of lexi-
cal aspect Explorations in Semantics. Blackwell. URL http://tinyurl.com/
rothstein-aktionsarten, Chapter : “Verb Classes and Aspectual Classification”,
pp. –, available online at http://tinyurl.com/rothstein-aktionsarten.
Concise statements of some of the issues surrounding dependent tenses:
von Stechow, Arnim. . On the proper treatment of tense. Proceedings of
Semantics and Linguistic Theory . URL http://www.sfs.uni-tuebingen.de/
~arnim/Aufsaetze/SALT.pdf.
von Stechow, Arnim. . Tenses in compositional semantics. To be published
in Wolfgang Klein (ed) The Expression of Time in Language. URL http:
//www.sfs.uni-tuebingen.de/~arnim/Aufsaetze/Approaches.pdf.
— T     —
 C S
DP  S  M C

We discuss ambiguities that arise when DPs occur in modal contexts.

. De re vs. De dicto as a Scope Ambiguity 

. Raised subjects 
.. Examples of de dicto readings for raised subjects 
.. Syntactic “Reconstruction” 
.. Some Alternatives to Syntactic Reconstruction 

. De re vs. De dicto as a Scope Ambiguity

When a DP appears inside the clausal or VP complement of a modal predicate ,
there is often a so-called de re-de dicto ambiguity. A classic example is (),
which contains the DP a plumber inside the infinitive complement of want.

() John wants to marry a plumber.

According to the de dicto reading, every possible world in which John gets what
he wants is a world in which there is a plumber whom he marries. According
to the de re reading, there is a plumber in the actual world whom John marries
in every world in which he gets what he wants. We can imagine situations in
which one of the readings is true and the other one false.
For example, suppose John thinks that plumbers make ideal spouses, because
they can fix things around the house. He has never met one so far, but he

 We will be using the terms “modal operator” and “modal predicate” in their widest sense here,
to include modal auxiliaries (“modals”), modal main verbs and adjectives, attitude predicates,
and also modalizing sentence-adverbs like possibly.
 DP  S  M C C 



definitely wants to marry one. In this scenario, the de dicto reading is true, but
the de re reading is false. What all of John’s desire-worlds have in common is
that they have a plumber getting married to John in them. But it’s not the same
plumber in all those worlds. In fact, there is no particular individual (actual
plumber or other) whom he marries in every one of those worlds.
For a different scenario, suppose that John has fallen in love with Robin and
wants to marry Robin. Robin happens to be a plumber, but John doesn’t know
this; in fact, he wouldn’t like it and might even call off the engagement if he
found out. Here the de re reading is true, because there is an actual plumber,
viz. Robin, who gets married to John in every world in which he gets what he
wants. The de dicto reading is false, however, because the worlds which conform
to John’s wishes actually do not have him marrying a plumber in them. In his
favorite worlds, he marries Robin, who is not a plumber in those worlds.
When confronted with this second scenario, you might, with equal justification,
say ‘John wants to marry a plumber’, or ‘John doesn’t want to marry a plumber’.
Each can be taken in a way that makes it a true description of the facts – although,
of course, you cannot assert both in the same breath. This intuition fits well
with the idea that we are dealing with a genuine ambiguity.
Let’s look at another example:

() John believes that your abstract will be accepted.

Here the relevant DP in the complement clause of the verb believe is your abstract.
Again, we detect an ambiguity, which is brought to light by constructing different
scenarios.

 What is behind the Latin terminology “de re” (lit.: ‘of the thing’) and “de dicto” (lit.: ‘of what
is said’)? Apparently, the term “de dicto” is to indicate that on this reading, the words which
I, the speaker, am using to describe the attitude’s content, are the same (at least as far as the
relevant DP is concerned) as the words that the subject herself would use to express her attitude.
Indeed, if we asked the John in our example what he wants, then in the first scenario he’d say
“marry a plumber”, but in the second scenario he would not use these words. The term “de re”,
by contrast, indicates that there is a common object (here: Robin) whom I (the speaker) am
talking about when I say “a plumber” in my report and whom the attitude holder would be
referring to if he were to express his attitude in his own words. E.g., in our second scenario, John
might say that he wanted to marry “Robin”, or “this person here” (pointing at Robin). He’d
thus be referring to the same person that I am calling “a plumber”, but wouldn’t use that same
description.
Don’t take this “definition” of the terms too seriously, though! The terminology is much older
than any precise truth-conditional analysis of the two readings, and it does not, in hindsight,
make complete sense. We will also see below that there are cases where nobody is sure how to
apply the terms in the first place, even as purely descriptive labels. So in case of doubt, it is
always wiser to give a longer, more detailed, and less terminology-dependent description of the
relevant truth-conditional judgments.
§. De re . De dicto   S A



(i) John’s belief may be about an abstract that he reviewed, but since the
abstract is anonymous, he doesn’t know who wrote it. He told me that
there was a wonderful abstract about subjacency in Hindi that is sure to
be accepted. I know that it was your abstract and inform you of John’s
opinion by saying (). This is the de re reading. In the same situation, the
de dicto reading is false: Among John’s belief worlds, there are many worlds
in which your abstract will be accepted is not true or even false. For all he
knows, you might have written, for instance, that terrible abstract about
Antecedent-Contained Deletion, which he also reviewed and is positive
will be rejected.
(ii) For the other scenario, imagine that you are a famous linguist, and John
doesn’t have a very high opinion about the fairness of the abstract selection
process. He thinks that famous people never get rejected, however the
anonymous reviewers judge their submissions. He believes (correctly or
incorrectly – this doesn’t matter here) that you submitted a (unique) abstract.
He has no specific information or opinion about the abstract’s content and
quality, but given his general beliefs and his knowledge that you are famous,
he nevertheless believes that your abstract will be accepted. This is the de
dicto reading. Here it is true in all of John’s belief worlds that you submitted
a (unique) abstract and it will be accepted. The de re reading of (),
though, may well be false in this scenario. Suppose – to flesh it out further
– the abstract you actually submitted is that terrible one about ACD. That
one surely doesn’t get accepted in every one of John’s belief worlds. There
may be some where it gets in (unless John is certain it can’t be by anyone
famous, he has to allow at least the possibility that it will get in despite its
low quality). But there are definitely also belief-worlds of his in which it
doesn’t get accepted.
We have taken care here to construct scenarios that make one of the
readings true and the other false. This establishes the existence of two
distinct readings. We should note, however, that there are also many
possible and natural scenarios that simultaneously support the truth of both
readings. Consider, for instance, the following third scenario for sentence
().
(iii) John is your adviser and is fully convinced that your abstract will be
accepted, since he knows it and in fact helped you when you were writing
it. This is the sort of situation in which both the de dicto and the de re
reading are true. It is true, on the one hand, that the sentence your abstract
will be accepted is true in every one of John’s belief worlds (de dicto reading).
And on the other hand, if we ask whether the abstract which you actually
wrote will get accepted in each of John’s belief worlds, that is likewise true
(de re reading).
In fact, this kind of “doubly verifying” scenario is very common when
we look at actual uses of attitude sentences in ordinary conversation. There
 DP  S  M C C 



may even be many cases where communication proceeds smoothly without

either the speaker or the hearer making up their minds as to which of
the two readings they intend or understand. It doesn’t matter, since the
possible circumstances in which their truth-values would differ are unlikely
and ignorable anyway. Still, we can conjure up scenarios in which the two
readings come apart, and our intuitions about those scenarios do support
the existence of a semantic ambiguity.

In the paraphrases by which we have elucidated the two readings of our examples,
we have already given away the essential idea of the analysis that we will adopt:
We will treat de dicto-de re ambiguities as ambiguities of scope. The de dicto
readings, it turns out, are the ones which we predict without further ado if
we assume that the position of the DP at LF is within the modal predicate’s
complement. (That is, it is either in situ or QRed within the complement clause.)
For example:

() John wants [ [ a plumber] [  to marry t ]]

() John believes [ the abstract-by-you will-be-accepted]

To obtain the de re readings, we apparently have to QR the DP to a position

above the modal predicate, minimally the VP headed by want or believe.

() [ a plumber] [ John wants [  to marry t ]]

() [ the abstract-by-you] [ John believes will-be-accepted]]
E .: Calculate the interpretations of the four structures in ()–(),
and determine their predicted truth-values in each of the (types of ) possible
worlds that we described above in our introduction to the ambiguity.
Some assumptions to make the job easier: (i) Assume that () and () are
evaluated with respect to a variable assignment that assigns John to the number
. This assumption takes the place of a worked out theory of how controlled
PRO is interpreted. (ii) Assume that abstract-by-you is an unanalyzed one-place
predicate. This takes the place of a worked out theory of how genitives with a
non-possessive meaning are to be analyzed. 

. Raised subjects

In the examples of de re-de dicto ambiguities that we have looked at so far, the
surface position of the DP in question was inside the modal predicate’s clausal or
VP-complement. We saw that if it stays there at LF, a de dicto reading results,
§. R 



and if it covertly moves up above the modal operator, we get a de re reading. In

the present section, we will look at cases in which a DP that is superficially higher
than a modal operator can still be read de dicto. In these cases, it is the de re
reading which we obtain if the LF looks essentially like the surface structure, and
the de dicto reading for which we apparently have to posit a non-trivial covert
derivation.

.. Examples of de dicto readings for raised subjects

Suppose I come to my office one morning and find the papers and books on my
desk in different locations than I remember leaving them the night before. I say:

() Somebody must have been here (since last night).

On the assumptions we have been making, somebody is base-generated as the

subject of the VP be here and then moved to its surface position above the modal.
So () has the following S-structure, which is also an interpretable LF.

() somebody [ λ [ [ must R] [ t have-been-here]]]

What does () mean? The appropriate reading  for must here is epistemic, so
suppose the variableR is mapped to the relation λw.λw . w 0 is compatible with
0

what I believe in w . Let w be the utterance world. Then the truth-condtion

calculated by our rules is as follows.

() ∃x[x is a person in w &

∀w 0 [w 0 is compatible with what I believe in w → x was here in w 0 ]]

But this is not the intended meaning. For () to be true, there has to be a
person who in every world compatible with what I believe was in my office. In
other words, all my belief-worlds have to have one and the same person coming
to my office. But this is not what you intuitively understood me to be saying
about my belief-state when I said (). The context we described suggests that I
do not know (or have any opinion about) which person it was that was in my
office. For all I know, it might have been John, or it might have been Mary, or it
have been this stranger here, or that stranger there. In each of my belief-worlds,
somebody or other was in my office, but no one person was there in all of them.
I do not believe of anyone in particular that he or she was there, and you did not
understand me to be saying so when I uttered (). What you did understand
me to be claiming, apparently, was not () but ().
 DP  S  M C C 



() ∀w 0 [w 0 is compatible with what I believe in w

→ ∃x [x is a person in w 0 & x was here in w 0 ]]

In other words – to use the terminology we introduced in the last section – the
DP somebody in () appears to have a de dicto reading.
How can sentence () have the meaning in ()? The LF in (), as we saw,
means something else; it expresses a de re reading, which typically is false when
() is uttered sincerely. So there must be another LF. What does it look like
and how is it derived? One way to capture the intended reading, it seems, would
be to generate an LF that’s essentially the same as the underlying structure we
posited for (), i.e., the structure before the subject has raised:

() [IP e [I 0 [ must R] [ somebody have-been-here]]]

() means precisely () (assuming that the unfilled Spec-of-IP position is
semantically vacuous), as you can verify by calculating its interpretation by our
rules. So is () (one of ) the LF(s) for (), and what assumption about syntax
allow it to be generated? Or are there other – perhaps less obvious, but easier to
generate – candidates for the de dicto LF-structure of ()?
Before we get into these question, let’s look at a few more examples. Each of
the following sentences, we claim, has a de dicto reading for the subject, as given
in the accompanying formula. The modal operators in the examples are of a
variety of syntactic types, including modal auxiliaries, main verbs, adjectives, and
adverbs.

() Everyone in the class may have received an A.

∃w 0 [w 0 conforms to what I believe in w &
∀x[x is in this class in w 0 → x received an A in w 0 ]].

() At least two semanticists have to be invited.

∀w 0 [w 0 conforms to what is desirable in w
→ ∃ x [x is a semanticist in w 0 & x is invited in w 0 ]].

() Somebody from New York is expected to win the lottery.

∀w 0 [w 0 conforms to what is expected in w
→ ∃x[x is a person from NY in w 0 & x wins the lottery in w 0 ]]

() Somebody from New York is likely to win the lottery.

∀w 0 [w 0 is as likely as any other world, given I know in w
→ ∃x[x is a person from NY in w 0 & x wins the lottery in w 0 ]]

 Hopefully the exact analysis of the modal operators likely and probably is not too crucial for
§. R 



() One of these two people is probably infected.

∀w 0 [w 0 is as likely as any other world, given what I know in w
→ ∃x[x is one of these two people & x is in infected in w 0 ]]

To bring out the intended de dicto reading of the last example (to pick just one)
imagine this scenario: We are tracking a dangerous virus infection and have
sampled blood from two particular patients. Unfortunately, we were sloppy and
the blood samples ended up all mixed up in one container. The virus count is
high enough to make it quite probable that one of the patients is infected but
because of the mix-up we have no evidence about which one of them it may
be. In this scenario, () appears to be true. It would not be true under a de re
reading, because neither one of the two people is infected in every one of the
likely worlds.
A word of clarification about our empirical claim: We have been concentrating
on the observation that de dicto readings are available, but have not addressed
the question whether they are the only available readings or coexist with equally
possible de re readings. Indeed, some of the sentences in our list appear to be
ambiguous: For example, it seems that () could also be understood to claim

the present discussion, but you may still be wondering about it. As you see in our formula,
we are thinking of likely (probably) as a kind of epistemic necessity operator, i.e., a universal
quantifier over a set of worlds that is somehow determined by the speaker’s knowledge. (We are
focussing on the “subjective probability” sense of these words. Perhaps there is a also an “objective
probability” reading that is circumstantial rather than epistemic.) What is the difference then
between likely and e.g. epistemic must (or necessary or I believe that)? Intuitively, ‘it is likely that
p’ makes a weaker claim than ‘it must be the case that p’. If both are universal quantifiers, then,
it appears that likely is quantifying over a smaller set than must, i.e., over only a proper subset of
the worlds that are compatible with what I believe. The difference concerns those worlds that
I cannot strictly rule out but regard as remote possibilities. These worlds are included in the
domain for must, but not in the one for likely. For example, if there was a race between John and
Mary, and I am willing to bet that Mary won but am not completely sure she did, then those
worlds where John won are remote possibilities for me. They are included in the domain of must,
and so I will not say that Mary must have won, but they are not in the domain quantified over
by likely, so I do say that Mary is likely to have won.
This is only a very crude approximation, of course. For one thing, probability is a gradable
notion. Some things are more probable than others, and where we draw the line between
what’s probable and what isn’t is a vague or context-dependent matter. Even must, necessary etc.
arguably don’t really express complete certainty (because in practice there is hardly anything we
are completely certain of ), but rather just a very high degree of probability. For more discussion
of likely, necessary, and other graded modal concepts in a possible worlds semantics, see e.g.
Kratzer , .
A different approach may be that likely quantifies over the same set of worlds as must, but
with a weaker, less than universal, quantificational force. I.e., ‘it is likely that p’ means something
like p is true in most of the worlds conforming to what I know. A prima facie problem with this
idea is that presumably every proposition is true in infinitely many possible worlds, so how can
we make sense of cardinal notions like ‘more’ and ‘most’ here? But perhaps this can be worked
out somehow.
 DP  S  M C C 



that there is a particular New Yorker who is likely to win (e.g., because he has
bribed everybody). Others arguably are not ambiguous and can only be read de
dicto. This is what von Fintel & Iatridou () claim about sentences like ().
They note that if () also allowed a de re reading, it should be possible to make
coherent sense of ().

() Everyone in the class may have received an A. But not everybody did.

In fact, () sounds contradictory, which they show is explained if only the
de dicto reading is permitted by the grammar. They conjecture that this is a
systematic property of epistemic modal operators (as opposed to deontic and
other types of modalities). Epistemic operators always have widest scope in their
sentence.
So there are really two challenges here for our current theory. We need to account
for the existence of de dicto readings, and also for the absence, in at least some of
our examples, of de re readings. We will be concerned here exclusively with the
first challenge and will set the second aside. We will aim, in effect, to set up the
system so that all sentences of this type are in principle ambiguous, hoping that
additional constraints that we are not investigating here will kick in to exclude
the de re readings where they are missing.
To complicate the empirical picture further, there are also examples where raised
subjects are unambiguously de re. Such cases have been around in the syntactic
literature for a while, and they have recently received renewed attention in the
work of Lasnik and others. To illustrate just one of the systematic restrictions,
negative quantifiers like nobody seem to permit only surface scope (i.e., wide
scope) with respect to a modal verb or adjective they have raised over.

() Nobody from New York is likely to win the lottery.

() does not have a de dicto reading parallel to the one for () above, i.e., it
cannot mean that it is likely that nobody from NY will win. It can only mean
that there is nobody from NY who is likely to win. This too is an issue that we
set aside.
In the next couple of sections, all that we are trying to do is find and justify
a mechanism by which the grammar is capable to generate both de re and de
dicto readings for subjects that have raised over modal operators. It is quite
conceivable, of course, that the nature of the additional constraints which often
exclude one reading or the other is ultimately relevant to this discussion and
that a better understanding of them may undermine our conclusions. But this is
something we must leave for further research.
§. R 



.. Syntactic “Reconstruction”

Given that the de dicto reading of () we are aiming to generate is equivalent to
the formula in (), an obvious idea is that there is an LF which is essentially the
pre-movement structure of this sentence, i.e., the structure prior to the raising of
the subject above the operator. There are a number of ways to make such an LF
available.
One option, most recently defended in Sauerland & Elbourne (), is to
assume that the raising of the subject can happen in a part of the derivation which
only feeds PF, not LF. In that case, the subject simply stays in its underlying
VP-internal position throughout the derivation from DS to LF. (Recall that
quantifiers are interpretable there, as they generally are in subject positions.)
Another option is a version of the so-called Copy Theory of movement intro-
duced in Chomsky (). This assumes that movement generally proceeds in
two separate steps, rather than as a single complex operation as we have assumed
so far. Recall that in H& K, it was stipulated that every movement effects the
following four changes:
(i) a phrase α is deleted,
(ii) an index i is attached to the resulting empty node (making it a so-called
trace, which the semantic rule for “Pronouns and Traces” recognizes as a
variable),
(iii) a new copy of α is created somewhere else in the tree (at the “landing site”),
and
(iv) the sister-constituent of this new copy gets another instance of the index i
adjoined to it (which the semantic rule of Predicate Abstraction recognizes
as a binder index).
If we adopt the Copy Theory, we assume instead that there are three distinct
operations:
“Copy”: Create a new copy of α somewhere in the tree, attach an index i to the
original α , and adjoin another instance of i to the sister of the new copy
of α . (= steps (ii), (iii), and (iv) above)
“Delete Lower Copy”: Delete the original α . (= step (i) above)
“Delete Upper Copy”: Delete the new copy of α and both instances of i.
The Copy operation is part of every movement operation, and can happen
anywhere in the syntactic derivation. The Delete operations happen at the end
of the LF derivation and at the end of the PF deletion. We have a choice of
applying either Delete Lower Copy or Delete Upper Copy to each pair of copies,
and we can make this choice independently at LF and at PF. (E.g., we can do
Copy in the common part of the derivation and than Delete Lower Copy at
LF and Delete Upper Copy at PF.) If we always choose Delete Lower Copy at
LF, this system generates exactly the same structures and interpretations as the
 DP  S  M C C 



one from H& K. But if we exercise the Delete Upper Copy option at LF, we are
effectively undoing previous movements, and this gives us LFs with potentially
new interpretations. In the application we are interested in here, we would apply
the Copy step of subject raising before the derivation branches, and then choose
Delete Lower Copy at PF but Delete Upper Copy at LF. The LF will thus look
as if the raising never happened, and it will straightforwardly get the desired de
dicto reading.
If the choice between the two Delete operations is generally optional, we in
principle predict ambiguity wherever there has been movement. Notice, however,
first, that the two structures will often be truth-conditionally equivalent (e.g.
when the moved phrase is a name), and second, that they will not always be both
interpretable. (E.g., if we chose Delete Upper Copy after QRing a quantifier
from object position, we’d get an uninterpretable structure, and so this option
is automatically ruled out.) Even so, we predict lots of ambiguity. Specifically,
since raised subjects are always interpretable in both their underlying and raised
locations, we predict all raising structures where a quantificational DP has raised
over a modal operator (or over negation or a temporal operator) to be ambiguous.
As we have already mentioned, this is not factually correct, and so there must be
various further constraints that somehow restrict the choices. (Similar comments
apply, of course, to the option we mentioned first, of applying raising only on
the PF-branch.)
Yet another solution was first proposed by May (a): May assumed that QR
could in principle apply in a “downward” fashion, i.e., it could adjoin the moved
phrase to a node that doesn’t contain its trace. Exercising this option with a
raised subject would let us produce the following structure, where the subject
has first raised over the modal and then QRed below it.

() tj λi [ must-R [ someone λj [ ti have been here]]]

As it stands, this structure contains at least one free variable (the trace t j ) and
can therefore not possibly represent any actual reading of this sentence. May
further assumes that traces can in principle be deleted, when their presence is
not required for interpretability. This is not yet quite enough, though to make
() interpretable, at least not within our framework of assumptions, for () is
still not a candidate for an actual reading of ().

() λi [ must-R [ someone λj [ ti have been here]]]

We would need to assume further that the topmost binder index could be deleted
along with the unbound trace, and also that the indices i and j can be the same,
so that the raising trace t j is bound by the binding-index created by QR. If these
§. R 



things can be properly worked out somehow, then this is another way to generate
the de dicto reading. Notice that the LF is not exactly the same as on the previous
two approaches, since the subject ends up in an adjoined position rather than in
its original argument position, but this difference is obviously without semantic
import.
What all of these approaches have in common is that they place the burden of
generating the de dicto reading for raised subjects on the syntactic derivation.
Somehow or other, they all wind up with structures in which the subject is
lower than it is on the surface and thereby falls within the scope of the modal
operator. They also have in common that they take the modal operator (here the
auxiliary, in other cases a main predicate or an adverb) to be staying put. I.e.,
they assume that the de dicto readings are not due to the modal operator being
covertly higher than it seems to be, but to the subject being lower. Approaches
with these features will be said to appeal to “syntactic reconstruction” of the
subject.

.. Some Alternatives to Syntactic Reconstruction

Besides (some version of ) syntactic reconstruction, there are many other ways in
which one try to generate de dicto readings for raised subjects. Here are some
other possibilities that have been suggested and or readily come to mind. We
will see that some of them yield exactly the de dicto reading as we have been
describing it so far, whereas others yield a reading that is very similar but not
quite the same. We will confine ourselves to analyses which involve no or only
minor changes to our system of syntactic and semantic assumptions. Obviously,
if departed from these further, there would be even more different options, but
even so, there seem to be quite a few.

. R   ,  :   Conceivably, an LF for

the de dicto reading of () might be derived from the S-structure (=()) by
covertly moving must (and its covert R-argument) up above the subject. This
would have to be a movement which leaves no (semantically non-vacuous) trace.
Given our inventory of composition rules, the only type that the trace could
have to make the structure containing it interpretable would be the type of the
moved operator itself (i.e. hst, ti). If it had that type, however, the movement
would be semantically inconsequential, i.e., the structure would mean exactly

 This is a very broad notion of “reconstruction”, where basically any mechanism which puts a
phrase at LF in a location nearer to its underlying site than its surface site is called “reconstruction”.
In some of the literature, the term is used more narrowly. For example, May’s downward QR
is sometimes explicitly contrasted with genuine reconstruction, since it places the quantifier
somewhere else than exactly where it has moved from.
 DP  S  M C C 



the same as (). So this would not be a way to provide an LF for the de dicto
reading. If there was no trace left however (and also no binder index introduced),
we indeed would obtain the de dicto reading.
E .: Prove the claims we just made in the previous paragraph. Why
is no type for the trace other than hst, ti possible? Why is the movement
semantically inert when this type is chosen? How does the correct intended
meaning arise if there is no trace and binder index? 

. R   ,  :     [Requires slightly

modified inventory of composition rules. Derives an interpretation that is not
quite the same as the de dicto reading we have assumed so far. Rather, it is a
“narrow-Q, R-de-re” interpretation in the sense of Section ?? below.]

. H     ,  :  het, ti [Before reading
this section, read and do the exercise on p./ in H& K]

So far in our discussion, we have taken for granted that the LF which corresponds
to the surface structure, viz. (), gives us the de re reading. This, however, is
correct only on the tacit assumption that the trace of raising is a variable of type
e. If it is part of our general theory that all variables, or at least all interpretable
binder indices (hence all bound variables), in our LFs are of type e, then there
is nothing more here to say. But it is not prima facie obvious that we must or
should make this general assumption, and if we don’t, then the tree in () is not
really one single LF, but the common structure for many different ones, which
differ in the type chosen for the trace. Most of the infinitely many semantic
types we might assign to this trace will lead to uninterpretable structures, but
there turns out to be one other choice besides e that works, namely het, ti:

() somebody λ,het,ti [ [ must R] [ t,het,ti have-been-here]]

() is interpretable in our system, but again, as above, the predicted interpreta-
tion is not exactly the de dicto reading as we have been describing it so far, but a
“narrow-Q, R-de-re” reading.
E .: Using higher-type traces to “reverse” syntactic scope-relation is a
trick which can be used quite generally. It is useful to look at a non-intensional
example as a first illustration. () contains a universal quantifier and a negation,
and it is scopally ambiguous between the readings in (a) and (b).

() Everything that glitters is not gold.

a. ∀x[x glitters → ¬x is gold] “surface scope”
b. ¬∀x[x glitters → x is gold] “inverse scope”
§. R 



We could derive the inverse scope reading for () by generating an LF (e.g. by
some version of syntactic reconstruction") in which the every-DP is below not.
Interestingly, however, we can also derive this reading if the every-DP is in its
raised position above not but its trace has the type hhe, ti, ti.
Spell out this analysis. (I.e., draw the LF and show how the inverse-scope
interpretation is calculated by our semantic rules.) 

E .: Convince yourself that there are no other types for the raising
trace besides e and het, ti that would make the structure in () interpretable.
(At least not if we stick exactly to our current composition rules.) 

. H     ,  :  hs, het, tii If we want
to get exactly the de dicto reading that results from syntactic reconstruction out
of a surface-like LF of the form (), we must use an even higher type for the
raising trace, namely hs, hhe, ti, tii, the type of the intension of a quantifier. As
you just proved in the exercise, this is not possible if we stick to exactly the
composition rules that we have currently available. The problem is in the VP:
the trace in subject position is of type hs, hhe, ti, tii and its sister is of type he, ti.
These two connot combine by either FA or IFA, but it works if we employ
another variant of functional application.

() Extensionalizing Functional Application (EFA)

If α is a branching node and {β, γ} the set of its daughters, then, for any
world w and assignment g:
if JβKw,g (w) is a function whose domain contains JγKw,g ,
then JαKw,g = JβKw,g (w)(JγKw,g ).

E .: Calculate the truth-conditions of () under the assumption that
the trace of the subject quantifier is of type hs, hhe, ti, tii. 

C      ? Two of the methods we tried
derived readings in which the raised subject’s quantificational determiner took

 Notice that the problem here is kind of the mirror image of the problem that led to the
introduction of “Intensional Functional Application” in H& K, ch. . There, we had a function
looking for an argument of type hs, ti, but the sister node had an extension of type t. IFA allowed
us to, in effect, construct an argument with an added “s” in its type. This time around, we have
to get rid of an “s” rather than adding one; and this is what EFA accomplishes.
So we now have three different “functional application”-type rules altogether in our system:
0
ordinary FA simply applies JβKw to JγKw ; IFA applies JβKw to λw 0 .JγKw ; and EFA applies
JβKw (w) to JγKw . At most one of them will be applicable to each given branching node,
depending on the type of JγKw .
Think about the situation. Might there be other variant functional application rules?
 DP  S  M C C 



scope below the world-quantifier in the modal operator, but the raised subject’s
restricting NP still was evaluated in the utterance world (or the evaluation world
for the larger sentence, whichever that may be). It is difficult to assess whether
these readings are actually available for the sentences under consideration, and
we will postpone this question to a later section. We would like to argue here,
however, that even if these readings are available, they cannot be the only readings
that are available for raised subjects besides their wide-scope readings. In other
words, even if we allowed one of the mechanisms that generated these sort of
hybrid readings, we would still need another mechanism that gives us, for at
least some examples, the “real” de dicto readings that we obtain e.g. by syntactic
reconstruction. The relevant examples that show this most clearly involve DPs
with more descriptive content than somebody and whose NPs express clearly
contingent properties.

() A neat-freak must have been here.

If I say this instead of our original () when I come to my office in the morning
and interpret the clues on my desk, I am saying that every world compatible
with my beliefs is such that someone who is a neat-freak in that world was here
in that world. Suppose there is a guy, Bill, whom I know slightly but not well
enough to have an opinion on whether or not he is neat. He may or not be, for
all I know. So there are worlds among my belief worlds where he is a neat-freak
and worlds where he is not. I also don’t have an opinion on whether he was or
wasn’t the one who came into my office last night. He did in some of my belief
worlds and he didn’t in others. I am implying with (), however, that if Bill
isn’t a neat-freak, then it wasn’t him in my office. I.e., () is telling you that,
even if I have belief-worlds in which Bill is a slob and I have belief-worlds in
which (only) he was in my office, I do not have any belief-worlds in which Bill is
a slob and the only person who was in my office. This is correctly predicted if
() expresses the “genuine” de dicto reading in (), but not if it expresses the
“hybrid” reading in ().

() ∀w 0 [w 0 is compatible with what I believe in w →

∃x[x is a neatfreak in w 0 and x was here in w 0 ]]

() ∀w 0 [w 0 is compatible with what I believe in w →

∃x[x is a neatfreak in w and x was here in w 0 ]]

We therefore conclude the mechanisms  and  considered above (whatever

there merits otherwise) cannot supplant syntactic reconstruction or some other
mechanism that yields readings like ().
This leaves only the first and fourth options that we looked at as potential com-
petitors to syntactic reconstruction, and we will focus the rest of the discussion
on how we might be able to tease apart the predictions that these mechanisms
imply from the ones of a syntactic reconstruction approach.
§. R 



As for moving the modal operator, there are no direct bad predictions that we are
aware of with this. But it leads us to expect that we might find not only scope
ambiguities involving a modal operator and a DP, but also scope ambiguities
between two modal operators, since one of them might covertly move over the
other. It seems that this never happens. Sentences with stacked modal verbs
seem to be unambiguous and show only those readings where the scopes of the
operators reflect their surface hierarchy.

() a. I have to be allowed to graduate.

b. #I am allowed to have to graduate.

Of course, this might be explained by appropriate constraints on the movement

of modal operators, and such constraints may even come for free in a the right
synatctic theory. Also, we should have a much more comprehensive investigation
of the empirical facts before we reach any verdict. If it is true, however, that
modal operators only engage in scope interaction with DPs and never with each
other, then a theory which does not allow any movement of modals at all could
claim the advantage of having a simple and principled explanation for this fact.
What about the “semantic reconstruction” option, where raised subjects can leave
traces of type hs, het, tii and thus get narrow scope semantically without ending
up low syntactically? This type of approach has been explored quite thoroughly
and defended with great sophistication. We can only sketch the main objections
to it here and must leave it to the reader to consult the literature for an informed
opinion.

S   C C An example from Fox () (build-

ing on Lebeaux  and Heycock ):

() a. A student of his seems to David to be at the party.

OK
de re, OK de dicto
b. A student of David’s seems to him to be at the party.
OK
de re, *de dicto

Sketch of argument: If Cond. C is formulated in terms of c-command relations

and applies at LF, it will distinguish between de re and de dicto readings only if
those involve LFs with different hierarchical relations.

R   

() The cat seems to be out of the bag.

() ?Advantage might have been taken of them.
 DP  S  M C C 



Sketch of argument: If idioms must be constituents at LF in order to receive

their idiomatic interpretations, these cases call for syntactic reconstruction. An
additional mechanism of semantic reconstruction via high-type traces is then at
best redundant.

Tentative conclusion: Syntactic reconstruction (some version of it) provides the

best account of de dicto readings for raised subjects.
 C E
B de re — de dicto : T T R

In this chapter, we will see that quantificational noun phrases in the

scope of a modal operator can receive a reading where their restrictive
predicate is not interpreted in the worlds introduced by the modal
operator (which is what happens in de re readings as well) while at
the same time their quantificational force takes scope below the modal
operator (which is what happens in de dicto readings as well). This
seemingly paradoxical situation might force whole-sale revisions to our
architecture. We discuss the standard solution (which involves supplying
predicates with world-arguments) and some alternatives.

. A Problem: Additional Readings and Scope Paradoxes 

. The Standard Solution: Overt World Variables 
.. Semantic Values 
.. Lexical entries 
.. Composition Rules 
.. Syntax 
.. The Need for a Binding Theory for World Vari-
ables 
.. Two Kinds of World Pronouns 
.. Excursus: Semantic reconstruction for de dicto
raised subjects? 
. Alternatives to Overt World Variables 
.. Indexed Operators 
.. Scoping After All? 
. Scope, Restrictors, and the Syntax of Movement 
. A Recurring Theme: Historical Overview 
 B de re — de dicto : T T R C 



. A Problem: Additional Readings and Scope

Paradoxes
Janet Dean Fodor discussed examples like () in her dissertation ().

() Mary wanted to buy a hat just like mine.

Fodor observes that () has three readings, which she labels “specific de re,”
“non-specific de re,” and “non-specific de dicto.”

(i) On the “specific de re” reading, the sentence says that there is a particular
hat which is just like mine such that Mary has a desire to buy it. Say, I
am walking along Newbury Street with Mary. Mary sees a hat in a display
window and wants to buy it. She tells me so. I don’t reveal that I have one
just like it. But later I tell you by uttering ().
(ii) On the “non-specific de dicto” reading, the sentence says that Mary’s desire
was to buy some hat or other which fulfills the description that it is just
like mine. She is a copycat.
(iii) On the “non-specific de re” reading, finally, the sentence will be true, e.g.,
in the following situation: Mary’s desire is to buy some hat or other, and
the only important thing is that it be a Red Sox cap. Unbeknownst to her,
my hat is one of those as well.

The existence of three different readings appears to be problematic for the scopal
account of de re-de dicto ambiguities that we have been assuming. It seems that
our analysis allows just two semantically distinct types of LFs: Either the DP
a hat just like mine takes scope below want, as in (), or it takes scope above
want, as in ().

() Mary wanted [ [a hat-just-like-mine] [  to buy t ]]

() [a hat-just-like-mine] [ Mary wanted [  to buy t ]]

In the system we have developed so far, () says that in every world w 0 in
which Mary gets what she wants, there is something that she buys in w 0 that’s a
hat in w 0 and like my hat in w 0 . This is Fodor’s “non-specific de dicto” reading.
(), on the other hand, says that there is some thing x which is a hat in the
actual world and like my hat in the actual world, and Mary buys x in every
one of her desire worlds. That is Fodor’s “specific de re.” But what about the
“non-specific de re”? To obtain this reading, it seems that we would have to
evaluate the predicate hat just like mine in the actual world, so as to obtain its
actual extension (in the scenario we have sketched, the set of all Red Sox caps).
§. A P: A R  S P



But the existential quantifier expressed by the indefinite article in the hat-DP
should not take scope over the modal operator want, but below it, so that we can
account for the fact that in different desire-worlds of Mary’s, she buys possibly
different hats.
There is a tension here: one aspect of the truth-conditions of this reading
suggests that the DP a hat just like mine should be outside of the scope of want,
but another aspect of these truth-conditions compels us to place it inside the
scope of want. We can’t have it both ways, it would seem, which is why this has
been called a “scope paradox”
Another example of this sort, due to Bäuerle (), is ():

() Georg believes that a woman from Stuttgart loves every member of the
VfB team.

Bäuerle describes the following scenario: Georg has seen a group of men on
the bus. This group happens to be the VfB team (Stuttgart’s soccer team), but
Georg does not know this. Georg also believes (Bäuerle doesn’t spell out on
what grounds) that there is some woman from Stuttgart who loves every one
of these men. There is no particular woman of whom he believes that, so there
are different such women in his different belief-worlds. Bäuerle notes that ()
can be understood as true in this scenario. But there is a problem in finding
an appropriate LF that will predict its truth here. First, since there are different
women in different belief-worlds of Georg’s, the existential quantifier a woman
from Stuttgart must be inside the scope of believe. Second, since (in each belief
world) there aren’t different women that love each of the men, but one that loves
them all, the a-DP should take scope over the every-DP. If the every-DP is in
the scope of the a-DP, and the a-DP is in the scope of believe, then it follows
that the every-DP is in the scope of believe. But on the other hand, if we want to
capture the fact that the men in question need not be VfB-members in Georg’s
belief-worlds, the predicate member of the VfB team needs to be outside of the
scope of believe. Again, we have a “scope paradox”.
Before we turn to possible solutions for this problem, let’s have one more
example:

() Mary hopes that a friend of mine will win the race.

This again seems to have three readings. In Fodor’s terminology, the DP a friend
of mine can be “non-specific de dicto,” in which case () is true iff in every
world where Mary’s hopes come true, there is somebody who is my friend and
wins. It can also have a “specific de re” reading: Mary wants John to win, she
doesn’t know John is my friend, but I can still report her hope as in (). But
 B de re — de dicto : T T R C 



there is a third option, the “non-specific de re” reading. To bring out this rather
exotic reading, imagine this: Mary looks at the ten contestants and says I hope
one of the three on the right wins - they are so shaggy - I like shaggy people. She
doesn’t know that those are my friends. But I could still report her hope as in
().

. The Standard Solution: Overt World Variables

The scope paradoxes we have encountered can be traced back to a basic design
feature of our system of intensional semantics: the relevant “evaluation world”
for each predicate in a sentence is strictly determined by its LF-position. All
predicates that occur in the (immediate) scope of the same modal operator must
be evaluated in the same possible worlds. E.g. if the scope of want consists of
the clause a friend of mine (to) win, then every desire-world w 0 will be required
to contain an individual that wins in w 0 and is also my friend in w 0 . If we want
to quantify over individuals that are my friends in the actual world (and not
necessarily in all the subject’s desire worlds), we have no choice but to place
friend of mine outside of the scope of want. And if we want to accomplish this
by means of QR, we must move the entire DP a friend of mine.
Not every kind of intensional semantics constrains our options in this way. One
way to visualize what we might want is to write down an LF that looks promising:

() Mary wantedw [λw 0 [ a hat-just-like-mine w ]λx [  to buyw 0 x ]]

We have annotated each predicate with the world in which we wish to evaluate
it. w is the evaluation world for the entire sentence and it is the world in
which we evaluate the predicates want and hat-just-like-mine. The embedded
sentence contributes a function from worlds to truth-values and we insert an
explicit λ-operator binding the world where the predicate buy is evaluated. The
crucial aspect of () is that the world in which hat-just-like-mine is evaluated is
the matrix evaluation world and not the same world in which its clause-mate
predicate buy is evaluated. This LF thus looks like it might faithfully capture
Fodor’s third reading.
Logical forms with overt world variables such as () are in fact the standard
solution to the problem presented by the third reading. Let us spell out some of
the technicalities. Later, we will consider a couple of alternatives.
§. T S S: O W V



.. Semantic Values

In this new system, we do not relativize the interpretation function to a possible
world. As in the old extensional system, the basic notion is just “JαK,” i.e., “the
semantic value of α”. (Or “JαKg ,” “the semantic value of α under assignment
g”, if α contains free variables.) However, semantic values are no longer always
extensions; some of them still are, but others are intensions. Here are some
representative examples of the types of semantic values for various kinds of words.

.. Lexical entries

() a. JsmartK = λw ∈ Ds . λx ∈ De . x is smart in w
b. JlikesK = λw ∈ Ds . λx ∈ De . λy ∈ De . y likes x in w
c. JteacherK = λw ∈ Ds . λx ∈ De . x is a teacher in w
d. JfriendK = λw ∈ Ds . λx ∈ De . λy ∈ De . y is x’s friend in w

() a. JbelieveK = λw ∈ Ds . λp ∈ Dhs,ti . λx ∈ D.

∀w 0 [w 0 conforms to what x believes in w → p(w 0 ) = ]
b. JmustK = λw ∈ Ds . λR ∈ Dhs,sti . λp ∈ Dhs,ti .
∀w 0 [R(w)(w 0 ) =  → p(w 0 ) = ]

() a. JAnnK =Ann

b. JandK = λu ∈ Dt . [λv ∈ Dt . u = v = ]
c. JtheK = λf ∈ Dhe,ti : ∃!x. f(x) = . the y such that f(y) = .
d. JeveryK = λf ∈ Dhe,ti . λg ∈ Dhe,ti . ∀x[f(x) =  → g(x) = ]

The entries in () (for words whose extensions are constant across worlds)
have stayed the same; their semantic values are still extensions. But the ones for
predicates (ordinary ones and modal ones) in () and () have changed; these
items now have as their semantic values what used to be their intensions.

.. Composition Rules

We abolish the special rule of Intensional Functional Application (IFA) and
go back to our old inventory of Functional Application, λ-Abstraction, and
Predicate Modification .

 We also abolish the Extensional Functional Application rule (EFA), if we had that one (see
section .. “Semantic Reconstruction”).
 Actually, PM requires a slightly revised formulation: Jα βKg = λw ∈ Ds . λx ∈
De . JαKg (w)(x) = JβKg (w)(x) = . But we will not be concerned with the compositional
interpretation of modifier-structures here, so you won’t be needing this rule.
 B de re — de dicto : T T R C 



.. Syntax

What we have at this point does not allow us to interpret even the simplest
syntactic structures. For instance, we can’t interpret the tree in ().

() [ VP John leave]

The verb’s type is hs, eti, so it’s looking for a sister node which denotes a world.
John, which denotes an individual, is not a suitable argument.
We get out of this problem by positing more abstract syntactic structures (at the
LF level). Specifically, we assume that there is a set of covert “world pronouns”
which are generated as sisters to all lexical predicates in LF-structures. Officially,
the variable would be a pair of an index and the type s. Inofficially, we will use
“w” with a subscripted index, with the understanding that the “w” indicates we
are dealing with a variable of type s. So, the syntax would generate something
like ():

() [ John [ leave w ]]

The sentence would then obviously have an assignment-dependent extension (a

truth-value), depending on what world the variable assignment assigns to the
world variable with index . In our intensional system of Chapter  — , we
were assuming the following principle:

() An utterance of a sentence (=LF) φ in world w is true iff JφKw = .

To achieve the same in our new system, we would have to ensure that the variable
assignment assign the utterance world to the free world variable(s) in the sentence.
Another possibility, which we will adopt here, is to introduce a variable binder
on top of the sentence. We will assume the following kind of syntactic structure
at LF:

() [λw [ John [ leave w ]]]

The sentence now has as its extension what used to be its intension, a proposition.
The principle of utterance truth now is this:

() An utterance of a sentence (=LF) φ in world w is true iff JφK(w) = .

Now, we have to look at more complex sentences. First, a simple case of

embedding. The sentence is John wants to leave, which now as an LF like this:
§. T S S: O W V



() [ λw [ John [[ wants w [[ λw [  [ leave w ]]]]]]]]

E .: Calculate the semantic value of (). 

Next, look at an example involving a complex subject, such as the teacher left.

() [ λw [[ the [ teacher w ]][ left w ]]]

The verb will need a world argument as before. The noun teacher will likewise
need one, so that the can get the required argument of type he, ti (not hs, eti!).
If we co-index the two world variables, we derive as the semantic value for ()
what its intension would have been in old system. But nothing we have said
forces us to co-index the two world variables, which is what will allow us to
derive the third reading for relevant examples.
Consider what happens when the sentence contains both a modal operator and
a complex DP in its complement.

() Mary wants a friend of mine to win.

There are now three predicates that need world arguments. Furthermore, there
will be two λ-operators binding world variables. We can now represent the
three readings (to make the structures more readable, we’ll leave off most of the
bracketing and start writing the world arguments as subscripts to the predicates):

() a. non-specific de dicto:

λw Mary wantsw [λw a friend-of-minew leavew ]
b. specific de re:
λw [a friend-of-minew ]λ x Mary wantsw [λw x leavew ]]
c. non-specific de re:
λw Mary wantsw [λw a friend-of-minew leavew ]

In this new framework, then, we have a way of resolving the apparent “scope
paradoxes” and of acknowledging Fodor’s point that there are two separate
distinctions to be made when DPs interact with modal operators. First, there is
the scopal relation between the DP and the operator; the DP may take wider
scope (Fodor’s “specific” reading) or narrower scope (“non-specific” reading) than
the operator. Second, there is the choice of binder for the world-argument of
the DP’s restricting predicate; this may be cobound with the world-argument of
the embedded predicate (Fodor’s “de dicto”) or with the modal operator’s own
world-argument (“de re”). So the de re-de dicto distinction in the sense of Fodor
is not per se a distinction of scope; but it has a principled connection with scope
in one direction: Unless the DP is within the modal operator’s scope, the de
 B de re — de dicto : T T R C 



dicto option (= co-binding the world-pronoun with the embedded predicate’s

world-argument) is in principle unavailable. (Hence “specific” implies “de re”,
and “de dicto” implies “non-specific”.) But there is no implication in the other
direction: if the DP has narrow scope w.r.t. to the modal operator, either the
local or the long-distance binding option for its world-pronoun is in principle
available. Hence “non-specific” readings may be either “de re” or “de dicto”.
For the sake of clarity, we should introduce a different terminology than Fodor’s.
The labels “specific” and “non-specific” especially have been used in so many
different senses by so many different people that it is best to avoid them altogether.
So we will refer to Fodor’s “specific readings” and “non-specific readings” as
“wide-quantification readings” and “narrow-quantification readings”, or “narrow-
Q/wide-Q readings” for short. For the distinction pertaining to the interpretation
of the restricting NP, we will keep the terms “de re” and “de dicto”, but will
amplify them to “restrictor-de re” and “restrictor-de dicto” (“R-de re”/”R-de
dicto”).

E .: For DPs with extensions of type e (specifically, DPs headed by
the definite article), there is a truth-conditionally manifest R-de re/R-de dicto
distinction, but no truth-conditionally detectable wide-Q/narrow-Q distinction.
In other words, if we construct LFs analogous to (a-c) above for an example
with a definite DP, we can always prove that the first option (wide scope DP)
and the third option (narrow scope DP with distantly bound world-pronoun)
denote identical propositions. In this exercise, you are asked to show this for the
example in ().

() John believes that your abstract will be accepted. 

.. The Need for a Binding Theory for World Variables

One could in principle imagine some indexings of our LFs that we have not
considered so far. The following LF indexes the predicate of the complement
clause to the matrix λ-operator rather than to the one on top of its own clause.

() λw John wantsw [λw  leavew ]

Of course, the resulting semantics would be pathological: what John would be

claimed to stand in the wanting relation to is a set of worlds that is either the
entire set W of possible worlds (if the evaluation world is one in which John
leaves) or the empty set (if the evaluation world is one in which John doesn’t
leave). Clearly, the sentence has no such meaning. Do we need to restrict our
system to not generate such an LF? Perhaps not, if the meaning is so absurd that
§. T S S: O W V



the LF would be filtered out by some overarching rules distinguishing sense from
nonsense.
But the problem becomes real when we look at more complex examples. Here is
one discussed by Percus in important work (Percus ):

() Mary thinks that my brother is Canadian.

Since the subject of the lower clause is a type e expression, we expect at least two
readings: de dicto and de re, cf. Exercise .. The two LFs are as follows:

() a. de dicto
λw Mary thinksw [ (that) λw my brotherw (is) Canadianw ]
b. de re
λw Mary thinksw [ (that) λw my brotherw (is) Canadianw ]

But as Percus points out, there is another indexing that might be generated:

() λw Mary thinksw [ (that) w my brotherw (is) Canadianw ]

In (), we have co-indexed the main predicate of the lower clause with the
matrix λ-operator and co-indexed the nominal predicate brother with the em-
bedded λ-operator. That is, in comparison with the de re reading in (b), we
have just switched around the indices on the two predicates in the lower clause.
Note that this LF will not lead to a pathological reading. So, is the predicted
reading one that the sentence actually has? No. For the de re reading, we can
easily convince ourselves that the sentence does have that reading. Here is Percus’
scenario: “My brother’s name is Allon. Suppose Mary thinks Allon is not my
brother but she also thinks that Allon is Canadian.” In such a scenario, our
sentence can be judged as true, as predicted if it can have the LF in (b). But
when we try to find evidence that () is a possible LF for our sentence, we fail.
Here is Percus:

If the sentence permitted a structure with this indexing, we would

take the sentence to be true whenever there is some actual Canadian
who Mary thinks is my brother — even when this person is not my
brother in actuality, and even when Mary mistakenly thinks that he
is not Canadian. For instance, we would take the sentence to be
true when Mary thinks that Pierre (the Canadian) is my brother and
naturally concludes — since she knows that I am American — that
Pierre too is American. But in fact we judge the sentence to be false
on this scenario, and so there must be something that makes the
 B de re — de dicto : T T R C 



indexing in () impossible.

Percus then proposes the following descriptive generalization:

() G X: The situation pronoun that a verb selects for must
be coindexed with the nearest λ above it.

We expect that there will need to be a lot of work done to understand the deeper
sources of this generalization. For fun, we offer the following implementation
(devised by Irene Heim).

.. Two Kinds of World Pronouns

We distinguish two syntactic types of world-pronouns. One type, w-, behaves
like relative pronouns and  in the analysis of H& K, ch. . (pp. ff.): it
is semantically vacuous itself, but can move and leave a trace that is a variable.
The only difference between w- and  is that the latter leaves a variable
of type e when it moves, whereas the former leaves a variable of type s. The
other type of world-pronoun, w-pro, is analogous to bound-variable personal
pronouns, i.e., it is itself a variable (here of type s). Like a personal pronoun, it
can be coindexed with the trace of an existing movement chain.
With this inventory of world-pronouns, we can capture the essence of Gener-
alization X by stipulating that w-pro is only generated in the immediate scope
of a determiner (i.e., as sister to the determiner’s argument). Everywhere else
where a world-pronoun is needed for interpretability, we must generate a w-
and move it. This (with some tacit assumptions left to the reader to puzzle over)
derives the result that the predicates inside nominals can be freely indexed but
that the ones inside predicates are captured by the closest λ-operator.
As we said, there is plenty more to be explored in the Binding Theory for world
pronouns. The reader is referred to the paper by Percus and the references he
cites.

.. Excursus: Semantic reconstruction for de dicto raised

subjects?
Let us look back at the account of de dicto readings of raised subjects that we
sketched earlier in Section ... We showed that you can derive such readings

 Percus works with situation pronouns rather than world pronouns, an immaterial difference
for our purposes here.
§. T S S: O W V



by positing a high type trace for the subject raising, a trace of type hs, het, tii.
Before the lower predicate can combine with the trace, the semantic value of the
trace has to be extensionalized by being applied to the lower evaluation world
(done via the EFA composition principle). Upstairs the raised subject has to
be combined with the λ-abstract (which will be of type hhs, het, tii, ti) via its
intension.
We then saw recently discovered data suggesting that syntactic reconstruction is
actually what is going on. This, of course, raises the question of why semantic
reconstruction is unavailable (otherwise we wouldn’t expect the data that we
observed).
Fox (: p. , fn. ) mentions two possible explanations:
(i) “traces, like pronouns, are always interpreted as variables that range over
individuals (type e)”,
(ii) “the semantic type of a trace is determined to be the lowest type compatible
with the syntactic environment (as suggested in Beck ())”.

In this excursus, we will briefly consider whether our new framework has some-
thing to say about this issue. Let’s figure out what we would have to do in the new
framework to replicate the account in the section on semantics reconstruction.
Downstairs, we would have a trace of type hs, het, tii. To calculate its extension,
we do not need recourse to a special composition principle, but can simply give it
a world-argument (co-indexed with the abstractor resulting from the movement
of the w- in the argument position of the lower verb).
Now, what has to happen upstairs? Well, there we need the subject to be of type
hs, het, tii, the same type as the trace, to make sure that its semantics will enter
the truth-conditions downstairs. But how can we do this?
We need the DP somebody from New York to have as its semantic value an inten-
sion, the function from any world to the existential quantifier over individuals
who are people from New York in that world. This is actually hard to do in our
system. It would be possible if (i) the predicate(s) inside the DP received w-
as their argument, and if (ii) that w- were allowed to moved to adjoin to the
DP. If we manage to rule out at least one of the two preconditions on principled
grounds, we would have derived the impossibility of semantic reconstruction as
a way of getting de dicto readings of raised subjects.
(i) may be ruled out by the Binding Theory for world pronominals, when it
gets developed.
(ii) may be ruled out by principled considerations as well. Perhaps, world-
abstractors are only allowed at sentential boundaries. See Larson () for
some discussion of recalcitrant cases, one of which is the object position of
so-called intensional transitive verbs, the topic of another section.
 B de re — de dicto : T T R C 



. Alternatives to Overt World Variables

We presented (a variant of ) what is currently the most widely accepted solution to
the scope paradoxes, which required the use of non-locally bound world-variables.
There are some alternatives, one of which is to some extent a “notational variant”,
the others involved syntactic scoping after all.

.. Indexed Operators

It is possible to devise systems where predicates maintain the semantics we origi-
nally gave them, according to which they are sensitive to a world of evaluation
parameter. The freedom needed to account for the third reading and further
facts would be created by assuming more sophisticated operators that shift the
evaluation world. Here is a toy example:

() Mary wants [ a [  friend-of-mine ] leave ]

The idea is that the  “temporarily” shifts the evaluation world back to
what it was “before” the abstraction over worlds triggered by want happened.
This kind of system can be spelled out in as much detail as the world-variable
analysis. Cresswell () proves that the two systems are equivalent in their
expressive power. The decision is therefore a syntactic one. Does natural
language have a multitude of indexed world-shifters or a multitude of indexed
world-variables? Cresswell suspects the former, as did Kamp () who wrote:

I of course exclude the possibility of symbolizing the sentence by

means of explicit quantification over moments. Such a symbolization
would certainly be possible; and it would even make the operators P
and F superfluous. Such symbolizations, however, are a considerable
departure from the actual form of the original sentences which they
represent — which is unsatisfactory if we want to gain insight into
the semantics of English. Moreover, one can object to symbolizations
involving quantification over such abstract objects as moments, if
these objects are not explicitly mentioned in the sentences that are
to be symbolized.

There is some resistance to world-time variables because they are not phonetically
realized. But in an operator-based system, we’ll have non-overt operators all over
the place. So, there is no a priori advantage for either system. We will stick with
the more transparent LFs with world variables.
§. A  O W V



.. Scoping After All?

Suppose we didn’t give up our previous framework, in which the evaluation-
world for any predicate was strictly determined by its LF-position. It turns out
that there is a way (actually, two ways) to derive Fodor’s non-specific de re reading
in that framework after all.
Recall again what we need. We need a way to evaluate the restrictive predicate
of a DP with respect to the higher evaluation world while at the same time
interpreting the quantificational force of the DP downstairs in its local clause.
We saw that if we move the DP upstairs, we get the restriction evaluated upstairs
but we also have removed the quantifier from where it should exert its force.
And if we leave the DP downstairs where its quantificational forces is felt, its
restriction is automatically evaluated down there as well. That is why Fodor’s
reading is paradoxical for the old framework. In fact, though there is no paradox.

Way  Raise the DP upstairs but leave a hhe, ti, ti trace. This way the restriction
is evaluated upstairs, then a quantifier extension is calculated, and that
quantifier extension is transmitted to trace position. This is just what we
needed.
Way  Move the NP-complement of a quantificational D independently of the
containing DP. Then we could generate three distinct LFs for a sentence
like Mary wants a friend of mine to win: two familiar ones, in which the
whole DP a friend of mine is respectively inside and outside the scope
of want, plus a third one, in which the NP friend of mine is outside the
scope of want but the remnant DP a [NP t] has been left behind inside
it:
() [ [ NP f-o-m] λ [ Mary [ want [ [ DP a the,ti, ] win]]]]

E .: Convince yourself that this third LF represents the narrow-
quantification, restrictor-de re reading (Fodor’s “non-specific de re"). 

We have found, then, that it is in principle possible after all to account for narrow-
Q R-de re readings within our original framework of intensional semantics.

E .: In (), we chose to annotate the trace of the movement of the
NP with the type-label he, ti, thus treating it as a variable whose values are
predicate-extensions (characteristic functions of sets of individuals). As we just
saw, this choice led to an interpretable structure. But was it our only possible
choice? Suppose the LF-structure were exactly as in (), except that the trace had

 Something like this was proposed by Groenendijk & Stokhof () in their treatment of
questions with which-DPs.
 B de re — de dicto : T T R C 



been assigned type hs, eti instead of he, ti. Would the tree still be interpretable?
If yes, what reading of the sentence would it express? 

E .: We noted in the previous section about the world-pronouns frame-
work that there was a principled reason why restrictor-de dicto readings necessarily
are narrow-quantification readings. (Or, in Fodor’s terms, why there is no such
thing as a “specific de dicto” reading.) In that framework, this was simply a con-
sequence of the fact that bound variables must be in the scope of their binders.
What about the alternative account that we have sketched in the present section?
Does this account also imply that R-de dicto readings are necessarily narrow-Q?


. Scope, Restrictors, and the Syntax of

Movement
To conclude our discussion of the ambiguities of DPs in the complements of
modal operators, let us consider some implications for the study of LF-syntax.
This will be very inconclusive.
Accepting the empirical evidence for the existence of narrow-Q R-de re readings
which are truth-conditionally distinct from both the wide-Q R-de re and the
narrow-Q R-de dicto readings, we are facing a choice between two types of
theories. One theory, which we have referred to as the “standard” one, uses a
combination of DP-movement and world-pronoun binding; it maintains that
wide-quantification readings really do depend on (covert) syntactic movement,
but de re interpretations of the restrictor do not. The other theory, which we may
dub the “scopal” account, removes the restrictor from the scope of the modal
operator, either by QR (combined with an het, ti type trace) or by movement of
the NP-restrictor by itself.
In order to adjudicate between these two competing theories, we may want to
inquire whether the R-de re — de dicto distinction exhibits any of the properties
that current syntactic theory would take to be diagnostic of movement. This is a
very complex enterprise, and the few results to have emerged so far appear to be
pointing in different directions.
We have already mentioned that it is questionable whether NPs that are com-
plements to D can be moved out of their DPs. Even if it is possible, we might
expect this movement to be similar to the movement of other predicates, such
as APs, VPs, and predicative NPs. Such movements exist, but — as discussed
by Heycock, Fox, and the sources they cite — they typically have no effect on
semantic interpretation and appear to be obligatorily reconstructed at LF. The
§. S, R,   S  M



type of NP-movement required by the purely scopal theory of R-de re readings

would be exceptional in this respect.
Considerations based on the locality of uncontroversial instances of QR provide
another reason to doubt the plausibility of the scopal theory. May (b) argued,
on the basis of examples like (), that quantifiers do not take scope out of
embedded tensed clauses.

() a. Some politician will address every rally in John’s district.

b. Some politician thinks that he will address every rally in John’s
district.

While in (a) the universal quantifier can take scope over the existential quan-
tifier in subject position, this seems impossible in (b), where the universal
quantifier would have to scope out of its finite clause. Therefore, May suggested,
we should not attribute the de re reading in an example like our () to the
operation of QR.

() John believes that your abstract will be accepted.

As we saw above, the standard theory which appeals to non-locally bound world-
pronouns does have a way of capturing the de re reading of () without any
movement, so it is consistent with May’s suggestion. The purely scopal theory
would have to say something more complicated in order to reconcile the facts
about () and (). Namely, it might have to posit that DP-movement is
finite-clause bound, but NP-movement is not. Or, in the other version, it would
have to say that QR can escape finite clauses but only if it leaves a het, ti type
trace.
Both theories, by the way, have a problem with the fact that May’s finite-clause-
boundedness does not appear to hold for all quantificational DPs alike. If we
look at the behavior of every, no, and most, we indeed can maintain that there is
no DP-movement out of tensed complements. For example, () could mean
that Mary hopes that there won’t be any friends of mine that win. Or it could
mean (with suitable help from the context) that she hopes that there is nobody
who will win among those shaggy people over there (whom I describe as my
friends). But it cannot mean merely that there isn’t any friend of mine who she
hopes will win.

() Mary hopes that no friend of mine will win.

So () has R-de dicto and R-de re readings for no friend of mine, but no
wide-quantification reading where the negative existential determiner no takes
 B de re — de dicto : T T R C 



matrix scope. Compare this with the minimally different infinitival complement
structure, which does permit all three kinds of readings.

() Mary expects no friend of mine to win.

However, indefinite DPs like a friend of mine, two friends of mine are notoriously
much freeer in the scope options for the existential quantifiaction they express.
For instance, even the finite clause in () seems to be no impediment to a
reading that is not only R-de re but also wide-quantificational (i.e., it has the
existential quantifier over individuals outscoping the universal world-quantifier).

() Mary hopes that a friend of mine will win.

The peculiar scope-taking behavior of indefinites (as opposed to universal, pro-

portional, and negative quantifiers) has recently been addressed by a number of
authors (Abusch ; Kratzer ; Matthewson ; Reinhart ; Winter
), and there are good prospects for a successful theory that generates even
the wide-Q R-de re readings of indefinites without any recourse to non-local
DP-movement. You are encouraged to read these works, but for our current
purposes here, all we want to point out is that, with respect to the behavior of
indefinites, neither of the two theories we are trying to compare seems to have a
special advantage over the other. This is because wide-Q readings result from
DP-movement according to both theories.
As we mentioned in the previous chapter, a number of recent papers have been
probing the connection between de dicto readings and the effects of Binding
Condition C applying at LF. These authors have converged on the conclusion
that DPs which are read as de dicto behave w.r.t. Binding Theory as if they are
located below the relevant modal predicate at LF, and DPs that are read as de
re (i.e., wide-Q, R-de re) behave as if they are located above. It is natural to
inquire whether the same kind of evidence could also be exploited to determine
the LF-location of the NP-part of a DP which is read as narrow-quantificational
but restrictor-de re. If this acted for Condition C purposes as if it were below
the attitude verb, it would confirm the standard theory (non-locally bound
world-pronouns), whereas if it acted as if it was scoped out, we’d have evidence
for the scopal account. Sharvit () constructs some of the relevant examples
and reports judgments that actually favor the scopal theory. For example, she
observes that (a) does allow the narrow-Q, R-de re-reading indicated in (b).

() a. How many students who like John does he think every professor
talked to?

 Sharvit’s own conclusion, however, is not that her data supports the purely scopal theory.
§. S, R,   S  M



b. For which n does John think that every professor talked to n people
in the set of students who actually like John?

More research is required to corroborate this finding.

As a final piece of potentially relevant data, consider a contrast in Marathi
recently discussed by Bhatt ().

() [ji bai kican madhe ahe]i Ram-la watte ki [[ti [ti
 woman kitchen in is Ram thinks that that woman
bai]i ] kican madhe nahi]
kitchen in not is
‘Ram thinks that the woman who is in the kitchen is not in the
kitchen’
() Ram-la watte ki [ [ji bai kican madhe ahe]i [[ti [ti bai]i ] kican madhe
nahi] ]
Ram thinks that  woman kitchen in is that woman kitchen in not is
‘Ram thinks that the woman who is in the kitchen is not in the kitchen’

The English translation of both examples has two readings: a (plausible) de re

reading, on which Ram thinks of the woman who is actually in the kitchen
that she isn’t, and an (implausible) de dicto reading, on which Ram has the
contradictory belief that he would express by saying: “the woman in the kitchen
is not in the kitchen”. The Marathi sentence () also allows these two readings,
but () unambiguously expresses the implausible de dicto reading. Bhatt’s
explanation invokes the assumption that covert movement in Hindi cannot
cross a finite clause boundary. In (), where the correlative clause has moved
overtly, it can stay high or else reconstruct at LF, thus yielding either reading.
But in (), where it has failed to move up overtly, it must also stay low at LF,
and therefore can only be de dicto. What is interesting about this account is
that it crucially relies on a scopal account of the R-de re-R-de dicto distinction.
(Recall that with type-e DPs like definite descriptions, there is no additional
wide/narrow-Q ambiguity.) If the standard theory with its non-locally bindable
world-pronouns were correct, we would not expect the constraint that blocks
covert movement in () to affect the possibility of a de re reading.
In sum, then, the evidence appears to be mixed. Some observations appear
to favor the currently standard account, whereas others look like they might
confirm the purely scopal account after all. Much more work is needed.
 B de re — de dicto : T T R C 



. A Recurring Theme: Historical Overview

To recap, the main shape of the phenomenon discussed in this chapter is that the
intensional parameter (time, world) with respect to which the predicate restrict-
ing a quantifier is interpreted can be distinct from the one that is introduced
by the intensional operator that immediately scopes over the quantifier. The
crucial cases have the character of a “scope paradox”. This discovery is one that
has been made repeatedly in the history of semantics. It has been made both in
the domain of temporal dependencies and in the domain of modality. Here are
some of the highlights of that history. .

. The now-operator
Prior () noticed a semantic problem with the adverb now. The main
early researchers that addressed the problem were Kamp () and Vlach
(). A good survey was prepared by van Benthem (). Another early
reference is Saarinen (). The simplest scope paradox examples looked
like this:
() One day all persons now alive will be dead.
While for this example one could say that now is special in always having
access to the utterance time, other examples show that an unbounded
number of times need to be tracked. It became clear in this work that
whether one uses a multitude of indexed now and then-operators or allows
variables over times is a syntactic and not a deep semantic question.
. The actually-operator
The modal equivalent of the Prior-Kamp scope paradox sentence is:
() It might have been that everyone actually rich was poor.
Crossley & Humberstone () discuss such examples. Double-indexed
systems of modal logic were studied by Segerberg () and Åqvist ().
See also work by Lewis (a), van Inwagen (), and Hazen ().
Indexed actually-operators are discussed by Prior & Fine (), Peacocke
(), and Forbes (, , ).
. The time of nominal predicates
There is quite a bit of work that argues that freedom in the time-dependency
of nominals even occurs when there is no apparent space for temporal
operators. Early work includes Enç (, ). But see also Ejerhed
(). More recently Musan’s dissertation (Musan ) is relevant.
 Some of this history can be found in comments throughout Cresswell’s book (Cresswell ),
which also contains additional references
§. A R T: H O



() Every fugitive is back in custody.

. Tense in Nominals
There is some syntactic work on tense in nominals, see for example
Wiltschko ().
. The Fodor-Reading
Examples similar to the ones from Fodor and Bäuerle that we used at
the beginning of this chapter are discussed in many places (Abusch ;
Bonomi ; Farkas ; Hellan ; Ioup ). The point that all
these authors have made is that the NP-predicate restricting a quantifier
may be evaluated in the actual world, even when that quantifier clearly
takes scope below a modal predicate.
Heim (?) gives an example like this:
() Every time it could have been the case that the player on the left
was on the right instead.
Here, the player on the left must be evaluated with respect to the actual
world. But it is inside a tensed clause, which — as we saw earlier — is
usually considered a scope island for quantifiers.
. Explicit World Variables
Systems with explicit world/time variables were introduced by Tichy ()
and Gallin (). A system (Ty) with overt world-variables is used by
Groenendijk & Stokhof in their dissertation on the semantics of questions.
See also Zimmermann () on the expressive power of that system.
. Movement
The idea of getting the third reading via some kind of syntactic scoping
has not been pursued much. But there is an intriguing idea in a paper
by Bricker (), cited by Cresswell (: p. ). Bricker formalizes a
sentence like Everyone actually rich might have been poor as follows:
() ∃X(∀y(Xy ≡ rich y)&  ∀y(Xy → poor y))

This is apparently meant to be interpreted as ‘there is a plurality X all of

whose members are rich and it might have been the case that all of the
members of X are poor’. This certainly looks like somehow a syntactic
scoping of the restrictive material inside the universal quantifier out of the
scope of the modal operator has occurred.
— T     —
 B

Abusch, Dorit. . The scope of indefinites. Natural Language Semantics ().
–. doi:./BF.

Aikhenvald, Alexandra Y. . Evidentiality. Oxford: Oxford University Press.

Aloni, Maria. . Free choice, modals, and imperatives. Natural Language
Semantics (). –. doi:./s---.

Alonso-Ovalle, Luis. . Disjunction in alternative semantics: University of

Massachusetts at Amherst dissertation. URL http://alonso-ovalle.net/index.
php?page_id=.

Anand, Pranav & Valentine Hacquard. . Epistemics with attitude. Proceed-
ings of Semantics and Linguistic Theory . doi:/.

Åqvist, Lennart. . Modal logic with subjunctive conditionals and

dispositional predicates. Journal of Philosophical Logic (). –.
doi:./BF.

Asher, Nicholas. . A typology for attitude verbs and their anaphoric proper-
ties. Linguistics and Philosophy (). –. doi:./BF.

Barcan, Ruth C. . A functional calculus of first order based on strict

implication. Journal of Symbolic Logic (). –. doi:./.

Barwise, John & Robin Cooper. . Generalized quantifiers and natural
language. Linguistics and Philosophy (). –. doi:./BF.

Bäuerle, Rainer. . Pragmatisch-semantische aspekte der NP-interpretation.

In M. Faust, R. Harweg, W. Lehfeldt & G. Wienold (eds.), Allgemeine sprach-
wissenschaft, sprachtypologie und textlinguistik: Festschrift für peter hartmann,
–. Narr Tübingen.

Beck, Sigrid. . Wh–constructions and transparent logical form: Universität

Tübingen dissertation.

Belnap, Jr., Nuel D. . Conditional assertion and restricted quantification.

Noûs (). –. doi:./.
 B

Belnap, Jr., Nuel D. . Restricted quantification and conditional assertion.

In Hugues Leblanc (ed.), Truth, syntax and modality: Proceedings of the Temple
University conference on alternative semantics, vol.  Studies in Logic and the
Foundations of Mathematics, –. Amsterdam: North-Holland.
Bennett, Jonathan. . A philosophical guide to conditionals. Oxford University
Press.
Bennett, Michael & Barbara Partee. . Toward the logic of tense and aspect in
English. Indiana University Linguistics Club.
van Benthem, Johan. . Tense logic and standard logic. Logique et Analyse .
–.
Bhatt, Rajesh. . Obligation and possession. In Heidi Harley (ed.), Papers
from the upenn/mit roundtable on argument structure and aspect, vol.  MIT
Working Papers in Linguistics, –. URL http://people.umass.edu/bhatt/
papers/bhatt-haveto.pdf.
Bhatt, Rajesh. . Locality in apparently non-local relativization: Correlatives
in the modern indo-aryan languages. Handout for Talk Presented at UT
Austin and MIT.
Bhatt, Rajesh & Roumyana Pancheva. . Conditionals. In The Blackwell
companion to syntax, vol. , –. Blackwell. URL http://www-rcf.usc.edu/
~pancheva/bhatt-pancheva_syncom.pdf.
Blain, Eleanor M. & Rose-Marie Déchaine. . Evidential types: Evidence
from Cree dialects. International Journal of American Linguistics (). –.
doi:./.
Blumson, Ben. . Pictures, perspective and possibility. Philosophical Studies
doi:./s---.
Bonomi, Andrea. . Transparency and specificity in intensional contexts.
In P. Leonardi & M. Santambrogio (eds.), On quine, –. Cambridge
University Press.
Bonomi, Andrea & Sandro Zucchi. . A pragmatic framework for truth
in fiction. Dialectica (). –. doi:./j.-..tb.x.
Preprint http://filosofia.dipafilo.unimi.it/~bonomi/Pragmatic.pdf.
Bricker, Phillip. . Quantified modal logic and the plural de re. In French,
Uehling & Wettstein (eds.), Contemporary perspectives in the philosophy of
language ii Midwest Studies in Philosophy, Vol. , –. University of
Notre Dame Press.
 

Butler, Jonny. . A minimalist treatment of modality. Lingua (). –.

doi:./S-()-.

Carnap, Rudolf. . Meaning and necessity: A study in semantics and modal
logic. Chicago: University of Chicago Press.

Chierchia, Gennaro & Sally McConnell-Ginet. . Meaning and grammar:

An introduction to semantics (nd edition). MIT Press.

Chomsky, Noam. . A minimalist program for linguistic theory. In Kenneth

Hale & Samuel Jay Keyser (eds.), The view from building : Essays in linguistics
in honor of sylvain bromberger, –. Cambridge, MA: MIT Press.

Cooper, Robin. . Tense and discourse location in situation semantics.

Linguistics and Philosophy (). –. doi:./BF.

Copeland, B. Jack. . The genesis of possible worlds semantics. Journal of

Philosophical Logic (). –. doi:./A:.

Cormack, Annabel & Neil Smith. . Modals and negation in English. In
Sjef Barbiers, Frits Beukema & Wim van der Wurff (eds.), Modality and its
interaction with the verbal system, –. Benjamins.

Cresswell, Max. . Logics and languages. London: Methuen.

Cresswell, Max. . Entities and indices. Kluwer Dordrecht.

Crossley, J. N. & I. L. Humberstone. . The logic of actually. Reports on

Mathematical Logic . –.

DeRose, Keith. . Epistemic possibilities. The Philosophical Review ().

–. doi:./.

Dowty, David, Robert Wall & Stanley Peters. . Introduction to Montague
semantics. Kluwer.

Dowty, David R. . Toward a semantic analysis of verb aspect and the
english ‘imperfective’ progressive. Linguistics and Philosophy (). –.
doi:./BF.

Dowty, David R. . Tenses, time adverbs, and compositional semantic theory.
Linguistics and Philosophy (). –. doi:./BF.

Drubig, Hans Bernhard. . On the syntactic form of epistemic modality. Ms,
Universität Tübingen. URL http://www.sfb.uni-tuebingen.de/b/papers/
DrubigModality.pdf.
 B

Edgington, Dorothy. . On conditionals. Mind (). –. URL

./mind/...
Egan, Andy. . Epistemic modals, relativism, and assertion. Philosophical
Studies (). –. doi:./s---x.
Egan, Andy, John Hawthorne & Brian Weatherson. . Epistemic modals in
context. In Gerhard Preyer & Georg Peter (eds.), Contextualism in philosophy:
Knowledge, meaning, and truth, –. Oxford: Oxford University Press.
Ejerhed, Eva. . The syntax and semantics of English tense markers. Mono-
graphs from the Institute of Linguistics, University of Stockholm, No. .
Ejerhed, Eva. . Tense as a source of intensional ambiguity. In Frank Heny
(ed.), Ambiguities in intensional contexts, –. Reidel Dordrecht.
Enç, Mürvet. . Tense without scope: An analysis of nouns as indexicals: Univer-
sity of Wisconsin, Madison dissertation.
Enç, Mürvet. . Towards a referential analysis of temporal expressions.
Linguistics and Philosophy . –.
Farkas, Donka. . Evaluation indices and scope. In Anna Szabolcsi (ed.),
Ways of scope taking, –. Dordrecht: Kluwer.
von Fintel, Kai. . Restrictions on quantifier domains: University of Mas-
sachusetts at Amherst dissertation. URL http://semanticsarchive.net/Archive/
jANIwN/fintel--thesis.pdf.
von Fintel, Kai. . Quantifiers and ‘if ’-clauses. The Philosophical Quarterly
(). –. doi:./-.. URL http://mit.edu/fintel/
www/qandif.pdf.
von Fintel, Kai. . NPI licensing, Strawson entailment, and context depen-
dency. Journal of Semantics (). –. doi:./jos/...
von Fintel, Kai. . Counterfactuals in a dynamic context. In Michael
Kenstowicz (ed.), Ken Hale: A life in language, –. MIT Press. URL
http://mit.edu/fintel/counterfactuals.pdf.
von Fintel, Kai. . Modality and language. In Donald M. Borchert (ed.),
Encyclopedia of philosophy – second edition, MacMillan. URL http://mit.edu/
fintel/fintel--modality.pdf.
von Fintel, Kai. . If : The biggest little word. Slides from a plenary
address given at the Georgetown University Roundtable, March , . URL
http://mit.edu/fintel/gurt-slides.pdf.
 

von Fintel, Kai. . Conditionals. Ms, prepared for Semantics: An international
handbook of meaning, edited by Klaus von Heusinger, Claudia Maienborn, and
Paul Portner. URL http://mit.edu/fintel/fintel--hsk-conditionals.pdf.
von Fintel, Kai & Anthony S. Gillies. . An opinionated guide to epistemic
modality. In Tamar Szabó Gendler & John Hawthorne (eds.), Oxford studies
in epistemology: Volume , –. Oxford University Press. URL http://mit.
edu/fintel/fintel-gillies--ose.pdf.
von Fintel, Kai & Anthony S. Gillies. a. CIA leaks. The Philosophical Review
(). –. doi:./--.
von Fintel, Kai & Anthony S. Gillies. b. Might made right. To appear in a
volume on epistemic modality, edited by Andy Egan and Brian Weatherson,
Oxford University Press. URL http://mit.edu/fintel/fintel-gillies--mmr.
pdf.
von Fintel, Kai & Anthony S. Gillies. . Must . . . stay . . . strong! Natural
Language Semantics (). –. doi:./s---.
von Fintel, Kai & Sabine Iatridou. . If and when If -clauses can restrict
quantifiers. Ms, MIT. URL http://mit.edu/fintel/fintel-iatridou--ifwhen.
pdf.
von Fintel, Kai & Sabine Iatridou. . Epistemic containment. Linguistic
Inquiry (). –. doi:./.
von Fintel, Kai & Sabine Iatridou. . What to do if you want to go to
Harlem: Anankastic conditionals and related matters. Ms, MIT. URL
http://mit.edu/fintel/fintel-iatridou--harlem.pdf.
von Fintel, Kai & Sabine Iatridou. . How to say ought in Foreign: The
composition of weak necessity modals. In Jacqueline Guéron & Jacqueline
Lecarme (eds.), Time and modality (Studies in Natural Language and Linguistic
Theory ), –. Springer. doi:./----.
Fodor, Janet Dean. . The linguistic description of opaque contexts: Mas-
sachusetts Institute of Technology dissertation. Published in  by Indiana
University Linguistics Club and in  in the Series “Outstanding Disserta-
tions in Linguistics” by Garland.
Forbes, Graeme. . Physicalism, instrumentalism, and the semantics of modal
logic. Journal of Philosophical Logic . –.
Forbes, Graeme. . The metaphysics of modality. Oxford: Clarendon Press.
Forbes, Graeme. . Languages of possibility. Oxford: Blackwell.
 B

Fox, Danny. . Economy and semantic interpretation. MIT Press.

Fox, Danny. . Free choice and the theory of scalar implicatures. In
Uli Sauerland & Penka Stateva (eds.), Presupposition and implicature in
compositional semantics, –. New York: Palgrave Macmillan. URL
http://web.mit.edu/linguistics/people/faculty/fox/free_choice.pdf.

Frank, Anette. . Context dependence in modal constructions: Universität

Stuttgart dissertation. URL http://www.cl.uni-heidelberg.de/~frank/papers/
header.pdf.

Gajewski, Jon. . Neg-raising: Polarity and presupposition: Massachusetts

Institute of Technology dissertation. doi:./.

Gajewski, Jon. . Neg-raising and polarity. Linguistics and Philosophy

doi:./s---z.

Gallin, D. . Intensional and higher-order modal logic. North-Holland Ams-

terdam.

Gamut, L. T. F. . Logic, language, and meaning. Chicago University Press.

Garson, James. . Modal logic. In Edward N. Zalta (ed.), The Stanford
encyclopedia of philosophy, URL http://plato.stanford.edu/entries/logic-modal/.

Gergonne, Joseph Diaz. . Essai de dialectique rationnelle. Annales de

Mathématiques Pures et Appliquées . –. URL http://archive.numdam.
org/article/AMPA_-_____.pdf.

Geurts, Bart. . Presuppositions and anaphors in attitude contexts. Linguistics

and Philosophy (). –. doi:./A:.

Giannakidou, Anastasia. . Affective dependencies. Linguistics and Philosophy

(). –. doi:./A:.

Gillies, Anthony S. . Counterfactual scorekeeping. Linguistics and Philosophy

(). –. doi:./s---.

Gillies, Anthony S. . On truth-conditions for if (but not quite only if ).

The Philosophical Review (). –. doi:./--.

Gillies, Anthony S. . Iffiness. Semantics and Pragmatics (). –.

doi:./sp...

Groenendijk, Jeroen & Martin Stokhof. . Semantic analysis of Wh-

complements. Linguistics and Philosophy . –.
 

Hacking, Ian. . Possibility. The Philosophical Review (). –.

doi:./. URL http://www.jstor.org/stable/.
Hacquard, Valentine. . Aspects of modality: Massachusetts Institute of
Technology dissertation. URL http://people.umass.edu/hacquard/hacquard_
thesis.pdf.
Hacquard, Valentine. . Modality. Ms, prepared for Semantics: An in-
ternational handbook of meaning, edited by Klaus von Heusinger, Claudia
Maienbon, and Paul Portner. URL http://ling.umd.edu/~hacquard/papers/
HoS_Modality_Hacquard.pdf.
Hanley, Richard. . As good as it gets: Lewis on truth in fiction. Australasian
Journal of Philosophy (). –. doi:./.
Hawthorne, John. . Eavesdroppers and epistemic modals. Ms, Rutgers
University, to appear in the proceedings of the  Sofia Conference in
Mexico, in a supplement to Noûs.
Hazen, Allen. . One of the truths about actuality. Analysis . –.
Heim, Irene. . Presupposition projection and the semantics of attitude verbs.
Journal of Semantics (). –. doi:./jos/...
Heim, Irene & Angelika Kratzer. . Semantics in generative grammar. Oxford:
Blackwell.
Hellan, Lars. . On semantic scope. In Ambiguities in intensional contexts,
Dordrecht: Reidel.
Herzberger, Hans. . Counterfactuals and consistency. The Journal of
Philosophy (). –. doi:./.
Higginbotham, James. . Conditionals and compositionality. Philosophical
Perspectives (). –. doi:./j.-...x.
Hintikka, Jaako. . Semantics for propositional attitudes. In J.W. Davis, D.J.
Hockney & W.K. Wilson (eds.), Philosophical logic, –. Dordrecht: Reidel.
Hockett, Charles F. . The origin of speech. Scientific American . –.
Hockett, Charles F. & Stuart A. Altmann. . A note on design features.
In Thomas A. Sebeok (ed.), Animal communication: Techniques of study and
results of research, –. Indiana University Press.
Horn, Laurence R. . On the semantic properties of the logical operators in
English: UCLA dissertation.
 B

Hughes, G.E. & M.J. Cresswell. . An introduction to modal logic. London:
Methuen.
Hughes, G.E. & M.J. Cresswell. . A new introduction to modal logic. London:
Routledge.
Huitink, Janneke. . Modals, conditionals and compositionality: Radboud Uni-
versiteit Nijmegen dissertation. URL http://user.uni-frankfurt.de/~huitink/
Huitink-dissertation.pdf.
Huitink, Janneke. a. Domain restriction by conditional connectives.
Ms, Goethe-University Frankfurt. URL http://semanticsarchive.net/Archive/
zgMDMM/Huitink-domainrestriction.pdf.
Huitink, Janneke. b. Quantified conditionals and compositionality. Ms, to
appear in Language and Linguistics Compass. URL http://user.uni-frankfurt.
de/~huitink/compass-conditionals-final.pdf.
Iatridou, Sabine. . On the contribution of conditional Then. Natural
Language Semantics (). –. doi:./BF.
van Inwagen, Peter. . Indexicality and actuality. Philosophical Review .
–.
Ioup, Georgette. . Specificity and the interpretation of quantifiers. Linguistics
and Philosophy (). –.
Kadmon, Nirit & Fred Landman. . Any. Linguistics and Philosophy ().
–. doi:./BF.
Kamp, Hans. . Formal properties of now. Theoria . –.
Kamp, Hans. . Free choice permission. Proceedings of the Aristotelian Society,
New Series . –.
Knuuttila, Simo. . Medieval theories of modality. In Edward N. Zalta
(ed.), The stanford encyclopedia of philosophy, URL http://plato.stanford.edu/
archives/fall/entries/modality-medieval/.
Kratzer, Angelika. . What must and can must and can mean. Linguistics and
Philosophy (). –. doi:./BF.
Kratzer, Angelika. . Semantik der Rede: Kontexttheorie – Modalwörter –
Konditionalsätze. Königstein/Taunus: Scriptor.
Kratzer, Angelika. . The notional category of modality. In Hans-Jürgen
Eikmeyer & Hannes Rieser (eds.), Words, worlds, and contexts: New approaches
in word semantics (Research in Text Theory ), –. Berlin: de Gruyter.
 

Kratzer, Angelika. . Conditionals. Chicago Linguistics Society (). –.

Kratzer, Angelika. . Modality. In Arnim von Stechow & Dieter Wunderlich
(eds.), Semantics: An international handbook of contemporary research, –.
Berlin: de Gruyter.
Kratzer, Angelika. . Scope or pseudoscope? are there wide-scope indefinites?
In Susan Rothstein (ed.), Events and grammar, –. Kluwer Dordrecht.
Kratzer, Angelika. . Decomposing attitude verbs. Handout from a talk
honoring Anita Mittwoch on her th birthday at the Hebrew Univer-
sity of Jerusalem July , . URL http://semanticsarchive.net/Archive/
DcwYJkM/attitude-verbs.pdf.
Kripke, Saul. . Naming and necessity. Oxford: Blackwell.
Landman, Fred. . The progressive. Natural Language Semantics (). –.
doi:./BF.
Larson, Richard. . The grammar of intensionality. In G. Preyer (ed.), On
logical form, Oxford University Press.
Leslie, Sarah-Jane. . If, unless, and quantification. In Robert J. Stainton &
Christopher Viger (eds.), Compositionality, context and semantic values: Essays
in honour of Ernie Lepore, –. Springer. doi:./----_.
Lewis, Clarence Irving & Cooper Harold Langford. . Symbolic logic. New
York: Century.
Lewis, David. a. Anselm and actuality. Noûs . –. Reprinted (with a
postscript) in Lewis (Lewis : pp. –).
Lewis, David. b. General semantics. Synthese (-). –.
doi:./BF.
Lewis, David. . Counterfactuals. Oxford: Blackwell.
Lewis, David. . Adverbs of quantification. In Edward Keenan (ed.), Formal
semantics of natural language, –. Cambridge University Press.
Lewis, David. . Truth in fiction. American Philosophical Quarterly (). –
. URL http://www.jstor.org/stable/. Reprinted with postscripts in
Lewis (), pp. –.
Lewis, David. . Ordering semantics and premise semantics for counterfac-
tuals. Journal of Philosophical Logic (). –. doi:./BF.
Reprinted in David Lewis, Papers in Philosophical Logic, Cambridge: Cam-
bridge University Press, , pp. -.
 B

Lewis, David. . Logic for equivocators. Noûs (). –.

doi:./. Reprinted in Lewis (: pp. –).

Lewis, David. . Philosophical papers: Volume . Oxford: Oxford University

Press.

Lewis, David. . On the plurality of worlds. Oxford: Blackwell.

Lewis, David. . Papers in philosophical logic. Cambridge: Cambridge

University Press.

MacFarlane, John. . Logical constants. In Edward N. Zalta (ed.), The

stanford encyclopedia of philosophy, URL http://plato.stanford.edu/archives/
win/entries/logical-constants/.

MacFarlane, John. . Epistemic modals are assessment-sensitive. Ms,

University of California, Berkeley, forthcoming in an OUP volume on
epistemic modals, edited by Brian Weatherson and Andy Egan. URL
http://sophos.berkeley.edu/macfarlane/epistmod.pdf.

Matthewson, Lisa. . On the interpretation of wide-scope indefinites. Natural

Language Semantics ().

May, Robert. a. The grammar of quantification: Massachusetts Institute of

Technology dissertation.

May, Robert. b. The grammar of quantification: Massachusetts Institute of

Technology dissertation.

McCready, Eric & Norry Ogata. . Evidentiality, modality and probability.
Linguistics and Philosophy (). –. doi:./s---.

Montague, Richard. . The proper treatment of quantification in or-

dinary English. In Jaako Hintikka, Julius Moravcsik & Patrick Sup-
pes (eds.), Approaches to natural language, –. Dordrecht: Reidel.
URL http://www.blackwellpublishing.com/content/BPL_Images/Content_
store/Sample_chapter//Portner.pdf. Reprinted in Portner &
Partee (), pp. –.

Moulton, Keir. . Clausal complementation and the Wager-class. Proceedings

of the North East Linguistics Society . URL http://sites.google.com/site/
keirmoulton/Moultonnelswager.pdf.

Moulton, Keir. . Natural selection and the syntax of clausal complementation:
University of Massachusetts at Amherst dissertation. URL http://scholarworks.
umass.edu/open_access_dissertations//.
 

Musan, Renate. . On the temporal interpretation of noun phrases: Mas-

sachusetts Institute of Technology dissertation. Published in  in the Series
“Outstanding Dissertations in Linguistics” by Garland.
Musan, Renate. . Tense, predicates, and lifetime effects. Natural Language
Semantics (). –. doi:./A:.
Nauze, Fabrice. . Modality in typological perspective: Universiteit van Ams-
terdam dissertation. URL http://www.illc.uva.nl/Publications/Dissertations/
DS--.text.pdf.
Nute, Donald. . Conditional logic. In Dov Gabbay & Franz Guenthner
(eds.), Handbook of philosophical logic. volume ii, –. Dordrecht: Reidel.
Ogihara, Toshiyuki. . The semantics of tense in embedded clauses. Linguistic
Inquiry (). –. URL http://www.jstor.org/stable/.
Ogihara, Toshiyuki. . Tense and aspect in truth-conditional semantics.
Lingua (). –. doi:./j.lingua....
Partee, Barbara H. . Some structural analogies between tenses and pronouns
in English. The Journal of Philosophy (). –. doi:./.
Partee, Barbara H. . Nominal and temporal anaphora. Linguistics and
Philosophy (). –. doi:./BF.
Partee, Barbara H. . Reflections of a formal semanticist as of Feb .
Ms. (longer version of introductory essay in  book). URL http://people.
umass.edu/partee/docs/BHP_Essay_Feb.pdf.
Partee, Barbara H. & Herman L.W. Hendriks. . Montague grammar. In
Johan van Benthem & Alice ter Meulen (eds.), Handbook of logic and language,
–. Elsevier.
Peacocke, Christopher. . Necessity and truth theories. Journal of Philosophical
Logic . –.
Peano, Giuseppe. . Arithmetices principia: Nova methodo exposita. Torino:
Bocca.
Percus, Orin. . Constraints on some other variables in syntax. Natural
Language Semantics (). –.
Perry, John R. . Semantics, possible worlds. In E. Craig (ed.), Routledge
encyclopedia of philosophy, London: Routledge. Preprint http://www-csli.
stanford.edu/~john/PHILPAPERS/posswld.pdf.
Pollock, John. . Subjunctive reasoning. Dordrecht: Reidel.
 B

Portner, Paul. . The semantics of mood, complementation,

and conversational force. Natural Language Semantics (). –.
doi:./A:.

Portner, Paul. . The progressive in modal semantics. Language ().

–. doi:./.

Portner, Paul. . Modality. Oxford University Press.

Portner, Paul & Barbara H. Partee (eds.). . Formal semantics: The essential
readings. Oxford: Blackwell.

Prior, A. N. . Escapism: The logical basis of ethics. In A. I. Melden (ed.),

Essays in moral philosophy, –. Seattle: University of Washington Press.

Prior, A. N. & Kit Fine. . Worlds, times, and selves. London: Duckworth.

Prior, Arthur. . Now. Noûs . –.

Reinhart, Tanya. . Quantifier scope – how labor is divided between QR and
choice functions. Linguistics & Philosophy (). –.

van Rooij, Robert. . Free choice counterfactual donkeys. Journal of Semantics
(). –. doi:./jos/ffl.

Ross, Jeff. . The semantics of media (Studies in Linguistics and Philosophy
(SLAP) ). Dordrecht: Kluwer.

Rothstein, Susan. . Structuring events: A study in the semantics of lexi-

cal aspect Explorations in Semantics. Blackwell. URL http://tinyurl.com/
rothstein-aktionsarten.

Russell, Bertrand. . An inquiry into meaning and truth. London: George
Allen and Unwin.

Saarinen, Esa. . Backwards-looking operators in tense logic and in natural

language. In Jaako Hintikka, I. Niiniluoto & Esa Saarinen (eds.), Essays on
mathematical and philosophical logic, Reidel Dordrecht.

Sauerland, Uli & Paul Elbourne. . Total reconstruction, pf

movement, and derivational order. Linguistic Inquiry (). –.
doi:./.

Schlenker, Philippe. . Conditionals as definite descriptions (A refer-

ential analysis). Research on Language and Computation (). –.
doi:./s---.
 

Schulz, Katrin. . A pragmatic solution for the paradox of free choice
permission. Synthese (). –. doi:./s---y.
Segerberg, K. . Two-dimensional modal logic. Journal of Philosophical Logic
. –.
Sextus Empiricus. c. . Outlines of pyrrhonism.
Sharvit, Yael. . How many questions and attitude verbs. University of
Pennsylvania.
Speas, Peggy. . On the syntax and semantics of evidentials. Language and
Linguistics Compass (). –. doi:./j.-X...x.
Spencer, Mary. . Why the “s” in “intension”? Mind (). –.
doi:./mind/LXXX...
Stalnaker, Robert. . A theory of conditionals. In Nicholas Rescher (ed.),
Studies in logical theory (American Philosophical Quarterly Monograph Se-
ries ), –. Oxford: Blackwell.
Stalnaker, Robert. . Indicative conditionals. Philosophia (). –.
doi:./BF.
Stalnaker, Robert. . Inquiry. MIT Press.
Stalnaker, Robert. . Context and content. Oxford: Oxford University Press.
Stanley, Jason & Zoltán Gendler Szabó. . On quantifier domain restriction.
Mind and Language (/). –. doi:./-..
von Stechow, Arnim. . On the proper treatment of tense. Proceedings of
Semantics and Linguistic Theory . URL http://www.sfs.uni-tuebingen.de/
~arnim/Aufsaetze/SALT.pdf.
von Stechow, Arnim. . Tenses in compositional semantics. To be published
in Wolfgang Klein (ed) The Expression of Time in Language. URL http:
//www.sfs.uni-tuebingen.de/~arnim/Aufsaetze/Approaches.pdf.
Steiner, George. . After Babel: Aspects of language and translation. Oxford
University Press rd edn.
Stephenson, Tamina. a. Judge dependence, epistemic modals, and predicates
of personal taste. Linguistics and Philosophy (). –. doi:./s-
--.
Stephenson, Tamina. b. Towards a theory of subjective meaning: Mas-
sachusetts Institute of Technology dissertation. URL http://semanticsarchive.
net/Archive/QxMjkO/Stephenson--thesis.pdf.
 B

Steup, Matthias. . The analysis of knowledge. In Edward N. Zalta (ed.), The
Stanford encyclopedia of philosophy, Fall  edn. URL http://plato.stanford.
edu/archives/fall/entries/knowledge-analysis/.
Suber, Peter. . Paradoxes of material implication. An electronic hand-out for
the course “Symbolic Logic”. URL http://www.earlham.edu/~peters/courses/
log/mat-imp.htm.
Swanson, Eric. . Modality in language. Philosophy Compass (). –.
doi:./j.-...x.
Swanson, Eric. . Propositional attitudes. Ms, prepared for Semantics: An
international handbook of meaning, edited by Klaus von Heusinger, Claudia
Maienbon, and Paul Portner. URL http://tinyurl.com/swanson-attitudes.
Teller, Paul. . Epistemic possibility. Philosophia (). –.
doi:./BF.
Tichy, Pavel. . An approach to intensional analysis. Noûs . –.
Varzi, Achille. . Inconsistency without contradiction. Notre Dame Journal of
Formal Logic (). –. doi:./ndjfl/.
Vlach, Frank. . Now and then: A formal study in the logic of tense anaphora:
UCLA dissertation.
Warmbrod, Ken. . A defense of the limit assumption. Philosophical Studies
(). –. doi:./BF.
Willett, Thomas. . A cross-linguistic survey of the grammaticalization of
evidentiality. Studies in Language (). –.
Williamson, Timothy. . Knowledge and its limits. Oxford: Oxford University
Press.
Wiltschko, Martina. . On the interpretability of tense on D and its conse-
quences for case theory. Lingua (). –. URL http://dx.doi.org/.
/S-()-X.
Winter, Yoad. . Choice functions and the scopal semantics of indefinites.
Linguistics and Philosophy . –.
Wurmbrand, Susi. . Modal verbs must be raising verbs. West Coast Conference
on Formal Linguistics . –. URL http://citeseerx.ist.psu.edu/viewdoc/
download?doi=....&rep=rep&type=pdf.
Zimmermann, Thomas Ede. . Intensional logic and two-sorted type theory.
Journal of Symbolic Logic . –.
 

Zimmermann, Thomas Ede. . Free choice disjunction and epistemic possi-
bility. Natural Language Semantics (). –. doi:./A:.
Zucchi, Sandro. . Tense in fiction. In Carlo Cecchetto, Gennaro Chierchia
& Maria Teresa Guasti (eds.), Semantic interfaces: Reference, anaphora and
aspect, –. CSLI Publications. URL http://tinyurl.com/ulwxwg.