0% found this document useful (0 votes)
12 views26 pages

Thedefinitionofmodality

This chapter by Renaat Declerck defines modality as the reference to the actualization of situations in nonfactual worlds, distinguishing between root and epistemic modalities. It critiques existing definitions and categorizes various modal expressions, emphasizing the complexity and richness of modality in language. The chapter aims to provide a comprehensive understanding of modality, its types, and their relationships to factual and nonfactual worlds.

Uploaded by

Nam Do Minh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
12 views26 pages

Thedefinitionofmodality

This chapter by Renaat Declerck defines modality as the reference to the actualization of situations in nonfactual worlds, distinguishing between root and epistemic modalities. It critiques existing definitions and categorizes various modal expressions, emphasizing the complexity and richness of modality in language. The chapter aims to provide a comprehensive understanding of modality, its types, and their relationships to factual and nonfactual worlds.

Uploaded by

Nam Do Minh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 26

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/276273181

The definition of modality

Chapter · January 2011


DOI: 10.1075/hcp.29.05dec

CITATIONS READS

32 10,719

1 author:

Renaat Declerck
KU Leuven
93 PUBLICATIONS 2,018 CITATIONS

SEE PROFILE

All content following this page was uploaded by Renaat Declerck on 13 April 2016.

The user has requested enhancement of the downloaded file.


John Benjamins Publishing Company

This is a contribution from Cognitive Approaches to Tense, Aspect, and Epistemic Modality.
Edited by Adeline Patard and Frank Brisard.
© 2011. John Benjamins Publishing Company
This electronic file may not be altered in any way.
The author(s) of this article is/are permitted to use this PDF file to generate printed copies to
be used by way of offprints, for their personal use only.
Permission is granted by the publishers to post this file on a closed server which is accessible
to members (students and staff) only of the author’s/s’ institute, it is not permitted to post
this PDF on the open internet.
For any other use of this material prior written permission should be obtained from the
publishers or through the Copyright Clearance Center (for USA: www.copyright.com).
Please contact [email protected] or consult our website: www.benjamins.com
Tables of Contents, abstracts and guidelines are available at www.benjamins.com
The definition of modality

Renaat Declerck
University of Leuven

This chapter offers a definition of modality that is as concrete and complete as


possible. Modality means that there is reference to actualization of a situation
in a world that is not represented as being the factual world. All types of
modality are pigeonholed, regardless of whether the ‘modalizer’ is an auxiliary,
lexical verb, adverb, conditional clause, a morphological operation like ‘modal
backshifting’ or ‘conditionalization’, etc. Clear definitions are given of root and
epistemic modality, and it is shown that not all epistemic values are modal
values. On the basis of a rich modal world typology and a number of possible
relations between a modal world and the factual world, different subtypes of
modality are distinguished.

Keywords: modality, epistemic, root modality, condition

1. Introduction1

The entry for ‘modality’ in Leech’s (2006: 64) Glossary of English Grammar reads:
“see modal (auxiliary verb)”. In the entry ‘modal (auxiliary) (verb)’ we read: “a
member of a small class of verbs that have meanings relating to modality, that is to
such concepts as possibility or permission (can, may), obligation, necessity or like-
lihood (must, should), prediction, intention or hypothesis (will, would)”. This defi-
nition of modality reduces modality to the use of modal auxiliaries and gives us
little tangible to go by. This is not surprising, because modality is a most elusive
concept, which in pedagogical grammars is usually illustrated (rather than de-
fined) by a list of possible meanings that ‘modal auxiliaries’ can have. Of course,
more detailed linguistic treatments of the phenomenon of modality often do offer

1. I wish to thank Ilse Depraetere, Susan Reed, Bert Cappelle and An Verhulst for the fruitful
discussions we have had in connection with modality. Special thanks are due to Susan Reed and
Caroline Gevaert for their comments on a previous draft of this manuscript.

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

a kind of definition, but there does not seem to be a standard definition that is
widely accepted.
The issue is also clouded by the fact that the class of ‘modal auxiliaries’ is usu-
ally treated as a waste-paper basket in which all the verbs are put which have the
formal and syntactic characteristics of auxiliaries but are clearly not auxiliaries
used to build tense forms, aspectual forms or passive verb forms. At least two of
the auxiliaries that are often (even in major recent grammars) included in the set
of modal auxiliaries, viz. used to and dare, arguably do not belong to that class.2
The meaning expressed by used to is aspectual, not modal: My late uncle used to go
to bed early predicates of my uncle the past characteristic (habit) of going to bed
early.3 I daren’t do that has exactly the same nonmodal meaning as I don’t dare to
do that. Since to dare to is a lexical verb without a modal meaning, the auxiliary
dare in I daren’t do that should be treated as what we could call a ‘lexical auxiliary’
– something which to my knowledge nobody has ever suggested doing. A lexical
auxiliary combines the morpho-syntactic characteristics of an auxiliary with the
lexical meaning of a full verb (rather than with the meaning of an auxiliary ex-
pressing tense, voice, modality or aspect).
In this article, in which we aim to give a satisfactory definition of ‘modality’,
we have to find (among other things) what all the (true) modal auxiliaries have in
common, apart from their specific meanings, which mostly have to do with some
kind of possibility or necessity. It is this common core that constitutes the essence
of modality and thus offers a basis for a definition of modality. As we will see, that
common element of meaning is that a situation is represented as actualizing in a
nonfactual world.

2. Terminological preliminaries

a. We use ‘situation’ as a cover term for the various possible types of contents of
clauses, i.e. as a cover term for anything that can be expressed in a clause. Ac-
cording to Lyons (1977), a situation is either a state, an action, a process
(=change, development) or an event (=a nonagentive dynamic situation, e.g. a
fall). As a cover-term for the various verbs that are typically associated with
one of these kinds of situation we use the term actualize. The sentence The

2. Quirk et al. (1985: 137) treat used to and dare as “marginal modals”. Biber et al. (1999: 484)
call them “marginal auxiliary verbs”, adding that they “can behave like modals in taking auxil-
iary negation and yes-no question inversion”.
3. Huddleston & Pullum (2002) recognize that the meaning of used to is “aspectual, not mod-
al” (p. 115) but they still treat the auxiliary dare as a modal auxiliary (p. 110).

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

situation is actualizing can thus be said of a state that is holding, an action that
is being performed, a change that is taking place or an event that is happening.
In all four cases the sentence refers to the actualization of a situation.
b. A ‘possible world’ – we are using the term in its linguistic meaning rather than
in the way it is used in logic – is always a t-world, i.e. a world which is anchored
to a given time t. This means that a tensed proposition can be true at one time
but false at another, in other words, that it may be true of one t-world but false
of another. Thus, J.F. Kennedy is the President of the U.S. is true of any world
holding at (=anchored to) some time in the course of 1961 but is false of the
objective S-world, i.e. the actual world holding at S (=speech time) – see sec-
tion (d) below. By contrast, omnitemporal situations (referred to by generic or
universal sentences like A horse is an animal) actualize in every objective
world holding at any possible time.
The unmarked form of t-anchoring is S-anchoring (where S means ‘speech time’).
If the world referred to is an S-world, i.e. a world that is anchored to S, and the
sentence referring to it is in the present tense, no anchor time needs to be specified
in the sentence, nor in its context. This explains why, unlike The weather was nice
(which begs the question “When?”), The weather is nice is fully interpretable in
isolation: the hearer assumes that the speaker represents the situation as actualiz-
ing at S. It also explains why omnitemporal situations are as a rule referred to in
the present tense: this follows naturally from the fact that S is one of the times at
which the situation is factual and is the unmarked anchor time in the linguistic
mind of the speaker.
c. There is no essential difference between saying that a proposition is “true of ”
(Lyons 1977: 687) a particular t-world and saying that the situation denoted by
the proposition ‘actualizes in’ that t-world. Thus, if the tensed proposition John
is walking home is true of (=true with reference to) the objective S-world (=the
actual world holding at S), the situation referred to is represented as actualiz-
ing in that world. If the untensed proposition ‘John be walking home’ is not
true of the objective S-world, as in [I wish] John was walking home, the situa-
tion does not actualize in the objective S-world, i.e. it does not belong to the
actualizations making up the objective S-world. In this particular case John is
walking home is true of a counterfactual S-world (i.e. a world which is just the
reverse of the factual world), which means that the situation actualizes in that
counterfactual S-world.
d. We can distinguish between ‘objective’ and ‘subjective’ t-worlds. An objective
t-world is the unique real world that holds at a given time and which is judged
real by an (imaginary) ideal outside observer viewing the world as it is at that
given time. A subjective t-world is an alternative world which is not judged real

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

by such an ideal outside observer but which is conceived of as real by some


consciousness at a certain time. Such a t-world consists of the tensed
(=anchored in time by their finite verb form) propositions that are deemed
true by the world-creating consciousness at the given time. Thus, the situation
referred to by Amsterdam lies in Belgium (which is counterfactual in the objec-
tive S-world) actualizes in the counterfactual S-world existing in the mind of
a speaker who is convinced that this assertion is true at S.
For simplicity, we disregard subjective t-worlds in the discussion below.
e. Another distinction we can make is that between ‘narrow t-worlds’ and ‘ex-
tended t-worlds’. A narrow t-world is a t-world comprising all the situations
that are actualizing at a given time t. Tensed propositions can only be true of
such a world if the tense represents the (actualization of the) situation referred
to as simultaneous with t. An extended t-world is a world comprising all the
situations that are actualizing at t or have actualized before t. Various tenses
can be used to represent the actualization of a situation as factual in an ex-
tended t-world. Thus, Caesar was Emperor of Rome, There have been many
battles between the French and the English and Paris is the capital of France all
refer to actualizations that are factual in the speaker’s extended S-world.
Since an extended ti-world includes a narrow ti-world, any actualization that is
factual in an objective narrow t-world is also factual in the corresponding objective
extended t-world. Thus if John is in London now, the tensed proposition John is in
London is true not only of the narrow S-world but also of the extended S-world.
For the sake of simplicity, we ignore the distinction between narrow and ex-
tended t-worlds whenever it is not crucial. As regards S-worlds, we use ‘factual
world’ in the sense of ‘extended factual S-world’.
f. Whether a sentence is positive or negative is irrelevant to the question whether
the actualization of the situation is factual or not in a given t-world, or to the
question whether the sentence (=tensed proposition) is true or not of that
t-world. Thus both London lies on the Thames and London does not lie on the
Rhine represent the (positive or negative) situation referred to as actualizing in
the factual world, and both sentences are true of the factual world. It follows
that we must distinguish between ‘negative sentences’ and ‘counterfactual sen-
tences’, as well as between ‘negative situations’ and ‘counterfactual situations’.
London does not lie on the Rhine is a negative sentence, referring to the actual-
ization of a negative situation (i.e. a situation referred to in a negative sentence).
That negative situation is represented as actualizing in the factual world. By
contrast, If I were you... refers to a situation that actualizes in a counterfactual

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

S-world. This clause refers to a counterfactual situation (=a situation that does
not actualize in the factual world because it is incompatible with this world).
g. There are many different kinds of t-world, such as worlds that are factual, ‘the-
oretical’ (see below) or counterfactual at t. Situations that are located in such
worlds are represented as actualizing in them. Thus, the situations referred to
by John is ill, If only John was ill! and John is perhaps ill are each time repre-
sented as actualizing in a given S-world, but that S-world is, respectively, the
factual world, a counterfactual S-world (=an S-world which is conceived of by
the speaker as being incompatible with the factual world) and a ‘theoretical’
S-world which is represented as possibly coinciding with the factual world. If
there exist several ti-worlds (=worlds anchored to the same time ti), the situa-
tion referred to can only actualize in one of these ti-worlds. Thus, if it is indeed
the case that John is ill at S, the sentence John is ill represents the situation of
John being ill as actualizing in the factual world. In If only John {was/were} ill!,
the situation denoted by ‘John be ill’ is represented as actualizing in the coun-
terfactual S-world evoked by If only (=I wish...). And in John is perhaps ill, the
same situation is represented as actualizing in the theoretical S-world evoked
by the modal adverb perhaps (=It is possible that...).
h. A situation represented as actualizing in a nonfactual world is automatically
nonfactual in the factual world. Thus, the situation referred to by If only John
was ill! is represented as actualizing in the counterfactual world created by if
only, and hence as having the factuality value [+counterfactual] (rather than
[+factual]) in the factual world. Similarly, John is perhaps ill refers to a theo-
retical S-world (created by perhaps) in which the situation of John being ill
actualizes. The factuality value of this actualization in the factual world is ‘pos-
sibly factual’ (which, as we will see in Section 6.2, is an instance of a ‘relative’
factuality value.) ‘Possibly factual’ is a kind of nonfactuality.
i. All nonfactual worlds (i.e. possible worlds that are not represented and/or in-
terpreted as being the factual world) are created by some kind of ‘nonfactual-
world creating device’. Such a device can be, for example, an intensional verb
like believe, an attitudinal verb like intend, a modal auxiliary, a conditional
clause, a modal adverb like perhaps, etc. – see also Section 4(a). Since a non-
factual world is by definition a modal world (and vice versa – see Section 3),
we call the device in question the modalizer. A proposition underlying a clause
containing a modalizer is a modalized proposition. Thus, in Suppose John were
here, the modalizer is suppose and the modalized proposition is ‘John be here’.
In Bill must leave now, the modalizer is the modal auxiliary must and the mo-
dalized proposition is ‘Bill leave now’.

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

Huddleston (1984: 167–8) suggests the label ‘residue’ for what is left of a clause
with a modal auxiliary when the auxiliary is abstracted from it, viz. the subject
plus the (nontensed and nonmodal) proposition that has the form of an infinitival
clause. Thus, in Tom must swim, ‘Tom swim’ is the residue proposition or simply
‘residue’. We adopt this terminology. The situation denoted by the residue proposi-
tion we call the residue situation. In John may be ill, the residue situation is the state
denoted by the proposition ‘John be ill’. The speaker says that that state is actual-
izing in a nonfactual world that possibly coincides with the factual world, which
comes down to saying that the state in question is possibly actualizing in the fac-
tual world.
The term ‘residue’ corresponds to what we have called the ‘modalized proposi-
tion’, but it is a handier term because it is easier to speak of the ‘residue situation’
than of ‘the situation denoted by the modalized proposition’. We therefore adopt
the terms ‘residue’, ‘residue situation’ and ‘residue proposition’. Moreover, we ex-
tend the use of these terms to those cases in which the modalizer is not a modal
auxiliary. In I wish John was here, the verb wish is the modalizer and the proposi-
tion ‘John be here’ is the residue proposition.
j. The referent of a modalizer is a modal state, i.e. the state of a situation being
necessary, possible, permitted, wished for, etc., or the state of it being possible,
likely, necessary, etc. that a proposition is true in a given world.
(1) That rumor may be true. (The modal auxiliary asserts the existence of the
possibility that the rumor is true. The existence of something is a state. The
modalizer “may” thus refers to a modal state.)
(2) You must leave now. (The modal auxiliary expresses that it is obligatory for
the addressee to leave now. The existence of an obligation is a modal state.)
(3) [Don’t go through the woods.] You could get lost. (The modal auxiliary “could”
expresses the state of it being possible for the addressee to get lost.)
Like any state, a modal state is a situation actualizing in a particular world. In the
above examples, the world in question is the factual world, i.e. the extended
S-world – see Section 2(e). When the modal state is thus represented as actualizing
in the factual world, we can say that the modal world is grounded in the factual
world. This is usually the case, although there are also instances in which one
modal world is grounded in another. Thus, in I might have been killed if I had not
given in the modal suppositional world (created by if) is grounded in the modal
epistemic world (created by might) comprising theoretically possible actualiza-
tions: the sentence means ‘It is possible that I would have been killed if I had not
given in’. Note that this example shows that grounding is a different phenomenon
from anchoring, which is a question of the temporal relation between two worlds

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

– see Section 2(b) above – although the time to which a grounded t-world is an-
chored is usually also the time to which the grounding t-world is anchored. In the
above example, ‘It is possible that’ is anchored to S, while ‘I would have been killed
if I had not given in’ is anchored to a t earlier than S. Both t-worlds, however, are
anchored to the ‘extended S-world’ – see Section 2(e) – which we have agreed to
refer to as the ‘factual world’.
k. It should be stressed that by ‘nonfactual world’ we do not mean a world that is
necessarily different from the factual world but rather a possible world that is
not represented and/or interpreted as being the factual world. This is clear
from John may be here, which refers to a nonfactual world in which the state
of John being here actualizes and which may or may not coincide with the
factual world. A nonfactual world could not be represented as possibly coin-
ciding with the factual world if ‘nonfactual world’ were defined as ‘world that
is necessarily different from the factual world’.

3. The definition of modality

a. Modality can be defined as the phenomenon that a situation is located in a


nonfactual world (as defined above). This means that the residue situation is
represented as actualizing in a nonfactual world, in other words that the prop-
osition underlying the sentence is ‘true of ’ a modal world. The nonfactual
world in question can be related to (grounded in) the factual world in one of
several ways. For example:
(4) [For all we know,] John may be dead. (“May” is a modalizer that refers to a
present state: “It is possible...”. It creates a nonfactual world in which the
state of John being dead is actualizing. This nonfactual world is related to
the factual world in terms of the ‘factuality value’ (see Section 6.1.) ‘pos-
sibly factual’: the world in which John is dead possibly coincides with the
factual world. Hence the interpretation ‘It is possible that John is dead’.)
(5) Val must leave town as soon as possible. (“Must” is a modalizer that refers
to a present state. It creates a nonfactual world in which the situation of
Val leaving town actualizes. This nonfactual world is interpreted as a fu-
ture world because of “as soon as possible”. This modal world is not related
to the factual world in terms of a ‘factuality value’, because “must” here
expresses root (nonepistemic) modality – see Section 5(a). But there is a
relation between the two worlds because the sentence is interpreted as
‘Actualization of Val leaving town in a future nonfactual world is obliga-
tory’. The modal state thus concerns the obligation of the actualization of

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

the residue situation in a nonfactual (more specifically, not-yet-factual)


modal world. In other words, the relation between the two worlds is that
in the factual world the existence of the nonfactual world in which Val
leaves town is deemed necessary.)
b. What has been given in the previous subsection is the definition of modality
in a nutshell. In the following sections further elements are introduced, which
gradually render the definition more detailed and concrete.

4. Kinds of modalizers

a. As we have seen, any linguistic device creating a modal world is a modalizer.


In English, we can distinguish the following types of modalizer: a modal aux-
iliary (e.g. can, must, may...), a modal adverb (e.g. perhaps, possibly, duly, ob-
ligatorily), an intensional verb like believe, suppose, imagine, an attitudinal verb
like intend, want, hope, wish, the subjunctive mood, the imperative mood, a
conditional clause creating a ‘theoretical world’ (see Section 4(d)), a tense aux-
iliary creating a future world (e.g. will, be going to, be about to) or expressing
posteriority, an inserted comment clause with an intensional verb (e.g. I think),
‘modal backshifting’ (see Section 4(b) below) or ‘modal conditionalization’
(see Section 4(c) below) or a combination of the latter two. Some of these mo-
dalizing devices are well-known, others deserve some comment. The latter are
treated in the next subsections.
b. ‘Modal backshifting’ (or ‘formal distancing’) is a basic mechanism producing
either a tentative or a counterfactual interpretation. It is important to see that
this is not the same type of ‘backshifting’ (Jespersen 1931: 151) as occurs in
indirect speech, for example when I am ill is reported as She said she was ill. In
modal backshifting the conditional tense can be replaced by the conditional
perfect, e.g. I would be happy if she came → I would have been happy if she had
come. This kind of ‘backshift’ is never possible in indirect speech, where would
and would have remain unchanged: I would help him if I could → She said she
would help him if she could.
Modal backshifting can produce various results. In some cases only the verb is a
backshifted form. For example, It is time we leave can be modally backshifted to It
is time we left to stress the not-yet-factuality of our leaving. In other cases only the
infinitival clause representing the residue proposition is a backshifted form, as in
He was to have been in his office. The result of the modal backshifting here is that
the actualization of the residue situation is interpreted as counterfactual. Henry
was to have been in his office says that there was a past obligation (resulting from

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

some kind of arrangement or rule) for Henry to be in his office at a given time, but
that he was not there at that time. The existence of the obligation itself is not coun-
terfactual. Another example is the following:
(6) The decision should have been made tomorrow [but now we understand
that there is now another week’s grace]. (Google UK)
In this example should be made is modally backshifted to should have been made.
This does not affect the existence of the obligation but renders the representation
of the actualization of the residue situation counterfactual.
In some cases the result of modal backshifting is that both the infinitive and
the preceding verb are backshifted forms. Thus, He may be absent, which expresses
present epistemic possibility, can be modally backshifted to He might have been
absent, which represents the (past, present or future) actualization of the residue
situation as counterfactual. Similarly, in I {could/might} have been in Iraq now,
both the modal auxiliary and the infinitive are backshifted to create a counterfac-
tual interpretation. The sentence is interpreted something like ‘It would have been
possible for me to be in Iraq now, (but I am not)’. In this case, both the modal state
(root possibility) and the residue situation are represented as counterfactual. The
following are further examples in which both the verb and the verb of the residue
are backshifted forms. (It should be noted, however, that not all speakers of British
English find such instances acceptable.)
(7) I hoped to sit beside her. → I had hoped to sit beside her.
→ I had hoped to have sat beside her.
(8) I would like to see her tonight. → I would have liked to see her tonight. → I
would have liked to have seen her tonight.
There is also the mechanism of double modal backshifting. This means that the
same verb form undergoes modal backshifting twice. Examples of this are largely
restricted to a highly informal register and may be typical of a particular regional
variety of English. For instance:
(9) If he had a car, he would collect us. → If he’d had a car, he’d have collected
us. → If he had’ve had a car, he’d have collected us.
c. Another modalizing device is modal conditionalization. This means that the
present tense form of a verb is replaced by the corresponding conditional tense
form without there being an overt or implicit condition:
(10) [What a nice view!] We might be in Switzerland! (This is interpreted as ‘It
would be possible for us to be in Switzerland’, hence ‘It would seem as if we
were in Switzerland’. This is the conditionalized version of We may be in

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

Switzerland (on the interpretation ‘It is possible for us to be in Switzer-


land’, i.e. ‘The possibility exists that we are in Switzerland’).)
(11) [“John is very happy with his new girl-friend.”] – “It {seems/would seem} so”.
(The addition of would renders the reply more tentative or doubtful.)
(12) I {wish/should wish} to voice a different opinion. (id.)
(13) As a first step to the new life you desperately need, I {suggest/would suggest}
assertion training. (id.)
(14) I {think/should have thought} he could have left something for us, at least.
(Should have thought is an example of double modal conditionalization. It
stresses the counterfactuality of the actualization of the residue situation
and in doing so changes the sentence from a neutral statement into an
expression of criticism.)
d. Another modalizing device is the use of a conditional clause evoking a ‘non-
neutral theoretical world’. In Declerck & Reed (2001), conditionals with such
a clause are called ‘nonneutral theoretical-P conditionals’. This is because ‘con-
ditional clause’ and ‘head clause’ are abbreviated to ‘P-clause’ and ‘Q-clause’,
respectively, given that the standard logical representation of a conditional
sentence is ‘if P, (then) Q’.
Declerck and Reed (2001) propose the ‘possible-world typology of conditionals’
that is represented in Figure 1.
In this typology, P-propositions are either factual or theoretical. The theoreti-
cal-P conditionals are subclassified into ‘neutral theoretical-P’ and ‘nonneutral
theoretical-P’ conditionals. In the former, the P-proposition represents a mere sup-
position, while in the latter the supposition is accompanied by a presupposition
(assumption) concerning the relation between the theoretical P-world and the
factual world. The assumed relation may be that of identity (=closed condition),
possible identity (=open condition), likely nonidentity (=tentative condition) or
definite nonidentity (=counterfactual condition).

Possible P-worlds

Factual Theoretical (nonfactual)

Neutral theoretical Nonneutral theoretical

Closed Open Tentative Counterfactual

Figure 1.

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

In factual-P conditionals, the world referred to in the P-clause is the factual world.
This means that factual P-clauses are not modalizers. The following exemplify the
type:
(15) I (always) avoided her if I could.
(16) [Yet according to the government the number of homeless people in the UK
has now reached 300,000, over 50% of these are under 25 and this figure is
continuing to grow at an alarming rate.] If you add to these those who are at
risk or who have been considered as ‘no hopers’, the number of people who
need our help is truly of crisis proportions. (Cobuild corpus)
A conditional whose P-clause refers to a theoretical P-world – a theoretical-P con-
ditional – may refer to that theoretical world without specifying the relation be-
tween the suppositional world and the factual world. In this case we speak of a
neutral theoretical conditional clause. A speaker who uses a neutral-P conditional
does not convey a specific assumption (presupposition) about the likelihood that
the theoretical-P world coincides with the factual world. The following are some
examples:
(17) If a woman has a history of cancer in her family, she should have herself
checked at least once a year. (Because a woman has nonspecific reference,
the P-clause refers to a set of people in general. Though the speaker natu-
rally assumes that there are people belonging to this set, he does not as-
sume (nor denies, for that matter) that the hearer (or any other specific
person) forms part of the set. There is simply no presupposition that the
application of P to a specific person is true, false, unlikely to be true, or a
real possibility.)
(18) Emergency lighting should be fitted in the hall, if there is no glass in the front
door to let in light. (Cobuild corpus)
(19) If a shop’s good enough, it’s worth making a little effort to get to.
 (Cobuild corpus)
A nonneutral theoretical P-clause evokes a suppositional world which is epistemic
because it stands in a given factuality relation to the factual world. The factuality
values can be of four different kinds, which leads to a distinction between closed-P,
open-P, tentative-P and counterfactual-P conditionals. In these four types, differ-
ent factuality relations are assumed to hold between the theoretical P-world and
the factual world. To begin with, a P-world is ‘closed’ if it is assumed (or ostensibly
assumed) to coincide with the factual world:
(20) [“The picture you are now looking at is a Van Gogh.”] – “Well, if this is a Van
Gogh, I’m rather disappointed by it.”

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

(21) If I’m going to catch the 5.12 train, I’d better be on my way now. (The P-clause
presupposes that I want to catch that train.)
(22) If (as I am told now) he will come here himself tomorrow, I no longer need to
try and get into contact with him today.
In an ‘open-P conditional’, the theoretical world called up by the P-clause is as-
sumed to be such that it may or may not turn out to correspond with the factual
world or the expected future extension of the factual world. It follows that the P-
clause often refers to the future: since the future is unknown, the P-situation and
the Q-situation are easily viewed as an ‘open possibility’ (i.e. as neither certain,
probable nor unlikely).
(23) If the train is late, we will miss our connection in London.
(24) [I’ve been told that] it will be worth a lot more if we can get his signature on
it. (Cobuild corpus)
(25) [I don’t know if Liverpool won their match yesterday.] If they did, they must
be top of the League now.
The basic meaning of a ‘tentative-P conditional’ is that the theoretical P-world is
assumed to be such that it might correspond with the factual world, but that this
possibility is considered as rather unlikely (but not impossible) to be (or to have
been or to come) true.
(26) If he {did/were to do} that, he would be in real trouble. (tentative-P: the P-
actualization is contrary to expectation)
(27) I would be surprised if that man over there were the man we’re looking for.
Finally, in a ‘counterfactual-P conditional’, the theoretical P-world is assumed to be
contrary to fact, i.e. to be definitely different from the factual world (or, if the refer-
ence is to the future, from what is going to become the factual world). Thus, in I
would have been happy if she had come, the P-clause is presupposed to be true of a
counterfactual theoretical P-world but false of the factual world.
In sum, a P-clause is a modalizer creating an epistemic nonfactual world pro-
vided the condition is represented and interpreted as closed, open, tentative or
counterfactual.
e. Some nonfactual t-worlds can be characterized in terms of a modal concept
that has been neglected in the literature on modality, viz. the idea that the t-
world in question is envisaged by the speaker but not yet factual at the time t
to which the world in question is anchored. This kind of modal world, which
we call ‘not-yet-factual at t’, is evoked by any expression that has posteriority as
part of its meaning. The clearest cases are those in which the reference is ex-
clusively to a future world. Thus, John will trim the hedge evokes a t-world

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

which is nonfactual in the specific sense that it is not-yet-factual at S but is


predicted (at S) to become factual at some future time t. The proposition ‘John
trim the hedge’ is true of that not-yet-factual t-world (in which actualization
is envisaged by the speaker), but the not-yet-factual t-world itself is neither
factual nor counterfactual at S: it is an S-world which is partly counterfactual
(because there is no actualization at S) and partly theoretical (because actual-
ization at a time later than S is envisaged at S) – see Declerck (2009) for a more
detailed analysis.4,5
Another type of example in which a situation is located in a not-yet-factual world
is I saw Sam before she had seen me. Here the situation of the before-clause is rep-
resented as not-yet-factual at the time of the head clause situation. The sentence is
therefore interpreted as ‘I saw Sam at a time when she had not yet seen me’, or
‘When I saw Sam, it was not yet a fact that she had seen me’ (which leaves it vague
whether she ever saw me or not). In It’s high time they left and I’d rather they left
‘modal backshifting’ (i.e. the use of left rather than leave – see Section 4(b)) locates
the actualization of the guests’ leaving in a world that is not-yet-factual at S: though
the situation has not actualized yet, its actualization is envisaged for the future. In
I intended to call Bill up when he was at home, both the situation of my calling up
Bill and the when-clause situation are interpreted as not-yet-factual at the time
when my intention was a past fact.

5. Epistemic modality

a. We speak of epistemic modality when the degree of compatibility (or overlap)


between the modal world and the factual world is at stake. Epistemic utter-
ances express the speaker’s evaluation of the relation between the modal world
in which the residue situation actualizes and the factual world. Thus, Bill may
be stuck in a traffic jam expresses that the situation of Bill being stuck in a traf-
fic jam is possibly actualizing in the factual world.

4. Envisage actually refers to something weaker than prediction. It just means that the speaker
reckons with the possibility that a situation will actualize.
5. One might object that Rita will do it tomorrow does not rule out that Rita is already doing
it at S, as witnessed by the fact that we can say Rita will still be doing it tomorrow. However, the
sentences Rita will do it tomorrow, Rita is doing it and Rita will still be doing it tomorrow all refer
to different situations. Doing something tomorrow is not the same thing as doing it now or as
still be doing it tomorrow. In Rita will do it tomorrow, the situation referred to is located, not in
the factual S-world, but in a world that is expected to exist tomorrow. In other words, the situa-
tion is counterfactual in the factual S-world, but at the same time envisaged as possible in a fu-
ture world.

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

By contrast, utterances that are interpreted in terms of root modality (or non-
epistemic modality) are concerned with factors that determine the actualization of
the residue situation, as in the following examples:
(28) You may leave tomorrow. (The speaker is granting permission.)
(29) You {can/cannot} buy tickets in the office over there. (The addressee’s ability
or inability to buy tickets at a certain place depends on a circumstantial
factor.)
(30) The fugitives must leave the country because their visas expired last week.
(The expiration of the visas makes it necessary for the fugitives to leave the
country, because it is illegal to remain in the country without a valid visa.)
In sum, we can say that an epistemic modal world stands in a particular relation to
the factual world or to a future extension of it: there is an epistemic relation be-
tween the two of them. This relation concerns the chances that the epistemic world
coincides with, or will coincide with, the factual world or its future extension. By
contrast, as a rule – see, however, Section 7(b) – a root modal world is not epistem-
ically related to the factual world or to its future extension. It is an ‘epistemically
dangling’ world. Thus, The car can be repaired says nothing about the chances that
the car will be repaired.
b. Within epistemic modality, two major possibilities must be distinguished. In
most cases there is specified epistemic modality. This means that the epistemic
relation between the modal world and the factual world is specified. For exam-
ple, the modal world created by may in Dave may be lying is related to the fac-
tual world in terms of epistemic possibility: the residue proposition has the fac-
tuality value ‘possibly factual’. This means that the situation of Dave being lying
is possibly actualizing in the factual world, in other words that the modal world
(in which the situation actualizes) possibly coincides with the factual world.
When there is specified epistemic modality, one of various ‘factuality values’ on
the ‘epistemic scale’ – see Section 6 – is applicable. Specified epistemic modality
can be indicated by epistemic modal auxiliaries, epistemic modal adverbs, inten-
sional verbs (believe, imagine...), comment clauses with an intensional verb (e.g. I
think), expressions of posteriority (e.g. before in He died before I had seen him),
modal backshifting, modal conditionalization and ‘nonneutral theoretical’ condi-
tional clauses.
The second type of epistemic modality is indeterminate epistemic modality. In
this case the question of the epistemic status of the nonfactual world (i.e. its factu-
ality value vis-à-vis the factual world) is not specified (though it is not irrelevant).
Indeterminate epistemic modality is found in conditionals with a ‘neutral

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

theoretical’ conditional clause – see Declerck & Reed (2001: 72–80). The following
are some illustrations:
(31) If a woman has a history of cancer in her family, she should have herself
checked at least once a year. (subject NP with nonspecific reference)
(32) If water boils, it changes into steam. (general truth)
(33) If she has a problem, she goes to her grandmother. (permanent habit)
(34) Children are orphans if their parents are dead. (set-identifying conditional
clause)
(35) If he noticed Brand, he didn’t comment. (Cobuild corpus) (‘anchoring con-
ditional clause’, i.e. the function of the conditional clause is to anchor the
head clause into the ongoing discourse)
In none of these examples is there an assumption or presupposition on the part of
the speaker that the supposition is treated as true, untrue, unlikely to be true, or as
a real possibility. There is no specified epistemic relation (=relation in terms of a
factuality value) between the suppositional modal world and the factual world.
The same is true of imaginary worlds, which form another instance of neutral the-
oretical suppositional worlds. In the following examples, nothing is specified, im-
plicated or presupposed in connection with the (in)compatibility of the supposi-
tional world with the factual world. For example:
(36) In your place I wouldn’t react if he wrote me a threatening letter. (This sen-
tence neither affirms nor denies the speaker’s not reacting to a threatening
letter. In other words, the head clause situation is presented neither as
factual nor as counterfactual. The utterance is a pure supposition, i.e. a
mere thought experiment.)6
(37) [I didn’t go to the party, so I don’t know if I would’ve become as drunk as you
all appear to have been. In fact, I’m afraid] I would’ve been drunk too, unless
I’d had my wife with me.
(38) [It seemed unlikely that such a frail woman would have been capable of
strangling her husband. However] if it had been her, the clues that would
have been left would have been exactly the same as those we did find. [So we
could not cancel her from our list of suspects.]
(39) [Suppose the door was actually locked.] If the door had been locked, the
burglar would first have tried to open a window, [and that would explain
why we found those footprints in the flower-beds]
(40) [If you’re naughty, a wild dog might come in and] it would eat you.

6. Unlike Harder (1996: 456), who applies the term “thought experiment” to tentative-P and
counterfactual-P conditionals, we speak of ‘thought experiment’ in connection with imaginary-
P conditionals only.

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

As is clear from these examples, imaginary P-conditionals are not formally distin-
guishable from counterfactual conditionals. This explains why imaginary P-condi-
tionals have often been mistaken for counterfactuals.

6. The epistemic scale of specified factuality values

1. ‘Epistemic’ has to do with knowledge. We can say that there is ‘specified epistem-
ic’ meaning when either knowledge or a degree of (un)certainty is expressed
about whether there {is/was/has been/will be} actualization of a situation in the
factual world or in the expected future extension of it. In other words, a clause
has specified epistemic meaning if the speaker expresses his opinion on the
chances that the nonfactual world in which the situation actualizes coincides
with the factual world or its expected future extension. These chances represent
different values on an epistemic scale that ranges from ‘factual’ to ‘counterfac-
tual’. This scale consists of a number of factuality values. The exact number of
these values is irrelevant: we can split the continuum into as many categories as
we wish. However, to account for all the meanings of the modal auxiliaries in
English we need to distinguish between (at least) the following values:7
a. factuality: there is no doubt about the actualization of the (positive or nega-
tive) situation in the speaker’s factual world because that actualization is af-
firmed (asserted as being a fact in that world): John is ill./John is not ill.8
b. strong necessity (strong assumption of truth): [John is not here.] He must be ill.
c. weak necessity (weak assumption of truth): [The money is not in the till, so] it
should already be in the safe.
d. probability: John may well be ill.
e. possibility: John may be ill./It might be true.
f. improbability: It should not be difficult to find his address.
g. impossibility: It cannot be true./You cannot be serious!
h. not-yet-factuality: John will come tomorrow. [I read the letter] before he had
read it.
i. counterfactuality: [You shouldn’t have gone through the woods.] You might have
got lost. Because a counterfactual world is the opposite of the factual world,

7. The ordering of the values on the scale may be subject to debate, but that need not con-
cern us here.
8. It may be worth stressing that, contrary to what is sometimes claimed, negation is not a
modalizer. A negative clause like John is not ill does not express counterfactuality (but rather
what we might call ‘the factuality of nonactualization’.) This means that both John is ill and John
is not ill represent the actualization of a (positive or negative) situation as ‘factual’ – see also Sec-
tion 2(f). Both sentences should therefore be treated as nonmodal sentences.

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

there is no doubt about the actualization of a (positive or negative) situation in


a counterfactual world. (Note that, unlike the actualization of a negative situ-
ation, the actualization of a counterfactual situation is not located in the fac-
tual world. It is located in a counterfactual world, which is defined in relation
to the factual world.)
2. There are some further remarks to be made. Firstly, some of these factuality
values can also be expressed by conditional clauses – see also Section 4(d):
a. Factuality is expressed by ‘factual’ conditional clauses, as in If I had a prob-
lem, I always went to my grandmother.
b. Strong assumption of truth is expressed if the conditional clause is ‘closed’,
i.e. when the condition is assumed by the speaker to be fulfilled, as in If the
work will be done anyhow, I might as well have a lie down or If, as you say, you
were late this morning, I would like you to do some overtime this afternoon.
e. Possibility is expressed by ‘open’ conditional clauses, as in I’ll be relieved if
she phones me.
f. Improbability is expressed by ‘tentative’ conditional clauses, as in I’d be
relieved if she phoned me, Call the police if someone should try to break in or
I’d be very angry if she were to disturb me tomorrow.
i. Counterfactuality is expressed by ‘counterfactual’ conditional clauses like
If I were you... or If I had been there at the time...
Secondly, the factuality values in between the extremes ‘factual’ and ‘counterfac-
tual’ can be referred to as relative factuality values. (In the case of conditionals,
they are the nonneutral theoretical values ‘closed’, ‘open’, and ‘tentative’ – see
Section 4(d).) Unlike the extremes, they all imply some degree of uncertainty on
the part of the speaker: the modal world is represented neither as coinciding with
the factual world nor as being incompatible with it.
Thirdly, of the values on the specified epistemic scale, only the relative factual-
ity values and the counterfactuality value are modal values. Factuality is an
epistemic value, but a world that is represented as factual is not a modal (=nonfac-
tual) world. Obviously, a sentence with the epistemic value ‘factual’ cannot refer to
a nonfactual world. In Section 2(i), ‘nonfactual world’ has been defined as a world
that is not represented as factual.
Fourthly, neutral theoretical conditional clauses, such as those expressing an
‘imaginary’ condition – see Section 5(d) –, do not express a factuality value on the
specified epistemic scale, because the relation between the modal world and the
factual world remains unspecified (indeterminate).
Fifthly, counterfactuality differs from the other values on the specified epistem-
ic scale in that it is not a relative factuality value: it is the only nonfactual epistem-
ic modal value that does not imply a degree of uncertainty about the actualization

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

Specified epistemic scale


Nonmodal Modal
= factual = nonfactual
Purely theoretical Not-yet-factual Counterfactual
= absolute factuality = relative factuality = relative factuality = absolute factuality
value (a) values (b–g) value (h) value (i)

Figure 2.

Epistemic Values

Nonmodal Modal

Indeterminate Specified

Factual Neutral Purely Not- Counterfactual


theoretical theoretical yet-factual

Figure 3.

of the residue situation in the relevant modal world. Because counterfactuality is


defined in relation to the factual world – a situation is counterfactual if it is located
in a ti-world which is the reverse of the factual ti-world – there is no doubt about
the absolute factuality status of counterfactual situations. Thus, in [Kim should not
have gone through the woods.] She might have got lost, there is a clear understand-
ing that Kim did not get lost when she went through the woods. Unlike She may
get lost, the clause She might have got lost does not leave any room for uncertainty.
Figure 2 summarizes the conclusions arrived at in this section in connection
with the specified epistemic scale. (The indications (a), (b-g), (h) and (i) refer to
the examples in Section 6.1.)
This chart can be expanded by adding ‘indeterminate epistemic modality’ –
see Section 5(b). The result is Figure 3.

7. Root (nonepistemic) modality

a. As noted in Section 5(a), sentences that are interpreted in terms of root mo-
dality (or nonepistemic modality) are concerned with factors that determine
the actualization of the residue situation in a nonfactual world, as in the fol-
lowing examples:

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

(41) You may leave tomorrow. (The speaker is granting permission.)


(42) You {can/cannot} buy tickets in the office over there. (The addressee’s ability
or inability to buy tickets at a certain place depends on a circumstantial
factor.)
(43) The fugitives must leave the country because their visas expired last week.
(The expiration of the visas makes it necessary for the fugitives to leave the
country.)
In such sentences no idea is evoked of an epistemic world that stands in a particu-
lar factuality-value relation to the factual world or to a future extension of the
factual world. There is no ‘epistemic relation’ between two worlds, i.e. no relation
that concerns the chances that a given epistemic world coincides with, or will co-
incide with, the factual world or its future extension. In the above examples, the
nonfactual world in which the residue situation actualizes is an ‘epistemically dan-
gling’ world – see also Section 5(b). Thus, The car can be repaired says nothing
about the chances that the car will be repaired.
It follows that we may wonder what is modal about ‘root modality’. Can we not
say that My cat can swim is a simple statement of fact, without any reference to a
nonfactual world? The answer is no. In My cat can swim, the residue proposition
‘My cat swim’ is true of a world that is not represented as the factual world, i.e. of
a nonfactual world. The sentence says nothing about whether the actualization of
the residue situation (which actualizes in that nonfactual world) is factual, doubt-
ful, not-yet-factual, counterfactual, etc. in the factual world. The sentence My cat
can swim may well be true even if my cat has never swum in its life and is never
going to swim. This is because no epistemic relation is expressed between the non-
factual world in which my cat swims and the factual world.
It does not follow, however, that there is no relation whatever between the two
worlds. In the case of root modality, the relation is expressed by the lexical mean-
ing of the modalizer. In The fugitives must leave the country, it is must that ex-
presses the relation between the factual world and the nonfactual world in which
the fugitives leave the country: the link between the two worlds is the modal posi-
tion assumed by the speaker. In this example, the speaker expresses the modal
position that it is necessary for the nonfactual world to coincide with the factual
world. Similarly, in This problem can be solved, the speaker expresses the modal
position that it is possible for the nonfactual world in which the problem is solved
to coincide with the factual world.
It follows that there is only a minimal difference between root modality and
epistemic modality, whereas there is a lot that the two have in common. Let us
start by looking at what they have in common, which is the following:

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

a. The sentence contains a modalizer, e.g. must. (Compare the nonepistemic John
must leave with the epistemic John must be leaving.)
b. The auxiliary functioning as modalizer refers to a modal state (e.g. ‘X is neces-
sary’) which is located in the factual world.
c. The modalizer creates a world which is not represented as factual, i.e. a non-
factual world.
d. The residue situation (denoted by e.g. ‘John leave’ or ‘John be leaving’) is con-
ceptualized as actualizing in the nonfactual modal world.
e. The lexical semantics of the modal auxiliary expresses the relation (e.g. neces-
sity, possibility) between the modal world and the factual world.
The difference between epistemic modality and root modality is determined by
the interpretation-in-context of the auxiliary. It is only when the auxiliary is inter-
preted as an epistemic modal auxiliary that the modal world is an epistemic world
and that the relation between that world and the factual world is an epistemic rela-
tion (e.g. ‘necessarily factual’). In the case of root modality, the relation between
the nonfactual world and the factual one is not an epistemic relation. For example,
John must leave tells us nothing about a factuality relation between the nonfactual
world and the factual world, i.e. about the chances that the two worlds coincide.
The relation that there is is the modal position (e.g. volition, obligation, possibili-
ty) assumed by the speaker. The number of possible modal positions is restricted.
The set minimally comprises root necessity (including obligation) and root pos-
sibility (including ability and permission), probably also volition (willingness) and
possibly some others. The question how many root modal meanings there exist is
a moot question, which we will not go into here.
b. Root modality and epistemic modality are not necessarily mutually exclusive.
Sometimes an epistemic modalizer can be added to a clause expressing root
modality. In such a case the root modal world is no longer an ‘epistemically
dangling’ world, because the root modality is overlaid by epistemic modality.
For example:
(44) The soldiers must guard the ammunition bunker. (root modality, viz.
obligation)
(45) The soldiers must probably guard the ammunition bunker. (Probably ex-
presses a relative factuality value, so that the root modality is overlaid by
epistemic modality.)
Similarly:
(46) According to the arrangement, the decision should be made tomorrow. (root
necessity)

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

(47) According to the arrangement, the decision should have been made tomor-
row. (root necessity overlaid by epistemic modality, viz. a sense of coun-
terfactuality resulting from modal backshifting)
c. Root modality can be created not only by root modal auxiliaries but also by
root modal adverbs (duly, obligatorily...) and by ‘attitudinal verbs’ like wish,
want, etc. (see the next section).

8. World-evoking lexical verbs

There are three types of world-evoking lexical verbs: factive verbs, intensional
verbs and attitudinal verbs. The latter two are modalizers.
a. ‘Factive verbs’ (see Kiparsky & Kiparsky 1970) are verbs like realize, know, ac-
knowledge, etc. which evoke a world (expressed by their complement clause)
which is automatically interpreted as being the factual world:
(48) I {realize/know/acknowledge} that John is a better player than myself.
The world expressed by the that-clause is treated as factual, and hence as non-
modal. Factive verbs are therefore not modalizers.
b. ‘Intensional verbs’ like believe, suppose, think, dream, etc. create an ‘intensional’
world which may or may not coincide with the factual world. These verbs thus
create a ‘specified epistemic modal world’ – see Section 5(b) – which is related
to the factual world in terms of a factuality value. The value in question is
mostly one of the purely theoretical values (e.g. what John believes may or
may not be true, i.e. has the factuality value ‘possibly factual’). However, those
verbs that have an implication of posterior actualization of the residue situa-
tion (e.g. I {foresee/predict} that they will soon surrender./I forgot to talk to her.)
express the value ‘not-yet-factual’, whereas at least one of them, viz. imagine
can imply counterfactuality, as in He imagined he was Napoleon.
In sum, verbs like think, believe, etc. can create a world that is not represented as
being the factual world. The residue situation is conceived of as actualizing in that
nonfactual world. The stronger the intensional meaning of the verb, the less likely
it is for the intensional world to coincide with the factual world. In the case of the
strongest intensional verb (imagine) the intensional world is even normally inter-
preted as counterfactual. The semantics of the verb thus has to do with the likeli-
hood of compatibility with the factual world. This means that such an intensional
verb expresses a kind of factuality value, so that the intensional world can be called
an epistemic modal world. Intensional verbs are therefore epistemic modalizers.

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

c. ‘Attitudinal verbs’ are verbs like intend, wish, hope, expect, etc. that create a
world which is not represented as being the factual world but which has the
characteristics of a root modal world: there is little difference of meaning be-
tween I want him to do it and He must do it. Attitudinal verbs are thus modal-
izers creating a nonfactual world that is related to the factual world in terms of
root modality.
As in the case of root modality expressed by a modal auxiliary, the root modal
interpretation of a world created by an attitudinal verb can be supplemented with
an epistemic modal interpretation. Most of these nonfactual worlds have the
epistemic value ‘not-yet-factual’. Moreover, an epistemic modalizer can often be
inserted into the sentence, as in the following:
(49) Perhaps John wants to leave. (The epistemic modalizer is perhaps.)
(50) I had intended to have left this place earlier. (A counterfactual reading is
produced by modally backshifting intended to had intended and by mod-
ally backshifting leave to have left.)
(51) I wish Jane were here. (The use of the past subjunctive form were creates a
counterfactual interpretation of the residue.)

9. Overview of modal worlds

A nonfactual world has been defined as a possible world that is not represented
as being the factual world. A nonfactual world is by definition a modal world
(and vice versa). The world that is evoked by a ‘factive verb’ (realize, acknowl-
edge, etc.) or by a ‘factual conditional clause’ – see Section 4(d) – is the factual
world, not a modal world. The world that is evoked is a modal world in the fol-
lowing cases:
a. The evoked world is a modal world if the sentence expresses ‘indeterminate
epistemic modality’ – see Section 5(b) –, as is the case in ‘imaginary’ and other
‘neutral theoretical’ conditionals. Indeterminate epistemic modality means that
the question of the epistemic status of the nonfactual world (i.e. its factuality
value vis-à-vis the factual world) is relevant but is not answered because there
is no reference to a specific factuality value on the epistemic modal scale.
b. The evoked world is also a modal world if the sentence expresses ‘specified
epistemic modality’, which means that the question of the epistemic status of
the nonfactual world is relevant and is also answered: we know the factuality
value (on the epistemic scale) that is applicable. Specified epistemic modality
can be indicated by epistemic modal auxiliaries, epistemic modal adverbs,

© 2011. John Benjamins Publishing Company


All rights reserved
The definition of modality 

intensional verbs (believe, imagine...), comment clauses with an intensional


verb (e.g. I think), expressions of posteriority (e.g. before in He died before I
had seen him), modally backshifted indicative forms (e.g. [I wish] it was true),
modal conditionalization (e.g. I would suggest something different), the sub-
junctive (e.g. So be it, then./[I wish] it were true) and conditional sentences
with a ‘nonneutral theoretical’ conditional clause, i.e. a conditional clause ex-
pressing a condition that is ‘closed’, ‘open’, ‘tentative’ or ‘counterfactual’.
c. The evoked world is also a modal world if root modality is expressed. The
relevant relation between the modal world and the factual world can be speci-
fied by a root modal auxiliary, by a root modal adverb (e.g. duly, obligatorily),
by an attitudinal verb like want, intend, etc. or by the imperative mood.

10. Conclusion

In this article we have attempted to give a definition of modality that is as concrete


and as complete as possible. We have argued that there is modality whenever there
is reference to actualization of a situation in a world that is not represented as be-
ing the factual world. All types of modality have been pigeonholed, regardless of
whether the modalizer is a modal auxiliary, a lexical verb, an adverb, a conditional
clause, a morphological operation like modal backshifting or conditionalization,
etc.
Clear definitions have been given of root modality and epistemic modality,
and it has been shown that not all epistemic values are modal values: ‘factuality’ is
one of the two extremes of the epistemic scale but is not a modal value. On the
basis of a rich modal world typology and a relatively great number of possible rela-
tions between a modal world and the factual world, many different subtypes of
modality have been distinguished, such as ‘indeterminate epistemic modality’ ver-
sus ‘specified epistemic modality’, the latter of which can be ‘purely theoretical’,
‘not-yet-factual’ or ‘counterfactual’. ‘Purely theoretical epistemic modality’ has
been shown to comprise a variety of relative factuality values, which in the case of
conditionals give rise to ‘closed’, ‘open’ and ‘tentative’ P-clauses.
It has been shown that root modality and epistemic modality have much in
common, in fact everything except the kind of relation existing between the mod-
al world and the factual world.
World-evoking verbs, which are usually disregarded in treatments of modali-
ty, have been drawn into the picture too. Whereas factive verbs are not modalizers,
intensional verbs like believe or imagine can be used as epistemic modalizers, and
attitudinal verbs like want or wish as root modalizers.

© 2011. John Benjamins Publishing Company


All rights reserved
 Renaat Declerck

References

Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, & Edward Finnegan. 1999. Long-
man Grammar of Spoken and Written English. Harlow: Pearson Education Ltd.
Declerck, Renaat & Susan Reed. 2001. Conditionals. Berlin & New York: Mouton de Gruyter.
Declerck, Renaat. 2009. ‘Nonfactual at t’: A neglected modal concept. In Modality in English, R.
Salkie, P. Busuttil, & J. van der Auwera, (eds.), 31–54. Berlin: de Gruyter.
Harder, Peter. 1996. Functional Semantics: A Theory of Meaning, Structure and Tense in English.
Berlin & New York: Mouton de Gruyter.
Huddleston, Rodney. 1984. English Grammar: An Outline. Cambridge: Cambridge University
Press.
Huddleston, Rodney, & Geoffrey K. Pullum. 2002. The Cambridge Grammar of the English Lan-
guage. Cambridge: Cambridge University Press.
Jespersen, Otto. 1931. A Modern English Grammar on Historical Principles. Part 4: Syntax (3rd
volume): Time and Tense. London: Allen & Unwin/Copenhagen: Ejnar Munksgaard.
Kiparsky, Paul & Carol Kiparsky. 1970. Fact. In Progress in Linguistics, M. Bierwisch & K. E.
Heidolph (eds.), 143–173. The Hague: Mouton.
Leech, Geoffrey. 2006. A Glossary of English Grammar. Edinburgh: Edinburgh University
Press.
Lyons, John. 1977. Semantics. Cambridge: Cambridge University Press.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, & Jan Svartvik. 1985, A Comprehensive
Grammar of the English Language, London: Longman.

© 2011. John Benjamins Publishing Company


All rights reserved
View publication stats

You might also like