An1 Elements of English Morphology Crainiceanu

ILINCA CRĂINICEANU
ELEMENTS OF ENGLISH MORPHOLOGY
1
Descrierea CIP a Bibliotecii Naţionale a României
CRĂINICEANU, ILINCA
Elements of english morphology, Ilinca Crăiniceanu –
Bucureşti, Editura Fundaţiei România de Mâine, 2007
Bibliogr.
200 p.; 20,5cm.
ISBN 978-973-725-759-8
811.111'366
© Editura Fundaţiei România de Mâine, 2007
Redactor: Andreea DINU

Tehnoredactor: Marcela OLARU
Coperta: Marilena BĂLAN
Bun de tipar: 5.02.2007; Coli tipar: 12,5

Format: 16/61×86
Editura Fundaţiei România de Mâine

Bulevardul Timişoara nr.58, Bucureşti, Sector 6
Tel./Fax: 021/444.20.91; www.spiruharet.ro
e-mail: contact@edituraromaniademaine.ro
2
UNIVERSITATEA SPIRU HARET
FACULTATEA DE LIMBI ŞI LITERATURI STRĂINE
ILINCA CRĂINICEANU
ELEMENTS OF ENGLISH
MORPHOLOGY
EDITURA FUNDAŢIEI ROMÂNIA DE MÂINE

Bucureşti, 2007
3
4
CONTENTS
1. ON LEXICAL AND GRAMMATICAL CATEGORIES

THE LEXICON
1.0. Introductory Remarks ………………………………………… 11
1.1. What does Morphology study? ………………………………... 12
1.1.1. Derivational Morphology ………………………………. 12
1.1.2. Inflectional Morphology ………………………………... 12
2.1. Lexical Categories vs. Functional Categories …………………. 14
2.2. The Morpheme ……………………………….……………….. 15
3.1. The Lexicon ……………………………….………………….. 16
3.2. Syntactic Phrases ……………………………….…………….. 17
3.3. The Morpho-syntactic Nature of Functional Categories ………. 18
2. THE CATEGORY OF NUMBER

1.0. Singular-Plural Opposition. Countable-Uncountable Distinction 19
1.1. Morpho-syntactic Reflexes of the Number Category …………. 20
1.2. Sortals vs. Non-Sortals ………………………………………... 20
2.0. On Morpho-Syntactic Properties of Sortals/Non-Sortals ……… 21
2.1. Sortals ………………………………………………………… 21
2.2. Non-Sortals …………………………………………………… 22
2.3. Classifiers ……………………………….……………………. 23
3.0. The Plural Morpheme ……………………………….………... 26
3.1. Collective Nouns. Distributive and Collective Plurals ………… 26
3.1.1. Properties of Collective Nouns …………………………. 27
3.1.2. Agreement Patterns …………………………………….. 28
3.1.3. Collective Nouns and Predicates ……………………….. 29
3.1.4. Collective Nouns in the Plural ………………………….. 29
4.0. Morphologically Defective Nouns ……………………………. 30
5.1. Number Recategorization of Mass Nouns into Countable Nouns 32
6.0. General Terms with No Corresponding Mass Lexicalization …. 34
5
7.0. Plurale Tantum Nouns ……………………………….………... 34
3. THE CATEGORY OF DETERMINATION

1.0. Classes of Determiners ……………………………….……….. 37
2.0. Philosophical Background ……………………………….……. 38
3.0. Characteristics of Indexicals ……………………………….…. 43
3.1. Indexicals or Deictic Elements ………………………………... 44
3.2. Resource Situation ……………………………….…………… 45
4.0. The Values/Uses of Demonstrative Descriptions ……………… 45
4.1. The Deictic/Gestural Value ……………………………….…... 45
4.2. The Symbolic Value ……………………………….…………. 45
4.3. Non-deictic Functions of Demonstratives ……………………. 46
4.3.1. The Anaphoric Value ……………………………….…. 46
4.3.2. The Cataphoric Value ……………………………….…. 47
4.3.3. The Emotional Value ……………………………….….. 47
5.0. Definite Descriptions ……………………………….…………. 47
5.1. General Remarks ……………………………….……………... 47
5.2. The Values/Uses of Definite Descriptions …………………….. 49
5.2.1. The Deictic / Gestural Use of Definite Descriptions ……. 49
5.2.2. The Symbolic/Deictic Use of Definite Descriptions …….. 50
5.2.3. Discourse Functions of Definite Descriptions:
the Anaphoric and Cataphoric Functions ………………. 51
5.2.3.1. General Characteristics ………………………… 51
5.2.3.2. The Anaphoric Value/Use of the Definite Article .. 51
5.2.3.3. The Cataphoric Value/Use of the Definite Article 52
5.3. The Generic Function of Definite Descriptions ……………….. 52
5.4. The Non-Referential (Attributive) Use of the Definite Article … 53
6.0. Proper Names in the Form of Definite Descriptions …………... 53
7.0 Indefinite Descriptions ……………………………….………... 55
7.1. General Remarks ……………………………….……………... 55
7.1.1. The Ambiguity of Indefinite Descriptions ……………… 56
8.0. The Values/Uses of Indefinite Descriptions …………………... 58
8.1. The Epiphoric Value of Indefinite Descriptions ………………. 58
8.1.2. The Numerical Value of Indefinite Descriptions ……….. 58
8.1.3. The Generic Value of Indefinite Descriptions …………... 59
9.0. Generic Sentences ……………………………….……………. 59
10.0. A Note on Bare Plurals (BPs) ………………………………... 60
10.1. Semantic Differences between Indefinite DPs and BPs ……… 60
10.1.1. Anaphoric Processes ……………………………….… 60
10.1.2. Specific vs. Non-specific Readings …………………... 61
6
10.1.3. Predicates that Occur only with BPs …………………. 61
11.1. Bare Plural Generic Sentences ……………………………….. 62
11.2. Definite Article Generic Sentences ………………………….. 63
11.3. Indefinite Article Generic Sentences ………………………… 64
4. THE CATEGORY OF GENDER

1.0. General Remarks ……………………………….……………... 66
2.0. The Gender of Animate Entities ………………………………. 67
2.1. The Gender of Male/Female Beings ………………………….. 67
2.1.1. Nouns that Lack Gender Specification …………………. 69
3.0. Referential Gender …………………………………………….. 70
5. THE CATEGORY OF CASE

1.0. General Remarks ……………………………….…………….. 72
2.0. Formal Configurations: Government and Agreement.
Structural Cases ……………………………….………………. 73
3.0. Remarks on the English Cases in Traditional Grammar ………. 74
4.0. Case as a Conceptual Notion: Case Grammar ………………… 79
4.1. General Remarks ……………………………………………… 79
4.2. Predicates and their Argument Structure ……………………… 79
4.3. Semantic Fields of Several Verbs of Motion and Location ……. 82
4.4. On the Relevance of Theta-Roles in Syntax and Morphology …. 83
6. THE CATEGORY OF ASPECT

1.1. The Perfective – Imperfective Grammaticalized Aspectual
Opposition ……………………………………………………. 87
1.1.2. Grammaticalizations of the Perfective – Imperfective
Aspectual Opposition …………………………………... 87
2.1. Situation-Type Aspect ………………………………………… 88
3.1. Situation-Type Aspect versus Grammatical Aspect …………… 88
4.1. Temporal Structure and Aspectual Situation Types …………… 89
4.1.1. States …………………………………………………… 90
4.1.2. Processes/Activities …………………………………….. 90
4.1.3. Events ………………………………………………….. 90
5.0. The Temporal Structure of the Perfective – Imperfective
Aspectual Opposition …………………………………………. 91
6.0. Conceptual Features of Situation Types and Grammatical /
Viewpoint Aspects ……………………………………………. 92
7
6.1. [± Stativity] …………………………………………………… 93
6.2. [± Telicity] …………………………………………………… 93
6.3. [± Duration] …………………………………………………... 93
6.4. Compositionality ……………………………………………… 94
7.0. Aspectual Recategorization / Shift ……………………………. 94
8.0. The Perfective Viewpoint and Aspectual Situation Types ……... 98
8.1. General Characteristics of the Perfective Grammatical /
Viewpoint Aspect ……………………………………………... 98
8.2. States, Activities and Events in the Perfective Aspect …………. 99
9.0. The Imperfective Viewpoint and Aspectual Situation Types ….. 100
9.1. General Characteristics of the Imperfective Grammatical /
Viewpoint Aspect ……………………………………………... 100
9.2. Jespersen’s (1933) View on the Properties of the English
Progressive ……………………………………………………. 101
9.3. Bennett and Partee’s (1972/1978) Formalization of the
Progressive Aspect ……………………………………………. 103
9.4. The Imperfective Paradox (Dowty 1979) ……………………... 104
10.0. States in the Progressive Aspect ………………………………. 106
11.0. Activities / Processes in the Progressive Aspect ………………. 109
12.0. Events in the Progressive Aspect ……………………………… 110
7. THE CATEGORY OF TENSE

1.0. General Remarks ……………………………………………… 113
1.1. The Notion of Axis of Orientation …………………………….. 113
2.0. The Notions of Speech Time, Reference Time and Event Time .. 115
3.0. The General Classification of Temporal Adverbs and Their
Temporal Specification ……………………………………….. 119
3.1. Duration Adverbs ……………………………………………... 119
3.2. Completive Adverbs ………………………………………….. 122
3.3. Frequency Adverbs …………………………………………... 122
3.4. Locating Adverbs (or Frame Adverbs) ……………………….. 122
4.0. Temporal Values of the Main Tenses in English ……………… 123
4.1. General Properties of the English Simple Present Tense ………. 123
4.3. On the [+Perfective] Feature of English Simple Present Tense ... 127
4.4. Other Values/Uses of the Simple Present Tense ………………. 128
4.4.1. Generic and Habitual Sentences in the Simple Present Tense 128
4.4.2. The Instantaneous Use of the Simple Present Tense …….. 129
4.4.3. The Historical Value of the Simple Present Tense ……… 130
8
5.0. The Temporal Values / Uses of the Simple Past Tense ………... 131
5.1. The Simple Past Tense with Deictic Value ……………………. 131
5.2. The Non-Deictic Use of the Simple Past Tense ……………….. 133
5.3. The Habitual Value of the Simple Past Tense ………………… 133
5.4. The Simple Past Tense with Past Perfect Value ………………. 134
5.5. The Simple Past Tense Referring to Present Time ……………. 134
6.0. The Temporal Values / Uses of the Simple Present Perfect Tense 135
6.1. General Properties of the English Present Perfect Tense ……… 135
6.2. The Indefinite Past Theory on the Present Perfect …………….. 137
6.3. The Current Relevance Theory on the Present Perfect ………… 138
6.4. The Extended Now Theory on the Present Perfect ……………. 139
6.5. The Values / Uses of the Simple Present Perfect ……………… 140
6.5.1. The Existential Value of the Present Perfect ……………. 140
6.5.2. The Resultative or Continuative Value of the Present
Perfect …………………………………………………. 141
6.5.3. The ‘Hot News’ Value of the Present Perfect …………… 142
7.0. Other Temporal Uses of the Present Perfect …………………… 142
7.1. The Future Value of the Present Perfect ………………………. 142
8.0. Temporal Adverbs with the Present Perfect and the Past Tense .. 143
9.0. The Temporal Values / Uses of the Past Perfect ……………..... 146
9.1. General Properties of the Past Perfect Tense in Simple Sentences 146
9.2. General Properties of the Past Perfect Tense in Complex Sentences 146
10.0. Means of Expressing Future Time …………………………….. 147
10.1. General Characteristics ………………………………………... 147
10.2. The Simple Present Tense with Future Time Adverbs ………… 148
10.3. The Present Progressive with Future Time Adverbs …………… 148
10.4. Will and Shall plus the Infinitive ……………………………… 149
10.5. Will and Shall plus the Progressive Infinitive …………………. 150
10.6. Be Going To …………………………………………………... 150
10.7. The Future of Past Situations ………………………………….. 151
11.0. The Temporal Values / Uses of the Future Perfect Tense ……... 152
8. MODALITY AND MODAL VERBS

2.0. On Root and Epistemic Meanings of Modal Verbs …………… 155
3.0. More on the Deontic Meaning and Epistemic Meaning and the
Shift from One Meaning to the Other Meaning ………………. 159
4.0. Morpho-syntactic Properties of Modal Verbs …………………. 163
5.0. CAN …………………………………………………………. 168
5.1. General Remarks ……………………………………………… 168
9
5.2. Ability CAN (root meaning) ………………………………….. 170
5.3. Permission CAN (root meaning) ……………………………… 172
5.3.1. Hypothetical COULD ………………………………….. 174
5.4. Possibility CAN (epistemic meaning) …………………………. 174
5.4.1. Hypothetical COULD ………………………………….. 175
6.0. MAY ………………………………………………………….. 176
6.1. Permission MAY (root meaning) ……………………………... 176
6.1.1. Hypothetical MIGHT ………………………………….. 177
6.2. Possibility MAY (epistemic meaning) ………………………... 177
6.2.1. Hypothetical MIGHT ………………………………….. 178
6.3. A Parallel between CAN and MAY …………………………... 178
7.0. MUST ………………………………………………………… 179
7.1. Obligation MUST (root meaning) …………………………….. 179
7.1.1. Negation of Obligation MUST …………………………. 181
7.2. Necessity MUST (epistemic meaning) ………………………... 182
8.0. WILL …………………………………………………………. 184
8.1. Volition WILL (root meaning) ………………………………... 184
8.1.1. Hypothetical WOULD …………………………………. 185
8.2. Power WILL (root meaning) ………………………………….. 186
8.3. Habitual WILL (root meaning) ………………………………... 187
8.4. Probability WILL (epistemic meaning) ……………………….. 188
9.0. SHALL ……………………………………………………….. 189
9.1. General Remarks ……………………………………………… 189
9.2. Obligation SHALL (root meaning) ……………………………. 189
9.2.1. Hypothetical SHOULD ………………………………… 190
9.3. Probability, Prediction SHALL (epistemic meaning) ………….. 191
9.3.1. Hypothetical SHOULD ………………………………… 191
10.0. OUGHT TO …………………………………………………. 192
10.1. Moral Obligation OUGHT TO (root meaning) ……………… 192
10.2. Weak Necessity OUGHT TO (epistemic meaning) ………….. 193
11.0. NEED ……………………………………………………….. 194
11.1. Obligation NEED (root meaning) ……………………………. 194
12.0. DARE ……………………………………………………….. 195
12.1. Hypothetical DARE …………………………………………. 196
References ………………………………………………………… 197
10
1. ON LEXICAL AND GRAMMATICAL CATEGORIES
THE LEXICON
1.0. Introductory Remarks

Linguistics is the science that studies language. “Like any science, it
attempts to systematize and explain a domain of the empirical world”
(Cornilescu 1995:17). People speak languages, and linguistics is the
empirical science that describes them in terms of their various properties.
The properties that a language evinces are outlined in the grammar of that
language. Grammar is interested in the general, recursive sets of rules that
operate in a language. The relation between linguistics and grammar is
intimate, as in current opinion grammar is associated with teaching
practices (cf. Cornilescu 1995).
Roughly speaking, languages are made up of words that combine to
form sentences. In turn, words are made up of sounds (more technically
of phonemes) whose properties are studied by one of the branches of
linguistics, viz. Phonetics and Phonology.
The general conception is that language is organized into
hierarchical levels (e.g., the phonological level, the morphological level,
the syntactic level, the semantic level) and that there are relations that
hold between these levels of language.
An individual linguistic unit (e.g., a phoneme, a word, a sentence)
has linguistic significance only if it is the constitutive part of an
immediately higher level of analysis. Thus, the phoneme is of linguistic
interest because strings of phonemes make up words and words
themselves (if they are complex, e.g. childhood) are decomposable into
other linguistic units called morphemes (e.g., child-hood) (see section 2.2.
below). In turn, strings of morphemes make up syntactic phrases (e.g.,
the + boy, walk + s, see section 3.2. below) and syntactic phrases make
up sentences.
11
Generalizing, the concern of linguistic analysis will not be the
description of linguistic levels (e.g., the phonological level, the
morphological level, the syntactic level) in an atomistic fashion.
Linguistic analysis will be concerned with the study of the relations of
compatibility and isomorphism (defined as something identical with or
similar to something else in form or structure) between these levels in a
holistic perspective on language (cf. Ştefănescu 1984).
1.1. What does Morphology study?

Morphology is the branch of science that deals with the form and
structure of an organism or its parts (etymologically, it is a Greek word:
morph– + logie (-logy)). In linguistics, morphology is the branch that
studies the structure of words. Words may be simple (e.g., girl, walk,
rational) or complex (e.g., girl+s, walk+ed, ir+rational).
The complexity of word structure is due to two morphological
operations: derivation and inflection. Both operations add extra elements
(called markers) to the ‘naked’ word, which is called the base.
1.1.1. Derivational Morphology

Derivational operations are processes such as prefixation
(e.g., happy – unhappy), suffixation (e.g., sense – sensation – sensational)
and compounding (e.g., heartbreaking, sunflower). As can be noticed,
derivational markers can change/shift the lexical class of the base. For
example, sense is a noun while sensational is an adjective. Similarly,
realize is a verb while realization is a noun; professor is a noun while
professional is an adjective. Derivational words and compound words are
lexical items listed separately (in dictionaries).
1.1.2. Inflectional Morphology

Inflectional operations, which are changes in the word that indicate
case, number, gender, person, tense, aspect, mood and comparison
markers do not shift the lexical class/category of the words to which they
attach. For example, boy is a noun and boy-s is also a noun; jump is a verb
and jump-ed / jump-ing are also verbal forms.
12
Two important properties distinguish inflectional morphology from
derivational morphology.
Firstly, inflectional morphology is organized in paradigms while
derivational morphology registers many irregularities and idiosyncrasies.
For instance, each English verb exhibits a paradigmatic set that contains
four verbal inflected forms:
(1) V V-s V-ed V-en V-ing
jump jumps jumped jumped jumping
sing sings sang sung singing
Nearly all English nouns exhibit a number paradigmatic set of two

inflectional forms (singular-plural):
(2) N N-s
house houses
In English, there are few exceptions to this tight paradigmatic
organization of inflected forms. One of them is suppletion. That is, one or
several forms of some elements in the paradigm are not phonologically
related to the other forms in the paradigm, e.g., go: went / gone, be: am /
are / is / was / were. The forms went / gone, and am / are / is / was / were
are said to be suppletive forms of go and be, respectively. Moreover, a
paradigm can also be defective that is, it may contain elements that lack
one of the forms, e.g., the number paradigm: measles – *measle; mumps
– *mump; trousers – *trouser.
The second important property that characterizes inflectional
morphology is that an inflectional element once attached to the base form
of a word closes off the respective lexical item. The following examples
show that after the attachment of an inflectional element to a word no
other inflectional or derivational element can be attached to that word:
(3) boy boyhood
boys *boyshood
(4) jump jumps
jumps *jumpsed
In sum, inflection elements, technically called inflection markers or
grammatical markers attach to lexical items. Grammatical markers are
13
not separately listed in dictionaries since they do not function
independently from lexical items. Of course, suppletive forms are listed
separately and each suppletive form is specified for the context in which
it occurs (e.g., went, the past tense of the verb go).
From another perspective, inflection markers are formal markers
that help us identify the lexical class a word belongs to (in terms of form
not meaning). For instance, jump belongs to the category verb if it is
marked by –ed or –ing (jumped, jumping) but it is a noun if it is preceded
by the article the (the jump).
On formal criteria, lexical items fall into four major classes: nouns
(N), verbs (V), adjectives (A) and prepositions (P). These lexical classes
of words are what traditional grammars called parts of speech. In recent
studies of linguistics they are referred to as lexical categories (beginning
with Chomsky’s “Aspects of the Theory of Syntax” 1965).
In sum, grammatical markers attach to lexical categories and, in
connected speech, one cannot function in the absence of the other. After
all, when we learn to speak a language we learn both words (lexical
categories) and their specific grammatical markers.
2.1. Lexical Categories vs. Functional Categories

Aristotle and his followers labelled both lexical and grammatical
items ‘categories’. The naïve assumption was that the structure of language
reflects the structure of the world and that language consists of elements
(i.e., ‘substances’) which have certain properties (i.e., ‘accidents’).
Language words signified things according to their modes of ‘being’, as
substances or accidents. The categories of substance were later on
assimilated by traditional grammarians to the parts of speech (nouns,
verbs, adjectives and prepositions). The parts of speech have ‘meaning’
and they designate ‘objects’ in the physical world. As already said, in
modern grammar the parts of speech are called lexical categories
(Chomsky 1965, 1970, 1981, 1995).
Aristotelian ‘substances’ were classified with respect to their
‘accidental’ properties that is, in terms of their form. Certain ‘accidental
categories’ were considered to be typical for particular parts of speech:
nouns are inflected for case (e.g., nominative, accusative), number
(singular, plural) and gender (e.g., masculine, feminine) while verbs are
14
inflected for tense, number or person. ‘Accidental’ categories are abstract
categories and they do not have ‘meaning’, as they do not designate
‘objects’ in the real world. The Aristotelian accidental categories are what we
call today grammatical markers or functional categories (Chomsky 1981).
In the Aristotelian line of thought, traditional grammarians also
distinguished between major and minor parts of speech. Only the major
parts of speech (nouns, verbs, adjectives, adverbs, prepositions and
particles) were truly meaningful, while the other parts of speech
(determiners (e.g., the, this, a), quantifiers (e.g., many, much, few, a few),
conjunctions (e.g., and, but) known as minor parts of speech did not
signify anything by themselves. They merely contributed to the total
meaning of the sentence by imposing upon it a certain ‘form’ or
‘organisation’ (cf. Baciu 2004). The minor parts of speech are nowadays
assimilated to the class of functional categories as well.
In spite of the fact that both lexical items/lexical categories and
grammatical markers/functional categories are designated by the term
category, there are great differences between lexical categories and
functional categories. In a nutshell, the differences between the lexical
and grammatical categories are:
Lexical Categories:
• They form open classes of items: the number of lexical
categories is very large and can be made even larger by borrowing and
word formation processes.
• Lexical categories have descriptive meaning. For instance,
nouns designate classes of objects while verbs designate classes of events.
Functional Categories:
• They form closed classes of items. For example, the number of
tense markers is limited to -s and -ed in English.
• They lack descriptive meaning and are very idiosyncratic. For
example, the definite article the and the indefinite article a are distribu-
tionally very different and no meaning can be attached to them in isolation.
2.2. The Morpheme
The basic unit of analysis in morphology is the morpheme. The
morpheme is the minimal unit endowed with constant form (acoustic
15
image) and constant meaning. From this perspective, in grammar, we
speak of grammatical morphemes (e.g., -s, -ed, -ing) and lexical
morphemes (e.g., boy (N), write (V), red (A), on (P)).
Morphemes can be bound or free. Bound morphemes never occur
in isolation (for example, morphemes that form the classes of prefixes
such as un-, ir– and suffixes such as -s, -ed). Free morphemes can occur
in isolation (e.g., the, a, girl, read, can, may, shall).
3.1. The Lexicon

When we know a language we know when and how to use both
kinds of morphemes, lexical and grammatical. This knowledge is stored
in our mental Lexicon. The Lexicon is the mental dictionary of a
language: it contains the lexical and grammatical items of a language and
information about them. Lexical and grammatical items include a
collection of properties. We illustrate the collection of properties evinced
by some lexical items (cf. Chomsky 1995).
Take, for instance, the lexical items book and run. Some of these
properties cover aspects of phonology and morphology, which are
language specific, and thus idiosyncratic. The fact that book is
pronounced /b ü k/ and run is pronounced /r^n/ or the fact that the plural
form of book is books and the present tense of run is (he) runs are specific
to English. Lexical items also include information with respect to their
semantic properties. When we know to use the word book as a noun we
necessarily know what it means (i.e., it designates a certain kind of
object) and when we know what run means we know that it designates a
certain type of activity. Lexical items also list their properties with respect
to the categorial features they have: for instance, the fact that the lexical
item book has the categorial feature noun [N] and that the lexical item run
has the categorial feature verb [V].
However, we said above that in discourse lexical categories
function together with functional categories. Let us see how this
‘functioning together’ works. It is the case that to each lexical category
there corresponds at least one functional category. For instance, the
lexical category N has Determiner (Number and Case) as functional
categories. The lexical category V has Auxiliary as a functional category;
the auxiliary contains the markers of Tense, Aspect and Mood of the verb
16
(e.g., He has come; He may come). The lexical category A has Degree
words as functional category (e.g., the tallest). P and Prt have nothing as
functional category.
Let us illustrate this with some examples. The N ‘horse’ designates
the class of objects that have the property of being horse. The functional
category Det (i.e., the + horse) fixes the object we speak about in the class
of ‘horse’. Thus, the lexical item horse ‘functions’ together with the
functional category Determination (i.e., the + horse). The V designates a
class of events (e.g., Swimming in the pool is fun). The functional
category for verbs is Auxiliary, which includes Tense. When we utter a
sentence such as: I swam in the pool last night, we refer to one particular
event of swimming that took place last night. Thus, functional categories
close off lexical categories making it possible for lexical items to function
in discourse. For example, the following sentences, which contain lexical
categories but lack functional categories to articulate them in discourse,
are ill formed: *He break glass; *Child be hungry.
Thus, the term functional category designates the characteristic
inflectional variation/paradigm of a certain part of speech. Nouns are
characterized by the paradigms of the categories of case, number, gender
and determination; verbs are characterized by the paradigms of the
categories of tense, aspect and mood; adjectives and adverbs are
characterized by the category of comparison; prepositions and particles
are characterized by no inflectional paradigms.
3.2. Syntactic Phrases

Lexical categories together with their specific functional categories
form syntactic phrases (annotated as P(hrases)). That is, each lexical
category (word) grows (better said projects) into a syntactic phrase. In
syntax, lexical categories are taken as heads (i.e., N0, V0, A0, P0) that is, as
obligatory constitutive element of a phrase. The properties of a head
determine the structure of the phrase. For instance, the lexical category N
projects as an NP syntactic category (e.g., N (boy) – NP (the boy)) while
the lexical category V projects as a VP syntactic category (e.g., V (walk)
– VP (walked)). The property of lexical heads to project as syntactic
phrases is not a property specific to English but holds for all languages
spoken by people.
17
The property of lexical categories to grow (project) into phrases is
important not only because this is how connected speech functions but
also because meaning is attached only to syntactic phrases.
In other words, only full syntactic phrases (NP, VP, AP, PP)
correspond to semantic categories not just lexical categories (N, V, A, P).
As we are going to see when we discuss the category of aspect or the
category of number, for instance, it is syntactic phrases that are in a
relation of correspondence with semantic aspectual categories such as
events (e.g., John ate an apple (VP)), processes (e.g., John ate apples
(VP)), states (e.g., John is tall (VP)) or individual objects (e.g., John is a
young man (DP)), respectively.
3.3. The Morpho-syntactic Nature of Functional Categories

It is also important to notice that functional categories are morpho-
syntactic in nature: some syntactic relations that a word contracts with
another word in a sentence trigger a change of inflectional form of that
word. For example, the presence of singular or plural number morpheme
on the noun that syntactically functions as subject determines singular or
plural form on the predicate (e.g., The girl has called me / The girls have
called me). Another example is case marking on nouns. For instance, the
verb give triggers the presence of accusative and dative cases on its
objects (to give something to somebody).
In other words, functional categories (case, gender, number,
determination, tense, aspect, mood inflectional morphemes) encode
abstract syntactic information about the relations among lexical categories
when they occur in sentences.
3.4. The present textbook in morphology adopts the theoretical
framework of various versions of Generative Grammar (Chomsky’s
Aspects of the Theory of Syntax (1965), the Extended Standard Theory
(1970), Principles and Parameters (1981), the Minimalist Program
(1995)). It provides a detailed description of the following functional
(morphological) categories: the category of number, the category of
determination, the category of gender, the category of case, the category
of tense, the category of aspect and the category of modality and modal
verbs.
18
2. THE CATEGORY OF NUMBER
2.0. Singular-Plural Opposition. Countable-Uncountable Distinction

The category of number is related to the various ways that language
uses to individuate entities.
For example, common nouns such as man, house, and rabbit
designate entities that can be individuated and counted. The entities
designated by man, house, and rabbit have characteristic spatial shapes
and are bound in space. Notice that parts of these objects do not constitute
the object itself (i.e., a man’s arms or legs do not constitute a man).
On the other hand, common nouns such as iron, gold and ice
designate entities that are continuous in space and cannot be individuated.
They do not have a characteristic shape in space. Moreover, any part of
the entities designated by iron, gold and ice is still iron, gold or ice.
Traditional grammars deal with the category of number as reflected
in (i) the opposition singular – plural (e.g., boy – boys) and in (ii) the
distinction between countable – uncountable (mass) nouns (e.g., table/tables
– wine).
The first opposition is ‘grammatical’ in nature: one of the two terms
of the opposition, namely the plural one carries a morphological marker
(usually -s). The second distinction closely related to the first, is
‘semantic’ in nature and has to do with the way nouns denote entities in
the physical world. Countable nouns denote entities that can be divided
and thus counted while uncountable nouns denote entities that cannot be
divided and thus cannot be counted (Baciu 2004). The semantic feature of
the category of number is [+divisibility].
19
1.1. Morpho-syntactic Reflexes of the Number Category
There are at least three morpho-syntactic reflexes of the category
number on a countable nominal:
• number concord between the noun and its determiner (e.g., this
boy vs. these boys),
• subject-verb number agreement (e.g., That car is blue vs. Those
cars are blue) and
• pronoun-antecedent number agreement (e.g., I have seen the
boys. They were running in the street) (cf. Baciu 2004).
1.2. Sortals vs. Non-Sortals

Common countable nouns like man, house, and rabbit are also
called SORTALS. Uncountable / mass nouns like iron, gold and ice are
also called NON-SORTALS. The sortal / non-sortal distinction
corresponds to the COUNT / MASS distinction of nouns. “The purpose
of the sortal distinction was to be able to apply number to it in a definite
manner and not to permit any arbitrary division of the sortal term. Non-
sortals do not allow number to apply to them and arbitrary division into
parts is an identification test” (Pelletier, 1979: 19).
There are two other properties that distinguish sortals from non-
sortals.
Non-sortal (mass) nouns evince the properties of subdivisibility and
additivity while sortal (countable) nouns evince the properties of anti-
subdivisibility and anti-additivity.
For example, the division of a lump of gold will result into smaller
lumps of gold (the property of subdivisibility of non-sortals) and the
addition of a lump of gold to another lump of gold will form a larger
lump of gold (the property of additivity of non-sortals). In contrast, the
division of a sortal / count entity such as table into its top, legs and nails
does no longer result into a table (the property of anti-subdivisibility).
Sortals such as man, house and rabbit are also anti-additive, i.e., several
men do not form a larger man.
20
2.0. On Morpho-syntactic Properties of Sortals/Non-Sortals
The semantic properties that distinguish between sortals /
non-sortals have significant morpho-syntactic reflexes (cf. Ştefănescu
1988, Baciu 2004).
2.1. Sortals
Sortals (i.e., countable nouns) co-occur with:
• the indefinite article a
• the plural morpheme –s
• the quantifiers each, every, many, few
• take cardinal numerals
• they trigger plural verb agreement and plural anaphoric
pronouns
(1) a. Please, tell me isn’t there a ghost?

b. Many girls seem to be so ignorant.
c. Isabel, bending towards one picture after another, indulged
in little exclamations.
d. They have broken their promises.
The most important semantic property of countable nouns is that

they have individualized reference (since they can occur with the
indefinite article a). The fact that sortals divide their reference accounts
for the application of number to them.
Notice however that it is not number alone that individuate
countable nouns. Individuation is the combined function of the quantifier
count nouns occurs with (e.g., a, each, every, many, few, several) and of
the category number (Cornilescu p.c.).
Neither quantifier alone nor the plural marker by itself can
individualize entities. Consider the following examples:
(2) a. What does he do for a living?

b. John writes books.
c. In the classroom there were zero students.
21
In (2a) a living, in spite of the fact that it co-occurs with the
indefinite article (quantifier) a, a does not have an individuation function.
Similarly the plural countable nouns books in (2b) and students in (2c) do
not individualize their referents. In (2b) the plural marker on books does
not guarantee that there is any book John has written.
Countable or sortal nouns are also known as general terms (Quine
1960). They are set in contrast with singular terms such as mother, John,
the chair I sit on. While general terms name classes of objects, singular
terms name unique objects (Quine 1960).
2.2. Non-Sortals
Non-sortals (i.e., uncountable nouns) have the following morpho-
syntactic properties:
• resistance to pluralization
• resistance to co-occurrence with the indefinite article a
• co-occurrence with (amassive) quantifiers such as much, little,
a lot of, an amount of
• singular agreement with the verb and singular anaphoric
pronouns
(2) a. Water was brought to the rescued party.
b. She has little confidence in him.
c. I have had some cold meat / Was it good?
Conceptually, uncountable nouns designate indivisible substances
such as gas, water, footwear, gold, sugar, etc. Uncountable / non-sortal
nouns do not have criteria of individuation (they are only subdivisible and
additive).
Uncountable nouns do not show number variation. As a rule, they
exhibit only the (unmarked) singular form. They can be individuated with
the help of classifiers (e.g., a drop of water, a lump of sugar, a stock of
hay, etc) (see section 2.3. below).
From a semantic point of view non-sortal/mass nouns are distinct
from both general terms and singular terms. Unlike general terms, which
individuate their reference mass terms do not individuate their reference.
The opposition between mass terms and singular terms (e.g., London, the
book on the table) is established with respect to their ability to name a
22
unique object: singular terms do name unique objects while mass terms
are unable to do so.
2.3. Classifiers
Not all languages have morphologically marked number (e.g.,
Chinese, Japanese or Korean). To individuate entities languages that
have no morphologically marked number make use of classifiers.
Consider the following examples from Chinese:
(3) a. Yi li me
one CL rice
‘one grain of rice’ (-count)
b. Yi ge ren
one CL person
‘one person’ (+count)
Languages like English (and Romanian) use classifiers for dividing

into parts uncountable nouns (e.g., a grain of rice, un bob de orez). It is
clear that the meaning of morphological number is the same as the
meaning of the classifier. A classifier shows that a given substance can be
partitioned out into units of a certain type (Cornilescu p.c.)
As already shown, conceptually, countable nouns/sortals have
criteria of individuation while uncountable nouns/non-sortals lack criteria
of individuation. In English, uncountable nouns/non-sortals make use of
classifiers (roughly paraphrased as ‘part of’ a substance) to yield units of
a certain type; for instance, portions of water are individuated by means
of lexical items such as a lake, a sea, a pool, a drop.
In terms of morphological marking we have seen that only the
plural member of countable nouns is morphologically marked (usually by
-s). Being morphologically marked the interpretation of plural countable
nouns becomes fixed (i.e., they cannot be reinterpreted as uncountable
nouns in a different context). In exchange, the singular, countable
unmarked member of the singular-plural opposition becomes flexible in
interpretation (Borer 2004). This flexibility makes it possible for singular,
countable nouns to appear in two different configurations. Consider the
following examples:
23
(4) a. There are two apples on the table. (+count)
a’. There is too much apple in the salad. (-count)
b. There was a cat on the road. (+count)
b’. There was much cat on the road after the accident. (-count)
The conclusion is that the distinction [+/-count] is independent

from any objective division of structure (Borer 2004). The same structure
can be named either by a [+count] or a [-count] noun and conversely, the
same noun has [+/-count] interpretations. For example, shoes, sandals are
countable nouns while footwear, which designates the same objects is
uncountable. Predominantly, [-count] nouns can be made [+count]
acquiring predictable interpretations (e.g. wine – wines (= kinds of wine)).
Conversely, [+count] nouns in the singular can be made [-count] (e.g.,
There are two chickens in the backyard – There is too much chicken in
my soup) (see the sections below for a detailed discussion of the types of
shifts in countability interpretation).
Thus, the contrast between [+count] vs. [-count] lies in the terms
used and not in the stuff they name (cf. Ştefănescu 1988). “There is
clearly no difference between noodles and spaghetti that can be held
responsible for the fact that noodles is a plural countable noun but
spaghetti is a mass noun, nor is there any difference between garlic and
onions or between rice and beans. The same entities can be described as
footwear or as shoes, as furniture or as chairs” (McCawley 1975: 314).
Jespersen (1931) noticed that one and the same noun can evince
both general term behaviour (i.e., + countable noun) or mass behaviour
(i.e., -countable noun). Consider the following examples set in parallel
columns:
a tin of sardines an alloy of copper and tin

two big cheeses a little more cheese
a tall oak a table made of oak
various matters were discussed the relation of matter and space
various noises a great deal of noise
some sorrows/joys some sorrow/joy
all these dangers are past there is little or no danger
many experiences much experience
24
Moreover, the addition of classifiers, which individuate a certain
portion of the intended stuff/substance, contributes to the recategorization
of mass terms into general terms. Here are a few examples:
(5) a lump/piece of sugar; a loaf/piece of bread; a sheet/piece of

paper; a bar/cake of soap; a strip of land; a grain of rice; a
pile/heap of rubbish; a piece of evidence/information; a word
of advice; a piece of luggage; an article/piece of furniture; a
piece/item of news; a fall of snow; a reel of thread; a cup of
milk; a flutter of excitement; a pang of jealousy; a stroke of
luck; an act of kindness
It is clear that the classifier (i.e., the partitioning expression) can

occur in the plural. There are cases when the classifier is elided and the
mass noun appears in the plural or preceded by the indefinite article a:
(6) a. I had two coffees this morning.
b. I had a beer for lunch.
In (6) two coffees and a beer are elliptical for ‘two cups of coffee’
and ‘a glass/a bottle of beer’, respectively.
On the other hand, the use of determiners such as the, this, that with
mass terms turns them into singular terms, uniquely identified. The same
individuating effect is obtained: a certain ‘portion’ of the stuff/substance
is intended.
(7) a. Butter is healthy.
a’. This/the butter is stale.
b. Gold is a precious metal.
b’. That/the gold was found in the other room.
There are other means of turning a mass term into a count term that
do not involve classifiers. These other means are pluralization and the use
of the indefinite article a; they trigger semantic differences between the
interpretation of the mass term and the newly formed count term (see
section 5).
25
3.0. The Plural Morpheme
As a rule, countable nouns form the plural by adding the
inflectional morpheme #s whose pronunciation varies function of the
final consonant/vowel of the word it attaches to. #s is pronounced /s/ after
voiceless consonants (e.g., books, pots, beliefs), it is pronounced /z/ after
voiced consonants (e.g., pens, mirrors, dogs) and after vowels (e.g., days,
rays) and it is pronounced /iz/ after hissing sibilant consonants
(e.g., buses, pieces, sizes).
Many words entered into English from Latin or Greek and retained
their plural forms from the respective language. For example, alumnus –
alumni, focus – foci, cactus – cacti, radius – radii, curriculum –
curricula, millenium – millennia, symposium – symposia are words of
Latin origin. Some examples of Greek words are analysis – analyses,
thesis –theses, axis – axes, diagnosis – diagnoses, criterion –criteria,
phenomenon – phenomena.
Several English words do not make any formal distinction between
the singular and the plural sheep – sheep, fish – fish, quail – quail, shrimp
– shrimp while others still retain the Old English plural form child –
children, ox – oxen.
All the plural affixes exemplified above are allomorphs of the same
morpheme (which stands for the feature [+plural]).
Other problems raised by plural nouns, either morphologically
marked or not, are discussed in the subsections below.
3.1. Collective Nouns. Distributive and Collective Plurals

Collective nouns such as class, audience, government can be
defined as nouns designating a whole class of individuals. They consist of
individualizable elements but their morphological form is singular.
The list of collective nouns below reflects the most significant
socio-cultural groupings of society as found in the area of politics, trade
and industry, religion, sports, etc (cf. Ştefănescu 1988, Baciu 2004):
(a) Politics: assembly, air-force, cabinet, House of Commons,

senate, government, party, opposition, Foreign Office,
minority, majority, ministry, mass, council, congress, press,
26
jury, committee, public, people, police, proletariat, army,
troop, fleet, society, squadron, etc.
(b) Trade / Industry: firm, staff, board, sales division,
department, management, union, club, team, etc
(c) Religion: congregation, clergy, parish, choir, ministry, etc.
(d) Education / Sports: class, crew, group, school, audience, etc.
(e) Others: family, proportion, crowd, mob, company,
aristocracy, gang, data, nobility, media, household, flock,
herd, poultry, mess, swarm, pack, flight, livestock, money, etc.
To this list we can add the names of many organizations which also
display the behavior of inherently collective nouns: the NATO, the BBC,
the EU, etc. though these nouns are usually interpreted as singular terms,
since they refer to one unique body.
3.1.1. Properties of Collective Nouns

There are two important properties that characterize collective nouns:
(i) morphologically, they are singular nouns
(ii) from a semantic point of view these nouns designate sets of
individuals
In English (but not in other languages such as Romanian) collective
nouns in the singular evince a bizarre property: they can be used to denote
either the individual members of the set or the whole set as a body or
group.
When collective nouns in the singular designate the individual
members of the set they acquire a distributive interpretation and agree
with the verb in the plural (as in (8b)). When collective nouns in the
singular designate the whole set as a body or group they acquire a
collective interpretation and agree with the verb in the singular (as in
(8a)).
(8) a. This Government is trying to control inflation.
b. This Government are trying to control inflation.
Consider some more examples that illustrate the distributive and
collective readings of collective nouns (Poutsma 1926):
27
(9) a. The audience, which was a large one, was in its place by
7 p.m. (+ coll, -dis)
b. The audience, who were all waving their arms above their
heads, were clearly enjoying themselves. (-coll,+dis)
c. The board has issued its new rules for the equipment of vessels
at sea. (+ coll, -dis)
3.1.2. Agreement Patterns

The examples in (8) and (9) illustrate one match agreement pattern
and also two-mismatch agreement patterns characteristic of singular
collective nouns (Kim 2004).
The examples in (8) show that the subject noun always
matches/agrees in number with its determiner.
However, the examples in (8) and (9) also show that the subject is
in the singular but triggers either singular or plural agreement on the verb,
depending on the subject’s collective or distributive meaning. Moreover,
notice that the number value of co-referential pronouns in (9 a,b,c) also
varies function of the collective noun’s interpretation as collective or
distributive. As a matter of fact, the number value of co-referential
pronouns is the same as the number value of the verb (i.e., was / its place
in (9a), were / their arms above their heads in (9b) and has / its news in
(9c)).
These disparate agreement patterns in sentences with singular
collective nouns can be explained if we hypothesize that English
determiner-noun agreement is morpho-syntactic in nature while the
subject-verb agreement and pronoun-antecedent noun agreement are
semantic in nature (Wechsler and Zlatiċ 2000, Kim 2004) as illustrated in
the schema below:
(10) Morph-syntactic agreement
Det noun verb…
Semantic agreement
28
3.1.3. Collective Nouns and Predicates
The collective or distributive readings of singular collective nouns
generally depend on the predicate of the sentence.
Intransitive predications like be dispersed, be numerous, gather,
outnumber, collide, be alike, be a trio, be a couple, force a collective
reading. These predicates have a collective reading and obligatorily
require semantically non-singular subjects (Hausser 1974). Predicates like
be admired or be pleased force a distributive reading of collective noun
subjects.
3.1.4. Collective Nouns in the Plural

Collective nouns can also appear in the plural. Collective nouns in
the plural also allow for a double semantic interpretation: either a
collective or a distributive reading. In these cases as well, the predicate is
the one that forces one of the two possible readings. Consider the
examples below:
(11) a. Many armies are admired for their courage all over
Europe.⇒ Each army / navy is admired for its courage.
(+pl, – coll)
a’. The four armies gathered their men to meet the enemy.
(+pl + coll)
b. Many gangs will be shaken by this law ⇒ Each gang will
be shaken by this law. (+pl, – coll)
b’. Many gangs gathered in the same room. (+pl, + coll)
The examples in (11a, 11b) receive a distributive reading. This

interpretation is required by the predicate be admired. The predicate be
admired applies to each army (i.e. the predicate is true of each member of
the set); in (11a’, 11b’) we have to do with a collective interpretation
(i.e. the predicate is true of the sets of individuals). This interpretation is
imposed by the predicate of the sentence, namely gather.
As a matter of fact, all plural nouns, not only collective nouns
display a collective or distributive reading.
There are cases when the predicate may render a plural noun
ambiguous between the two possible readings:
29
(12) a. All the cars have a petrol gauge ⇒ Each car has a petrol
gauge. (– coll)
b. The cars collided / gathered in the parking lot. (+ coll)
c. All the men lifted the piano (Cornilescu, 1986: 303).
In (12a) the plural noun in subject position has a distributive
reading since the predicate is true of each member of the set of cars while
the subject of (12b) has a collective reading since the predicate is true
only of the set. In (12c) the subject can be given either interpretation; the
sentence may mean either: (i) each man lifted the piano in turn
(distributive reading) or (ii) all the men lifted the piano together
(collective reading).
4.0. Morphologically Defective Nouns

Other classes of nouns are plural in meaning but are morpho-
logically defective. They do not evince the singular – plural contrast.
4.1. Nouns like deer, sheep and swine have their plural form
identical with their singular form. Verb agreement is either in the singular
or in the plural. In fact, they are countable nouns and their morphological
irregular behavior is explained in terms of their historic, diachronic
evolution.
They take all the articles and quantifiers (plus cardinals) that
characterize bona-fide countable nouns:
(13) a. There are several deer/sheep grazing peacefully in the distance.
b. His reindeer are from Lapland.
c. The domestic swine fairly dotes on snakes.
4.2. Nouns that designate wild animals, wild fowl and fish have the
unmarked (singular) form used for both singular and plural contexts.
They are countable nouns and have count properties, except for the lack
of plural marker on the noun, which is again explained in terms of the
diachronic evolution of English. They co-occur with cardinals and plural
anaphoric pronouns.
In terms of verbal agreement, the verb is always in the plural:
30
(14) a. Fresh-water fish are more valuable for the sport they
provide than for the market.
b. Are these duck or mergansers?
c. Did you get many salmon after I left?
d. Between four and five moose are annually eaten at the forts.
These nouns are also called ‘collective singular’ nouns (Poutsma
1926). Some of these nouns such as fish, trout, carp take the plural
marker -s when reference is made to varieties of fish.
The use of the singular form of the noun is the general tendency but
there are also exceptions to the rule:
(15) a. You may kill a few antelope.
b. The true antelopes you saw are remarkable for the graceful
symmetry of their bodies.
The collective use of the singular form of these nouns is found
particularly with the shooting jargon. Compare:
(16) to shoot duck vs, to raise ducks
to shoot waterfowl vs, to keep fowls
4.3. Other countable nouns form their plural by Ablaut (or vowel
change): foot – feet; goose – geese; tooth – teeth; louse – lice; mouse –
mice; man – men; woman – women. Again, Ablaut preservation in plural
formation is explained on historical grounds.
Compounds of ‘man’ change to ‘men’ as in fireman – firemen,
postman – postmen. Similarly, compounds of ‘woman’ form the plural by
using ‘women’ as in house-woman – house-women, chairwoman –
chairwomen. Anglicized foreign derivatives of –man such as German,
Norman or Roman form the plural according to the general rule by
adding the suffix -s: Germans, Romans (Poutsma 1926).
Other survivals from Old English are a few nouns that form the
plural in -en: child – children, ox – oxen, brother – brethren. The plural
forms brethren (confraţi, from the singular brother) is nowadays used in
religious contexts:
(17) The persons least surprised at the Reverend’s deficiencies were
his clerical brethren.
31
4.4. Nouns such as cattle, livestock, poultry, people, folk, vermin are
morphologically not marked for plural (but are understood as designating
sets of individuals) and agree with the verb only in the plural. These
nouns lack the singular – plural contrast:
(18) a. These cattle belong to John.
b. *This cattle belongs to John.
Another peculiarity of these nouns is that they cannot occur with low
numerals but only with high numerals (as in (16a,b)). Moreover, distinct
lexical items must be used when individuation takes place (as in (19a)):
(19) a. *four cattle vs. four cows
b. two hundred cattle, poultry
Instead, the nouns folk and people can be used with low numerals:
(20) These seven people/city folk.
When the noun people means popor it displays regular singular –
plural forms:
(21) a. The people of Romania.
b. The peoples of Europe.
5.1. Number Recategorization of Mass Nouns into Countable Nouns

We turn now to the description of the classes of uncountable / mass
nouns that undergo number recategorization, usually with a change in
meaning. In this case pluralization and the use of the indefinite article a
trigger number recategorization of mass nouns into countable nouns.
5.1.1. Some mass nouns can be recategorized as count/general
terms when they occur in the plural but they mean ‘kinds of x’: wine –
wines, tea – teas, gas – gases, steel –steels, fruit – fruits, coffee – coffees,
fashion – fashions, etc. Consider the following examples:
(22) a. Four wines were served at dinner. They were dry wines.
b. Many different wines were served at dinner.
c. In the Customs list, all fruit is divided into three parts:
dried fruits, green fruits and nuts.
d. There was some gas left in the bag. Air is a mixture of gases.
32
The newly formed terms in (22) are count terms since they have
plural form, the verb agreement is in the plural, the quantifier is a count
quantifier (22b) and the anaphoric pronoun is plural in form (22a).
Proper names like Murfatlar, Bordeaux, Malaga, etc., represent
lexicalizations of ‘kinds’ of wine. Syntactically and semantically these
‘names’ function as mass terms:
(23) He drank too much Bordeaux; it went straight to his head.
5.1.2. Another class of mass terms that can be recategorized into
count terms contains mass terms, which through pluralization denote an
act / an instance / an occasion of the mass term. The act / instance /
occasion count terms have the quality of the mass term. Here are some
examples: attention – attentions, confidence – confidences, regard –
regards, curiosity – curiosities, novelty – novelties, tin – tins, paper –
papers, rubber – rubbers, silk – silks, implication – implications,
mentality – mentalities, respect – respects:
(24) a. He shouted in order to attract attention. They showed the old
lady numerous attentions.
b. She showed much confidence in life. The two girls were
exchanging many confidences.
c. He has lost all feeling in his leg. Are your feelings the same
for me?
As shown by the examples in (24) the recategorized plural nouns
evince count properties as they take count quantifiers and the agreement
with the verb is done in the plural.
5.1.3. There is another class of mass nouns that have a
corresponding plural form but this time the newly formed noun is also a
mass term.
What is peculiar of these newly formed mass nouns is that the sense
of the singular mass term includes the sense of the mass plural term. Here
are some examples: water – waters, snow – snows, sand – sands, wit –
wits, salt – salts.
These plural mass terms differ from their corresponding singular
term in that they trigger plural agreement with the verb (due to the plural
form of the noun), plural anaphoric pronouns but mass quantifiers:
33
(25) a. Water is a liquid. The waters of the Nile are essential for the
country’s agriculture. They make life possible. Much of
them make life possible.
b. There is plenty of snow in the Alps. The snows round the
Aiguilles are the least trodden.
6.0. General Terms with No Corresponding Mass Lexicalization

It is a well-known peculiarity of English that those count/general
terms which name animals have mass noun correspondents that designate
their meat:
(26) pig – pork; sheep – mutton; deer – venison; calf – veal; cow –
beef
However, the overwhelming majority of general terms have no
corresponding mass lexicalization, i.e., the same noun is used in both
cases:
(27) a. I saw many chickens in the garden. We had cold chicken for
dinner.
b. How many apples do you want? Put some more apple in
the salad.
c. There are two oaks in the garden. The table is made of oak.
d. The baby has two teeth. There is too much tooth about her.
7.0. Plurale Tantum Nouns

7.1.1. The term ‘plurale tantum’ is Latin in origin and is the
singular of ‘pluralia tantum’, which is roughly translated as plural-only.
The nouns that belong to this class have one single form, the plural one
but their morpho-syntactic behaviour is that of mass nouns. According
partly to their meaning and partly to their origin, pluralia tantum nouns
have been divided into:
7.1.2. Names of certain physical or mental illnesses such as: creeps,

hysterics, jerks, measles, mumps, shakes, shivers, megrims, tantrums,
sulks, shingles etc. These nouns have the behavior of mass nouns, except
for the plural marker:
34
(28) a. Mumps is a disease. It is infectious.
b. Shingles has severe complications.
c. The whole of her first year was one continual series of
sulks, quarrels and revolts.
This subclass of pluralia tantum nouns, which are mass nouns,
triggers singular agreement with the verb and singular anaphoric
pronouns. However, their individuation is possible by specialized lexical
expressions (i.e., classifiers): a series of (sulks), a fit of (hysterics).
7.1.3. Nouns that name sciences such as: aesthetics, acoustics,

athletics, dialectics, economics, ethics, linguistics, mathematics,
mechanics, optics, physics, phonetics, politics, statistics, tactics. These
nouns are basically mass nouns.
(29) a. Economics has come out into the open.
b. Ethics is the science of the laws that govern our actions as
moral agents.
c. Mathematics is the science of quantities; its students are
mathematicians.
d. Politics, as a profession, was of importance to him.
This subclass of pluralia tantum nouns triggers singular agreement
with the verb and singular anaphoric pronouns.
However, there are cases when the verb is in the plural; in this
instance, the mass noun is contextually changed into a count/general term
that designates ‘kinds of N’:
(30) a. Do mathematics make one’s manners masculine? Well, they
haven’t done so as yet in your case. But still they are not
womanly pursuits.
b. Statistics show that moderate consumers of alcoholic drinks
live considerably longer than drunkards and total abstainers.
c. It was those infernal physics that I have always neglected.
d. What is it? Nothing about politics, I hope. They don’t
interest me.
7.1.4. Nouns that name certain games such as: all fours, billiards,
cards, checkers, marbles, draughts, ninepins. They have mass noun
behavior except for their plural marker:
35
(31) a. John has given me a lesson in all fours.
b. Billiards is played in England.
c. Ninepins requires great skill.
This subclass of pluralia tantum nouns triggers singular agreement
with the verb and singular anaphoric pronouns.
7.1.5. Nouns that name certain articles of dress (which consist of
two symmetrical parts): breeches, knee-shorts, pantaloons, pajamas,
overalls, shorts, tights, trousers, braces, jeans, pants, suspenders. These
nouns evince count properties, they divide their reference; hence, they are
count/general terms and their individuation is done with the help of
classifiers (e.g., a pair of).
(32) a. Marley in his pigtail, usual waistcoat, tights and boots turned
up late.
b. He was dressed in a tarnished green traveling jacket and a
pair of overalls.
7.1.6. Nouns that name instruments formed of two parts such as:
scales, eyeglasses, spectacles, scissors, forceps, pliers, tongs, compasses,
binoculars, bellows, chains. All these nouns have count behavior: they
trigger plural agreement with the verb and in order to individualize the
lexical expression (i.e., classifier) a pair of is used:
(33) a. She took out a pair of scissors.
b. He had a large pair of bellows with long slender nozzle
of ivory.
c. Mr. Brooke stuck his eyeglasses on nervously.
7.1.7. Finally, nouns that name parts of the body (made up of two
parts) such as: bowels, entrails, guts, lungs, loins, whiskers. These nouns
also display countable properties, qualifying as general terms:
(34) a. Scrooge had often heard it said that Marley had no bowels
but he had never believed it until now.
b. The lungs, or as they are vulgarly called lights, are eaten as
parts of the pluck or fry.
When reference is made to one of the two parts that make up the
respective body part some of these nouns also have a singular form: lung –
lungs, whisker – whiskers, tit – tits, buttock – buttocks, eyelash – eyelashes.
36
3. THE CATEGORY OF DETERMINATION
1.0. Classes of Determiners

The category of determination is another functional category of the
lexical category noun.
The following groups of items are generally included in the category
of determiners (Jackendoff, 1977, Cornilescu 1995, 2006, Baciu 2004):
• articles: the definite article the (e.g., the man), the indefinite
article a / an (e.g., a man, an umbrella) and the negative indefinite article
no (e.g., no man).
• demonstrative articles: this, these, that, those book/s.
• article-like quantifiers (i.e., quantifying elements that have the
syntactic position of articles): e.g., every, each, all, some, another, any,
what, which book/s.
• cardinal numerals, ordinal numerals and lexical quantifiers:
e.g., two, three, the second, the third, many, a little, little, a few, few,
several (book/s), much (wine).
The category of determiners meets the two criteria for qualifying as
a functional category: (i) this category includes a limited number of
members and (ii) these members do not have descriptive meaning of their
own. Determiners are words like this in this new car, which modify the
noun they occur with just as the adjective new does. Determiners have
abstract meaning and are extremely idiosyncratic. They have specific
distributional and interpretative properties:
(1) a. The boys will all/each/both/*some/*every get a prize.

b. All/both/some/*each/*every boys sat down.
c. *All/*both/each/every/some boy sat down.
37
The determiner is the most important morpho-syntactic element in
the nominal group. For example, the determiner can dictate number
agreement with the noun and the verb (e.g., This boy is tall / Those boys
are tall, All students in my class have got a prize / Every student in my
class has got a prize). In a language such as Romanian their role is even
more conspicuous: besides determining agreement with the noun and the
verb, determiners are also carriers of case and gender features:
(2) Nom / Acc: un băiat (m) un scaun (n) o fată (f)
Gen / Dat: unui băiat unui scaun unei fete
Semantically, determiners play a crucial role: they close off the
lexical category noun and make it possible that nouns (which designate
classes of objects, e.g., house, dog) have a fixed referent (the house, the
dog). In this way, nouns can function in discourse (e.g., The house
belongs to John / *House belongs to John; The dog is in the backyard /
*Dog is in the backyard).
Our presentation focuses mainly on the syntactic, semantic and
pragmatic description of the English definite and indefinite articles.
In spite of the fact that these articles belong to the same paradigm
there are important differences between them (see the examples in (1)).
Moreover, historically, the two articles have distinct sources: while the
developed from the demonstrative this, a developed from the numeral one
(Chesterman 1992, Baciu 2004).
As we have already seen (see the Category of Number) the definite
article the is neutral with respect to the opposition countable-uncountable
or singular-plural, thus being able to occur with any kind of noun (except
for proper names and the abstract nouns nature and mankind) whereas the
indefinite article a(n) occurs only with countable singular terms.
2.0. Philosophical Background

2.1. Modern research on the English class of determiners is deeply
grounded on the results philosophers achieved in their studies on the
relation between (linguistic) definiteness and (extra-linguistic) reference.
Natural languages contain expressions with the help of which
people refer or point out to individuals, things, times or places. The act of
referring is accomplished by using expressions such as (i) proper names
38
like Fred, Mary, Elizabeth, John Brown, Hyde Park, Hampton Court, etc;
(ii) definite descriptions (i.e., expressions made up of a definite
determiner and a nominal) such as: the clever man, the house, the book on
the table, this house, my dog, etc; (iii) indefinite descriptions
(i.e., expressions that consist of an indefinite determiner like a/an or some and
a nominal: a clever man, a house, a book, a house, some boys; (iv) pronouns
like I, you, he, she, etc; (v) adverbs like here, there, yesterday, etc.
The problem is whether these referring expressions unambiguously
refer to an identifiable thing or individual in all circumstances. (The
following discussion on the philosophical tradition of definite
descriptions draws on Cornilescu p.c.).
2.2. Roughly speaking, the great difference between the and a is
that definite descriptions (e.g., the student, the window) unambiguously
identify the object or individual they refer to while indefinite descriptions
(e.g., a student, a window) do not unambiguously identify the object or
individual they refer to:
(3) a. The student called this morning.
b. A student called this morning.
In his seminal article “On Denoting” (1905) Russell tried to identify
the exact ‘meaning’ of definiteness.
On the one hand, human beings can talk meaningfully about things
that do not have a referent:
(4) He is seeking the Golden Mountain / the round square / the
Fountain of Youth.
In view of such examples Russell concludes that one meaning
dimension of definite descriptions is uniqueness of the referent. (These
definite descriptions acquire discursive referents, not extra-linguistic
referents. As we are going to see below, definite descriptions can be
resumed by anaphoric definite pronouns such as it: e.g., John is looking
for the pen but he hasn’t found it. Just like any definite description that
has a referent in the world, the definite descriptions in (4) can be resumed
by it: John is seeking the Golden Mountain but he hasn’t found it).
On the other hand, human beings are also able to meaningfully talk
about entities (that have referents) in terms of their non-existence:
39
(5) a. The King of France doesn’t exist.
b. The blackest man in the world doesn’t exist.
In view of these examples, Russell’s conclusion is that the other
meaning dimension of definiteness is the existence of the referent (which
can be negated).
Thus, the in the definite description the King of France is used to
abbreviate two statements: the uniqueness statement and the existential
statement:
(6) a. There is a King of France. (existential statement)
b. There is only one King of France. (uniqueness statement)
In Russell’s view, a sentence such as:
(7) The King of France is bald.
is analyzed in terms of three assertions related by the conjunction and:
(8) a. There is a King of France. (and)
b. There is only one King of France. (and)
c. The King of France is bald.
When we negate a conjunction it is enough to negate one conjunct
(a, b or c in (8)).
In a negative sentence like The King of France doesn’t exist, not
negates the King of France’s existence and implicitly his uniqueness. In a
sentence like The King of France is not bald, Russell argues, it is enough
for not to negate one assertion in (8), namely (8c), for the whole sentence
to be false. In other words, in Russell’s view, in The King of France is not
bald if one conjunct is negated (i.e., (8c)) then the properties of existence
and uniqueness of the King of France are also negated.
Strawson’s account of definiteness in “On Referring” (1950) raises
against the view that The King of France is not bald means that there is
no King of France. According to Strawson, an assertion (e.g., The King of
France is bald) presupposes an act of reference, an act of identifying the
entity about which we assert something. If there is no referent, we cannot
use the sentence communicatively and we cannot say whether the
sentence is true or false. Indeed, the statement in (7) may well have been
true in AD 1670 and false in 1770 but in 1950 the statement cannot
40
sensibly be said to be either true or false. Due to the non-existence of a
King of France in 1950 the question of the truth or falsity of sentence (7)
does not even arise. (Strawson does not discuss the truth conditions of
sentences such as The King of France does not exist).
Strawson’s analysis of definite reference is pragmatic in nature.
Searle (1969) offers an analysis of definite descriptions close to
Strawson’s: he postulates two axioms that must hold for a speech act of
definite reference to be successful:
Axiom of existence: there must exist one and only one object to
which the speaker’s utterance of the expression applies.
Axiom of identification: the hearer must be given sufficient means
to identify the object from the speaker’s utterance of the expression
(Searle 1969:82).
Although Russell’s analysis of sentences that contain definite
descriptions and not is incorrect his interpretation of the definite article
the as meaning both existence and uniqueness of the referent has proved
extremely valuable and has been retained by modern analyses of definite
descriptions.
Of late, philosophers have also become aware that definite
description expressions can be used not only on a referential
interpretation (i.e., the definite description has one and only one known
referent). In the example below:
(9) I’m grateful to the inventor of the computer.
the definite description the inventor (of the computer) is paraphrased as
meaning (I’m grateful to) ‘whoever invented the computer’. In this case,
the definite article the does not have a referential interpretation but an
attributive interpretation (Donnellan 1966). It refers to the unique but
unknown individual that has the attribute of having invented the computer.
Furthermore, for long time philosophers believed that definite
descriptions are used to refer to objects/individuals in the same way as
proper names do (Frege, Russell). Definite descriptions and proper names
fell in the class of definite descriptors.
However, Kripke (1973) argued that proper names different from
definite descriptions, always refer to the same individual in all contexts of
use: their reference remains constant and is ‘rigid’ (i.e., they are ‘rigid
41
designators’). This assumption can be tested in counter-factual statements
of the form:
(10) If John Smith were taller he could touch the ceiling.
In the counter-factual circumstance (world) in (10) the individual
designated by the proper name John Smith is the same as the individual
designated by this name in the actual world. It follows that the reference
of proper names is constant and does not change function of
circumstances (worlds).
In contrast, definite descriptions are non-rigid designators and their
reference can vary in time and space. For instance, the referent of the
definite description the man of the year refers to different individuals,
function of the year: the man of the year (2003), the man of the year
(2004) or the man of the year (2005). The same goes for superlatives,
which can vary in space: the most beautiful girl (in Romania), the most
beautiful girl (in France).
In all these examples we have to do with the attributive use of
definite descriptions identified by Donnellan (1966). Kripke’s analysis of
definite descriptions is also Russellian in spirit as definite descriptions
convey uniqueness and existence of the referent, only the referent may
change according to circumstances (worlds). Of course, the man of the year
(2004) is used referentially when it has a unique referent that also exists and
is known; the same goes for the most beautiful girl (in Romania).
The insights of these analyses will prove crucial in the
characterization of the linguistic values/uses of the definite article.
2.2.1. In present-day studies in linguistics, definiteness is defined in
relation with the notion of ‘familiarity’ of the referent (Heim 1982) in the
‘resource situation’ on the side of both the speaker and of the hearer. In a
given communicative situation the set of objects in relation to which the
reference of an expression is established is called ‘resource situation’
(Barwise and Perry 1983) or ‘shared set’ (Hawkins 1978). In other words,
the conventionally implies that there is a set of entities in the universe of
discourse which is mutually manifest to speaker and hearer
(i.e., familiarity) and within which the definite referent exists and is
unique (cf. Baciu 2004).
42
2.2.2. In contrast with definite descriptions (which are associated
with familiarity of the referent to both speaker and hearer) indefinite
descriptions (e.g., a student, a chair) are associated with the pragmatic idea
of ‘novelty’ (Heim 1982) on the part of the hearer. In a sentence such as:
(11) A student called in the morning.
a new referent (a student) is introduced in the universe of discourse (i.e.,
the resource situation) a referent which may be known to the speaker but
is unknown to the hearer and cannot be uniquely identified by the hearer.
However, in negative sentences the indefinite description may refer
to no object (i.e., has no referent):
(12) John didn’t read a book the whole summer.
The paraphrase of the sentence in (12) is not that there is a specific
book that John didn’t read the whole summer but that the set of books
John read is zero.
Moreover, there are contexts where the indefinite a doesn’t mean
one (Heim 1982):
(13) If a farmer has a donkey he beats it.
In (13) we interpret a farmer and a donkey as forming pairs (a
farmer-a donkey) but a is not interpreted as meaning ‘if one farmer has one
donkey he beats it’ but as ‘if all farmers have donkeys they beat them’.
3.0. Characteristics of Indexicals

As can be noticed, determiners, exemplified above with
definite/indefinite determiners may acquire a different interpretation
function of the context in which they are used. In this respect they are
very much like indexicals, which are items such as I, this, that, yesterday,
now, then, etc. (Kaplan 1973). Indexicals, also called deictic elements, are
expressions that have a constant linguistic meaning but acquire different
interpretations each time they are used by different people at different
times or locations (Kaplan 1973 in Baciu 2004). For example, a sentence
like I was insulted yesterday has, on the one hand a constant meaning
independent of the context in which it is used: the use of I is to point to
the speaker of the utterance, and yesterday constantly points to the day
43
before speech time. On the other hand, ‘what is said’ by the utterance
depends on who the speaker is and when the utterance is said, i.e., I will
point to the ever-changing speaker who says the utterance, yesterday
acquires different interpretations depending on the day the utterance is
made. In other words, ‘what is said’ depends on both the meaning of the
expression used and also on the context (place, time, different people) in
which the respective expression is used.
Thus, linguistic expressions (e.g., deictic elements, the class of
determiners) have constant meaning but their interpretations may vary
function of the context of use.
The components of the context of use are (Barwise and Perry 1983):
• indexicality or deixis
• resource situation
We briefly discuss them below.
3.1. Indexicals or Deictic Elements

The term deixis is of Greek origin and can be paraphrased as
‘pointing to’ or ‘indicating’ something. All languages grammaticalize (or
encode) information with respect to the persons that take part in the
conversation (encoded in English as I and you), with respect to the time
when the conversation takes place (grammaticalized by adverbs of time
such as now, yesterday, last year, tomorrow, then and tense markers) and
with respect to the place where the conversation takes place
(grammaticalized by demonstrative pronouns like this, that and adverbs
of place like here, there).
The traditional deictic categories are person deixis, time deixis and
place deixis. To these deictic categories discourse deixis has been added
(Levinson 1983). Discourse deixis deals with the relations between the
deictic elements of an utterance that is located in a written or spoken text.
As Levinson (1983:54) remarks “communication has an egocentric
organization which is encoded or grammaticalized by deictic elements”:
the central person is the speaker, the central time is the time at which the
speaker produces an utterance and the central place is the speaker’s
location at utterance time.
44
3.2. Resource Situation
Deictic elements are expressions that identify (i.e., establish the
reference of) an individual, place and time in a given communicative
situation. The set of objects within which the reference of an expression is
established is called ‘resource situation’ (Barwise and Perry 1983). The
resource situation is available to the speaker and/or the hearer in various
fashions:
• by being directly perceived by the speaker and/or the hearer
• by being the object of common knowledge about the world
• by being built up by previous or subsequent discourse
4.0. The Values/Uses of Demonstrative Descriptions

Depending on the type of resource situation available,
demonstrative descriptions have different values or uses.
4.1. The Deictic/Gestural Value

When the speaker has direct access to the resource situation and the
object referred to is close (+proximal) to him the demonstrative articles
this/these are used. When the object referred to is at a certain distance
from the speaker (+distal) the demonstrative articles that/those are used.
These objects may also be indicated with a gesture.
(14) a. This/that dog cannot walk.
b. This book is the one you should read.
c. That building is the British Embassy.
4.2. The Symbolic Value

If the speaker uses a demonstrative article that exploits some
common knowledge shared by him and the hearer about the world they
live in, the function of the demonstrative article is symbolic.
(15) This city is really beautiful when it is not dirty.
The reference of this city is identified on the basis of common
knowledge of both the speaker and the hearer (it may be Bucharest if both
45
of them are in Bucharest, it may be Amsterdam if both of them are in
Amsterdam, etc.).
The gestural and symbolic values of demonstratives are subsumed
under the general term ‘deictic’ value (Fillmore 1971, Levinson 1983).
If the demonstrative expression exploits a resource situation built
up on previous or subsequent discourse, the demonstrative expressions
display discourse/textual functions, which are non-deictic functions.
4.3. Non-deictic Functions of Demonstratives

The non-deictic or discourse/textual functions of demonstratives are
the anaphoric and the cataphoric functions.
These two functions are syntactic in nature because the objects or
individuals referred to by demonstratives are located at the level of
discourse: in this case, demonstratives serve as expressions that connect
the speaker/reader to the discourse, which becomes the resource situation.
4.3.1. The Anaphoric Value

The term anaphora is a Greek loan and means ‘backward looking’
or ‘pointing backwards’, i.e., to previous discourse. In this case,
demonstratives function as ‘anaphors’ since they point backwards to an
individual/object already introduced in the previous discourse, known as
‘antecedent’. The syntactic requirement for anaphoric processes is the
relation of co-reference: the antecedent and the anaphor refer to the same
individual/object.
(16) a. Bush made his long-awaited announcement yesterday.
(antecedent)
This statement confirmed the speculations of many observers.
(anaphor)
b. What looked like a white lace poncho covered him from

head to foot. Beneath this he was wearing a shirt.
c. To be or not to be: that is the question.
46
4.3.2. The Cataphoric Value
The term cataphora is also a Greek loan but it may be paraphrased
as ‘pointing forward’, i.e., to subsequent discourse. Cataphoric uses are
also syntactic in nature as they involve the presence of a specification in
the subsequent discourse that ensures the possibility of pointing to the
intended element.
(17) a. What I want to say is this: drive carefully!
b. What do you think of this idea: let’s take them all fishing.
In the examples above the demonstrative pronoun and the
demonstrative article point forward to subsequent information.
4.3.3. The Emotional Value

Another textual value of the demonstratives is their emotional
value. The use of a demonstrative of proximity shows interest or
emotional involvement of the speaker in the identification of the referent:
(18) Once upon a time there was a king. This king had three daughters.
The use of a demonstrative of distance shows lack of interest on the
part of the speaker in the identification of the referent:
(19) And how’s that leg doing?
5.0. Definite Descriptions

5.1. General Remarks
The meaning of definite descriptions such as the girl, the dog, the
butter in the fridge is always associated with the notion of familiarity of
the referent on both the speaker and the hearer side in a communicative
situation. The reference of a definite description also implies existence
and uniqueness of the referent.
Definite descriptions, alongside demonstrative descriptions,
personal pronouns and proper names function as singular terms.
As already mentioned, the definite article is neutral to the
opposition singular-plural or countable-uncountable. It can attach to any
kind of nouns except for proper names. This restriction of occurrence is
47
semantic in nature: since both proper names and the have the role of
uniquely identifying an individual, the presence of the on a proper name
would lead to redundancy, i.e., a double presence of the semantic feature
of uniqueness.
Definite descriptions, just like proper names, have ‘interpretive
independence’ in the sense that there is no ambiguity as to the referent
(the identified individual/object) designated by the respective expression.
Other items, which also belong to the category of determiners (e.g.,
every / each / all / no) do not have interpretive independence and they do
not uniquely identify the individual/object referred to:
(20) a. Everyone came in. *He sat down.
b. Every student has read a book by Chomsky.
b’. Every student has read ‘Syntactic Structures’.
b’’. Every student read the assigned book.
In (20a) everyone does not introduce a new referent in the discourse
and it cannot be resumed by the definite pronoun he, which shows co-
reference and referential identification. The sentence in (20b) is
ambiguous between two interpretations. The first interpretation/reading is
‘there is a particular book by Chomsky (i.e., Syntactic Structures) that
every student has read, as in (20b’). The second reading is ‘every student
has read some book by Chomsky, not necessarily the same’, as in (20b’’).
Elements such as every, each, all, a evince lack of interpretive
independence and are called quantifiers.
There is a further pragmatic restriction on the use of definite
descriptions. Hawkins (1978) shows that definite descriptions refer
‘inclusively’ to the totality of objects that satisfy the referring expression
in the pragmatically delimited set of the resource situation. Consider the
examples below:
(21) a. Bring the wickets in after the game of cricket is over!
b. I must ask you to move the sand from my gateway.
In (21a) the wickets refer to all the six wickets necessary for a game
of cricket. If the hearer brings in only five of them, the speaker is not
satisfied. Similarly, the expression the sand in (21b) refers to all the sand
that is in front of my gateway not just part of it.
48
As in the case of demonstrative expressions, the uniqueness of
reference of definite descriptions is ensured by the different fashions in
which the resource situation becomes available for exploitation:
• by direct perception
• by being the object of common knowledge about the world
• by being built up on previous or subsequent discourse
In what follows we briefly describe the values/uses of the definite
article the (and implicitly of definite descriptions).
5.2. The Values/Uses of Definite Descriptions

5.2.1. The Deictic / Gestural Use of Definite Descriptions
The deictic / gestural value of definite descriptions is based on both
the speaker and hearer’s perception of the resource situation. There are
several sub-cases that can be distinguished. Consider the following
examples:
(22) a. Close the door, please!
b. PC 48, catch the jailbird!
c. Don’t come into this house my friend or I’ll set the dog onto
you!
d. Don’t feed the deer!
In (22a) the definite description is used gesturally / deictically to
point to the unique door identifiable in the resource situation by the
speaker and the hearer. The example in (22b) illustrates the case in which
only the hearer has direct access to the resource situation. The utterance
could be spoken on the radio by a policeman who instructs his colleague
to chase an escaped prisoner. The speaker cannot see the intended unique
individual but can instruct the hearer to locate and catch him. The
example in (22c) illustrates the case when only the speaker has direct
access to the object designated by the definite description. For one reason
or another the hearer has no direct perception of the dog in question. The
example in (22d) illustrates the situation in which the individual
designated by the definite description may be invisible for the hearer.
(The sentence can be used as notice in a zoological garden). The speaker
49
instructs the hearer to uniquely identify an individual (i.e., the deer) that
exists among the individuals that make up the resource situation.
In all the examples above the can be replaced by the demonstrative
articles this with more or less the same meaning. However, the definite
article is more abstract than the demonstrative article since the hearer
need not actually see the designated object/individual. “The use of the
demonstrative requires that the hearer is actually able to see the object in
the immediate situation, while the definite article does not have a
visibility condition on its use” (Hawkins 1978:114).
5.2.2. The Symbolic/Deictic Use of Definite Descriptions

The symbolic/deictic use of definite descriptions is based on both
the speaker and hearer’s exploitation of the resource situation as province
of common knowledge of the world (or part of the world). Consider the
following examples:
(23) a. Can you give me a lift to the town hall?
b. The Prime Minister has just returned from a visit in the
countryside.
c. Where is the church?
d. Who are the bride-maids?
In all the above examples the use of the definite article is based on
the speaker and hearer’s mutual knowledge of the existence of certain
unique objects/individuals in certain resource situations (i.e., the unique
town hall in the town, the unique church in the village, the unique Prime
Minister in the country).
In the case of singular definite descriptions the definite article refers
inclusively to the totality of the objects in the pragmatically relevant set
which happens to be made up of one (unique) member. In the case of
plural definite descriptions (e.g., the bride-maids in (23d)) the definite
article still has unique reference, only this time the referent is made up of
more members (the speaker and hearer share the general knowledge that
weddings have bride-maids along the bride, the bridegroom, the best man).
In these examples the function of the definite description is to
instruct the hearer to locate the intended object/individual in the resource
situation.
50
5.2.3. Discourse Functions of Definite Descriptions: the
Anaphoric and Cataphoric Functions
5.2.3.1. General Characteristics
The discourse functions of definite descriptions are the same as
those of demonstrative descriptions. The linguistic discourse acts as
resource situation for both anaphoric and cataphoric functions of the
definite article.
Unlike the deictic functions of the definite article, which depend on
the possibility of relating the use of linguistic expression to the world, its
discourse functions are syntactic in nature. They are related to whether
reference by means of definite descriptions relies on previous discourse
(the anaphoric use) or subsequent discourse (the cataphoric use).
5.2.3.2. The Anaphoric Value/Use of the Definite Article

The anaphoric use of the definite article the is based on the relation
between an antecedent nominal and an anaphor. Consider the following
examples:
(24) a. Once upon a time a child was born in Bethlehem.
(antecedent)
The child was baptized Jesus.
(anaphor)
b. When she entered the office she saw a little man. The little
man was sitting in her armchair, scratching his nose.
In (24a,b) the indefinite descriptions a child, a little man function as
syntactic antecedents to the anaphors expressed by the corresponding
definite descriptions the child, the little man. Thus, the role of an
indefinite description is to introduce a new individual (referent) in the
domain of discourse, which is later on resumed by the use of a definite
description. The uniqueness of reference of the definite description is
ensured by the singleton meaning of the article a that precedes it.
Besides the pure anaphoric use of the definite article there are other,
more complex uses of ‘associative’ anaphora with several sub-species
(Christophersen 1939 in Baciu 2004).
The sub-species of ‘associative’ anaphora are based on:
51
(i) the semantic relation of partonymy (i.e., a part-whole relation):
(25) a. Ian inherited a house but unfortunately he could not make it
his house since the roof was leaking and the windows
needed repairing.
b. I’ve still got a book of nursery rhymes I had as a child but
unfortunately the cover is torn.
Associative anaphora of this type is based on our encyclopedic or
conceptual knowledge of the world we live in (whole objects have
constitutive parts: a house has a roof and windows, a book has a cover
and pages, etc.).
(ii) the semantic relation of hyponymy (an inclusion relation):
(26) Bill was working at a lathe the other day. All of a sudden the
machine stopped working.
(iii) the semantic relation of synonymy (a sense relation):
(27) Fred was wearing trousers. The pants had a big patch on them.
5.2.3.3. The Cataphoric Value/Use of the Definite Article

The cataphoric use of the definite article is also based on the
linguistic context that acts as resource situation. In this case, the definite
article signals the presence of a post-modifier. Consider the following
examples:
(28) a. The man who stands in the corner is my brother.
b. The milk you bought yesterday turned sour.
c. The comment of the publisher was completely at fault.
The role of the post-modifier relative clauses in (28a,b) and that of the
post-modifier of the publisher is to license the definite article which helps us
uniquely identify the objects in discourse (the man, the milk, the comment).
5.3. The Generic Function of Definite Descriptions

Definite descriptions may also have a generic function. (Definite
descriptions that occur in generic sentences are discussed in detail in
section 11.2. below).
52
(29) a. The dog is an intelligent animal.
b. The lion is a noble beast.
5.4. The Non-Referential (Attributive) Use of the Definite Article

As already mentioned above, definite descriptions do not have only
a referential use but also a non-referential (attributive) one (e.g., I’m
grateful to the inventor of the computer) (Donnellan 1966).
There are even contexts where a definite description is ambiguous
between a referential interpretation and a non-referential/attributive
interpretation. Consider the following example (cf. Baciu 2004):
(30) Mary believes that the man who lives upstairs is insane.
This sentence may mean either a) or b) below:
a). Mary believes that a certain individual, namely the man who
lives upstairs (say John Brown) is insane. (the referential
interpretation of the definite description)
b). Mary believes that whoever it is that lives upstairs is insane.
(the non-referential interpretation of the definite description)
The non-referential interpretations of definite descriptions are
created by ‘opaque’ linguistic contexts. In (30) the verb believe engenders
referential opacity.
6.0. Proper Names in the Form of Definite Descriptions

In principle, proper names cannot be used with definite articles
(*the London / *the John) because of the double presence of the semantic
feature of uniqueness: both proper names and the definite article have the
role of uniquely identifying an individual/object.
6.1. However, definite descriptions, which name geographical
areas, historical and cultural institutions, political, and administrative
divisions function as proper names preceded by the definite article. The
speaker uniquely relates them to his resource situation, based on his
cultural, historical, administrative, geographical or political conventions
(e.g., the British Museum, the Atlantic Ocean, the Empire State Building,
the House of Lords, the Roman Empire, the Holy Virgin, the Ritz, etc.).
53
This class of unique objects that function as proper names evinces
many idiosyncrasies. We mention below the following (Poutsma, 1926,
Ştefănescu, 1988):
---some of these definite descriptions have the nominal in the
plural: the Rocky Mountains, the Zoological Gardens, the Low Counties,
the Basque Provinces, etc.
---some of them can occur without the nominal (elliptical definite
descriptions): the Atlantic, the Baltic, the Thames, the Tate, the Bronx, the
Acropolis, the United States, etc.
---some others may keep or drop the definite article. The definite
article is mainly dropped before names of buildings, bridges and other
structures: Buckingham Palace, Westminster Abbey, Victoria Station,
St. Paul’s Cathedral, Lincoln’s Inn, Magdalene College, London Bridge, etc.
The definite article is also dropped before names of streets, squares,
parks or proper names containing nouns such as circus, cross, filed(s),
garden(s): Finsbury Circus, Oxford Street, Hyde Park, Charing Cross,
Soho Fields, Covent garden, Russell Square, etc. The definite article is
almost regularly used before names of hotels, museums and theatres: the
Clarendon Hotel, the Court Theatre, the Globe Theatre, the South
Kensington Museum.
6.2. There are instances when definite/indefinite articles and
cardinals can precede proper names. In this case, they behave like
countable/sortal nouns and they are interpreted as designating two or
more individuals or as collective nouns (with family names in the plural):
(31) a. I met two Maries at the party last night.
b. I am referring to the Napoleon who lost the battle of Sedan,
not the Napoleon who died on Saint Helena.
c. He has a Rembrandt at home.
d. This is a Ford.
e. The Bonnets were engaged to dine with the Lucases.
6.3. There is also a small class of common nouns that are
recategorized into proper names on our cultural and historical basis: the
Book, the Virgin, the Flood, the Lord, the Savior, the Devil, etc.
6.4. The following classes of common nouns are used without the
definite article and they designate unique objects based on our everyday,
54
routine basis (i) mother, father, brother, sister; (ii) breakfast, lunch,
dinner, supper; (iii) bed, church, school, home, market, college; (iv)
winter, spring, summer, autumn, harvest; (v) morning, night, evening,
noon, midnight, dawn, daybreak, dusk, sunrise, sunset. Consider the
following examples:
(32) We went to bed at midnight / I remember Allworthy at college
/ At one hour he was sure to be at church; at another at market;
in his office at a third; and at home when respectable men
should be at home / Dinner is served ! / I slept undisturbed till
morning / Aunt was always at law with her tenants.
7.0. Indefinite Descriptions

The meaning of indefinite descriptions such as a girl, a dog, some
children is associated with the notion of novelty of the referent on the part
of the hearer. Indefinite descriptions introduce new discourse referents.
As already mentioned, the indefinite article a is compatible only
with countable singular nouns (e.g., a boy, *a milk).
Pragmatically, indefinite descriptions refer ‘exclusively’ to a member
or a subset of members of the objects accessible to the speaker. It follows
that there are other members that belong to the same set but which are not
included in the reference of indefinite descriptions. When using an
indefinite article the speaker “refers to a proper sub-set, i.e., not-all, of the
potential referents of the referring expression” (Hawkins 1987:187).
(33) a. Fred bought a book from Heffer’s.
b. He was dismayed to find that a page was torn.
c. He was dismayed to find that some pages were torn.
A book or a page refers to only one member (arbitrarily chosen) of
the set of objects pragmatically relevant (i.e., the set of books in a
bookshop, the set of pages in the book). The nominal some pages refers
to a subset of the set of pages that make up the book introduced in (33a).
The necessary condition for the use of a/some indefinite
descriptions is that there must be at least one more such object in the
resource situation whose usage is excluded.
55
Whenever the set consists of only one member the indefinite article
cannot be used. Indeed, reference to a unique object/individual can only
be achieved by using a definite description. The pragmatic notion of
‘exclusiveness’ of indefinite descriptions explains the ill formedness of
the following examples:
(34) a. *Fred lost a nose in the war.
b. *I decided not to buy the house because a roof was leaking.
c. The / *a prettiest girl at the party was Sue.
d. The / *an only girl at the party was Sue.
e. Fred bought the / *a bigger dog of the two.
The possibility of choosing at random just one member out of the
relevant set of objects without uniquely identifying it points to the close
relationship of the indefinite article a with the numeral one, i.e., the
indefinite article is viewed as the unstressed variant of the numeral one.
7.1.1. The Ambiguity of Indefinite Descriptions

It has been noticed that indefinite nominals (i.e., nominals prefixed
by determiners like a, one, many, much, little, some, any, no) are
ambiguous between a specific and a non-specific reading/interpretation.
Consider the examples below:
(35) a. John caught a fish and ate it for dinner.
b. John caught a fish and Susan caught one too.
In (35a) the indefinite nominal a fish introduces a new referent in
discourse and may be referred to later on as it or the fish John caught (a
fish functions as an antecedent for the definite pronoun it). In this case,
we have to do with the specific reading of the indefinite description, i.e.,
the speaker has a particular fish in mind and the hearer can identify this
referent despite the indefiniteness of reference. In (35b) the indefinite
pronoun one indicates that John and Susan caught two different fish. In
this case, we have to do with the non-specific reading of the indefinite
description a fish.
Speakers may use other linguistic means to disambiguate between a
specific vs. non-specific reading of an indefinite description. Consider the
following examples:
56
(36) Mary played a sonata last night.
a. Mary played a certain sonata last night.
b. Mary played a single sonata last night.
(37) Susan bought a book on flowers.
a. Susan bought a certain book on flowers.
b. Susan bought a single book on flowers.
In (36) and (37) the specific reading of the indefinite description is
effected by the expression a certain sonata/book while the non-specific
reading is drawn by the expressions a single sonata/book.
Every time opacity creating elements such as quantifiers (e.g.,
every, many), verbs of propositional attitude (e.g., believe, want, wish,
think), modal verbs (e.g., can, may, would) are used with indefinite
descriptions, they give rise to ambiguity between a specific and a non-
specific reading (Cornilescu 1986, Baciu 2004).
(38) John reads a book every day.
a. There is a certain book that John reads every day.
b. Every day John reads another book.
(39) Susan wants to marry a Norwegian.
a. There is a certain Norwegian that Susan wants to marry.
b. Susan wants to marry any individual who is Norwegian.
In (38) and (39) the a) sentences have a specific reading while the
b) sentences have a non-specific interpretation. On the specific reading
(also called referential) the speaker has a particular individual in mind
about which the sentence is true. On the non-specific interpretation (also
called existential) the indefinite expression only asserts the existence of
whatever entity applies to it. The speaker does not have a particular
individual in mind and any individual that satisfies the description will do
(cf. Cornilescu 1986).
Moreover, contexts that contain verbs such as need, owe, expect or
negative contexts (Cornilescu 1986) favour the non-specific
interpretation of indefinite descriptions:
(40) a. You owe me an apology. You must make *it / one right away.
b. John needs a shirt. Give *it / one to him.
c. Bill doesn’t have a car, but he will have *it / one next year.
57
7.1.2. Quite often specificity is a matter of pragmatic context. As a
rule, it is the speaker who knows the reference of the indefinite
description, but not the hearer (i.e., speaker’s specificity)
(41) A cousin of mine called this morning.
It may also be the case that the reference of the indefinite
description is known to another participant in the event, but not
necessarily to the speaker or the hearer (i.e., epistemic specificity, Farcas
1995 in Cornilescu p.c.)
(42) John met a friend at the opera.
In (42) a friend is specific to John but not necessarily to the speaker
or the hearer.
8.0. The Values/Uses of Indefinite Descriptions

8.1. The Epiphoric Value of Indefinite Descriptions
Since the article a is mainly used to introduce a new referent in
discourse its main function/use is epiphoric:
(43) a. As it grew dark, a ruddy glare came out on the hilltop, and
out of the glare the diminished commotion of the flare.
b. These two men offered a contrast – the contrast no so much
of generations (although Appleby was by full twenty years
younger) as of two epochs of English life.
c. He was armed with a rapier and a dagger; the rapier he held
in his right hand, the dagger in his left.
On its epiphoric use the indefinite article has a specific reading: it
functions as an antecedent for a definite description (see the examples
above). The speaker has a certain object / individual in mind and this
object / individual is easily identified by the hearer with the help of the
subsequent definite description.
8.1.2. The Numerical Value of Indefinite Descriptions
Given the affinity between the indefinite article and the numeral
one (the indefinite article a/an is the unstressed variant of one) its second
value/use is the numerical value.
58
(44) a. He had learned a routine but he was essentially untrained
and unspecialized.
b. I need a new dress.
c. Mary offered John a cigarette.
The numerical / counting value of the indefinite article gives rise to
a non-specific interpretation of indefinite descriptions: the identity of the
referent is arbitrary to both speaker and hearer.
8.1.3. The Generic Value of Indefinite Descriptions

The indefinite article can also be used with a generic value (to be
discussed in detail in section 11.3. below):
(45) a. A lion is a noble beast.
b. A symphony has several parts.
9.0. Generic Sentences

The subject of a generic sentence can be a definite description, an
indefinite description or a plural noun with null determiner (called Bare
Plural):
(46) a. The dog is intelligent.
b. A dog is intelligent.
c. Dogs are intelligent.
Generic sentences have the following main properties:
The subject noun in generic sentences has a non-specific
interpretation (i.e., it does not refer to a particular individual).
The predicate in generic sentences designates a property of the
subject individual. The predicate is always used in the generic present.
Generic sentences are ‘atemporal’: they do not specify a particular
moment or interval of time at which the property predicated about holds.
They always contain adverbs such as generally, usually, typically (called
adverbs of quantification) which may but need not be lexically realized.
Since generally, usually, typically are different in interpretation from the
universal quantifier always, it follows that a generic sentence has to be
true in a significant number of cases (not necessarily always). That is, a
sentence such as Dogs are intelligent is true if there is a significant
59
number of dogs that are intelligent and the fact that my dog is not smart
does not make the sentence false.
10.0. A Note on Bare Plurals (BPs)

The term ‘bare plural’ is used to designate individuals / objects such
as dogs, cats, pillows, tigers or men. They are nominals with plural nouns
that lack a determiner (Carlson 1977).
Traditional grammarians called these expressions nominals with
zero or null determiner in the strong belief that the ‘zero’ article is the
plural counterpart of the indefinite article a(n):
(47) a. Spot is a dog.
b. Spot and Collie are dogs.
It was also noticed that both the indefinite singular nouns and bare
plurals (BPs) have generic uses:
(48) a. A mammal bears live young.
b. Mammals bear live young.
However, in spite of the above pairings between an indefinite noun
phrase and a BP, the two have different semantic properties.
10.1. Semantic Differences between Indefinite DPs and BPs

10.1.1. Anaphoric Processes
It is known that indefinite descriptions may have two anaphoric
pronominal correlates: it (which is anaphorically used) and one (which is
not anaphorically used):
(49) a. Kelly is seeking a unicorn and Millie is seeking it too.
b. Kelly is seeking a unicorn and Millie is seeking one too.
Consider now a sentence that contains a BP:
(50) Queenie is seeking unicorns and Phil is seeking them too.
Although them is the plural of it, in (50) them is not used
anaphorically: it is not the case that Queenie and Phil are looking for the
same unicorns.
60
10.1.2. Specific vs. Non-specific Readings
Another difference between indefinite descriptions and BPs has to
do with their distinct specificity interpretations. The sentence in (51)
below, which contains an indefinite description, is ambiguous between
these two readings:
(51) Millie wishes to talk with a young psychiatrist.
On the specific reading of (51) Millie wants to talk with a particular
psychiatrist who is young. On its non-specific reading, Millie wants to
talk with any psychiatrist who is young.
However, sentence (52), which contains a BP is not ambiguous in
interpretation: only the non-specific reading is possible:
(52) Millie wishes to talk to young psychiatrists.
All these semantic differences between indefinite descriptions and
BPs clearly show that plural nominals with null determiner (i.e., BPs) are
not the plural counterpart of a(n) nominals. Rather, the plural counterpart
of a(n) + noun are nominals containing some + nominal:
(53) a. A dog is sitting on the lawn.
b. Some dogs are sitting on the lawn.
In what follows more arguments against treating BPs as the plural
counterpart of a(n) nominals and further characterizing properties of BPs
are presented.
10.1.3. Predicates that Occur only with BPs

There is a specific class of predicates that combines only with BPs
but not with indefinite descriptions: be widespread, be common, be
extinct, be indigenous to, be in short supply, come in many sizes, be
everywhere. Consider the following examples:
(54) a. Cats are common / extinct / widespread / in short supply
b. *A cat is common / extinct / widespread / in short supply
Carlson (1977) suggests that what all BPs have in common is that
they are proper names for kinds of objects. Kinds are proper names for
classes of individuals because both proper names and kinds share several
61
distrubutional properties. For instance, both BPs and proper names can
occur in generic/habitual sentences:
(55) a. Jack is a drunkard.
b. Birds fly.
Moreover, both BPs and proper names can be substituted by so:
(56) a. Slim is so called because of his slender build.
b. Cardinals are so called because of their colour.
The linguistic fact that led Carlson (1977) call classes of
objects/individuals kinds is the close similarity in behaviour between
kinds and the noun kind. BPs occur in several structures where the noun
kind can be used:
(57) a. John repairs cars / this kind of car for a living.
b. Cats were / This kind of animal was seen everywhere.
Moreover, among other shared properties, structures with kind
evince the same anaphoric phenomena as BPs:
(58) a. Harriet caught this kind of animal yesterday and Max caught
it earlier today.
b. Harriet caught cats yesterday and Max caught them earlier
today.
In (58a) and (58b) the reading of the definite pronouns it and them
is non-anaphoric.
11.1. Bare Plural Generic Sentences

When used in generic sentences, BPs (which always designate kind
classes of objects) have two possible interpretations:
a) as expected, BPs receive a generic reading when they occur with
predicates (in the present tense) that designate permanent properties of
the objects that realize the kind:
(59) a. Whales are intelligent.
b. Tigers are ferocious.
62
The explicitly or implicitly present quantifier generally quantifies
over the objects that realize the kind (i.e., most whales are intelligent /
most tigers are ferocious).
b). when BPs occur with predicates that do not designate permanent
properties of the objects (that realize the kind) and these predicates are not
in the present tense, then the sentence containing the BP has an existential
reading:
(60) Whales have attacked the ship on several occasions.
Note that the sentence in (60) implicitly contains the existential
quantifier some and it is bound in time and space. In sentence (60), the BP
whales is still a kind designating expression but it is slightly different
from the BP in the generic reading (i.e., it designates temporal stages of
the kind ‘whales’).
11.2. Definite Article Generic Sentences

The sentences below unambiguously show that definite article
generics are kind level constructions:
(61) a. The dog is widespread / Dogs are widespread.
b. The tiger is striped / Tigers are striped.
c. The wolf is getting more rare as you move north / Wolves are
getting more rare as you move north.
d. The American has, on the average, two children / Americans
have, on the average, two children.
Generic definites show ‘conceptual’ uniqueness.
Somewhat surprisingly we come across generic sentences where
the definite article is followed by a noun in the plural. They are called
definite article plural generic sentences:
(62) a. The airlines charge too much.
b. The generals usually get their way.
c. The lions are noble beasts.
d. The Italians are lazy.
There is a slight difference in interpretation between bare plural
generics and definite article plural generics (cf. Cornilescu, 1986).
63
Hawkins (1977) makes the following comment on the difference between
his examples in (63):
(63) a. Italians are lazy.
b. The Italians are lazy.
“The former is more damning than the latter. (63a) claims that
laziness is an inherent attribute of Italians. By contrast, (63b) involves a
pragmatic restriction of the definite reference. The Italians, therefore,
generally refers to fewer individuals than Italians” (Hawkins 1977:217)
(i.e., a paraphrase of sentence (63b) would be ‘out of the group of
workers, the Italians are lazy).
11.3. Indefinite Article Generic Sentences

Indefinite article generics raise many intriguing properties.
Indefinite article generics do not involve ordinary indefinite reference.
In the generic sentence below:
(64) A good teacher loves all students.
the indefinite article generic expression a good teacher has no reading in
which a specific teacher is referred to.
Moreover, indefinite article generics cannot be anaphorically
resumed by definite descriptions:
(65) a. A pork chop is tender (indefinite generic = a pork chop)
b. *The pork chop is nourishing (definite description =
the pork chop)
The question is what property of indefinite descriptions makes it
possible that they receive a generic reading since they usually refer to one
individual. Nunberg (1976) claims that the generic property ascribed to an
indefinite generic nominal holds by virtue of class membership. The
property may be essential or accidental, but the individual acquires it in
virtue of its class membership:
(66) a. A unicorn has a single horn. (essential property)
b. A symphony has four movements. (essential property)
c. A baby-sitter gets $ 2.00 an hour. (accidental property)
64
Christophersen (1939) and Hawkins (1978) make the insightful
remark that a singular indefinite generic in the examples above still
involves reference to one individual (as opposed to the whole class), an
individual that is chosen at random or arbitrarily by virtue of class
membership.
Moreover, since the indefinite generic does not allow a particular
individual but an arbitrary one to be picked out of the set, the indefinite
generic can be used when the existence of individuals satisfying the
description is not presupposed:
(67) A / every perpetual machine runs forever.
Generic sentences with indefinite noun phrases also allow a
prescriptive interpretation. Consider the following contrasts (Nunberg
1976 in Cornilescu 1986):
(68) a. A Christian is forgiving.
a’. *?Any Christian is forgiving.
b. A pork-chop is tender.
b’. *?Any pork-chop is tender.
The sentences in (68a) and (68b) tell us only how a Christian and a
pork-chop are expected to be. (The any-sentences are descriptive, they do
not allow exceptions, and hence, the sentences are not well formed).
These sentences are understood as similar to sentences that involve
explicitly evaluative expressions as in the examples below:
(69) a. A true Christian is forgiving.
b. A good programmer is smart.
c. A good pork-chop is tender.
Generic sentences like those in (69) cannot be falsified in case it
turns out that, for instance, a Christian proves to be non-forgiving in his
acts. This is because this type of generics acquires a prescriptive
interpretation.
65
4. THE CATEGORY OF GENDER

Gender is another functional category of the lexical category noun
alongside number, determination and case.
Indo-European languages distinguish three genders of nouns:
masculine, feminine and neuter. This division reflects the concepts of
animacy, inanimacy and sex manifested in language.
Indo-European languages evince roughly two types of gender:
semantic gender (where the natural sex of an entity matches grammar) and
grammatical gender (where the natural sex of an entity is not necessarily
the criterion according to which a noun belongs to a certain gender).
Modern English has basically semantic gender. Gender information
is incorporated in the lexical information of certain words (e.g., man/he –
masculine, woman/she – feminine, stick/it – neuter). Moreover, the
systems of anaphoric pronominal reference by personal pronouns (he,
she, and it), possessives (his/her book), reflexive pronouns
(himself/herself) also encode gender information.
Other languages such as Romanian, French or German have
grammatical gender (given in the dictionaries). Grammatical gender
distributes gender according to other criteria: there are morpho-syntactic
markers of gender such as the quality of the stem of the word (consonant
or vowel) and the system of determiners (e.g., le/la in French, un/o in
Romanian). Languages that have grammatical gender do not observe
gender classification of nouns in terms of the sex distinction animate
(male-female) versus inanimate (neuter). For instance, in Romanian
covrig is grammatically masculine, corn is grammatically neuter while
sentinelă is grammatically feminine. In German das Fräulein and das
66
Mädchen are grammatically neuter. Jespersen (1931) offers the following
divisions of gender in Indo-European languages:
(1) Nature Grammar

(sex) (gender)
male beings masculine words

female feminine
sexless things neuter
In what follows, the discussion of gender in English mainly

concerns listing nouns that evince semantic gender. (The description
draws on Ştefănescu 1988).
In English, the distinctions between animate / inanimate and human
/ non-human cut across the classification of nouns according to gender.
2.0. The gender of Animate Entities
Nouns that denote human beings and animate entities may or may
not contain gender information.
2.1. The gender of Male/Female Beings

Nouns that designate human beings that are marked for gender are
organized in pairs: one member of the pair designates the male being and
the other the female being. Several of these nouns have a third member
that designates either member of the pair but is unmarked for gender.
Consider the list below:
(2) man woman person/human
husband wife spouse
father mother parent
boy girl child
son daughter child
lad lass youth
king queen sovereign/monarch
bridegroom bride
bachelor spinster/old maid
67
brother sister sibling
uncle aunt
nephew niece
lord lady
master mistress
monk/friar nun
wizard witch
Other nouns that denote human beings bearing gender information
form the feminine member by adding specific suffixes: #ess, #(t)rix, #ina,
#ette. Consider the list below:
(3) prophet – prophetess count – countess
peer – peeress shepherd – shepherdess
poet – poetess heir – heiress
host – hostess prior – prioress
baron – baroness god – goddess
lion – lioness prince – princess
executor – executrix hero – heroine
czar – czarina suffragette / usherette
Nouns that denote animals are also organized in pairs: one member of
the pair designates the male animal and the other the female animal. Several
of these nouns have a third member that designates either member of the pair
but is unmarked for gender (illustrated in the (4a) list below). Other nouns
denoting animals use the male animal to designate either sex (illustrated in
the (4b) list below). Other nouns use the female animal to designate either sex
(as in the (4c) list below) while still other nouns make use of compound
nouns with the pronouns she/he, the adjectives male/female or proper names
to indicate sex (as in the (4d) list below):
(4a) stallion mare horse
bull cow
ram ewe sheep
boar sow pig/swine
stag hind deer
cock hen fowl
68
(4b) dog bitch dog
ruff reeve ruff
(4c) gander goose goose
drake duck duck
(4d) otter dog-otter bitch-otter
fox dog-fox bitch-fox/vixen
cat tom-cat tabby-cat
ass jack-ass jenny-ass
goat billy-goat nanny-goat
hare buck-hare/jack-hare doe-hare
rabbit buck-rabbit doe-rabbit
pheasant cock-pheasant hen-pheasant
pigeon cock-pigeon hen-pigeon
bear he-bear she-bear
2.1.1. Nouns that Lack Gender Specification

There is a large class of nouns that designate human beings, which
does not contain gender information: the same noun is used to denote
both males and females. For these nouns gender information is attributed
at the level of the sentence generally by the presence of a personal
pronoun, a reflexive or a possessive (see the examples in (6)). Consider
the lists below:
(5) a. relative, friend, guest, enemy, servant, fool, criminal,
prisoner, thief, neighbour, stranger, foreigner, artist,
chairman, cook, engineer, professor, teacher, liar, student,
inhabitant, musician, etc.
b. European, American, Dane, Londoner, etc.
c. Christian, Lutheran, Mohammedan, heathen, atheist, republican,
democrat, candidate, member, partisan, etc.
(6) a. The teacher praised her students.
b. They asked me to send them to the author if I should hear
who she was.
69
3.0. Referential gender
In what follows we discuss the problems raised by referential gender:
the anaphoric use of the third person singular pronouns he, she and it.
3.1. In the normative pattern the third person singular pronouns he,
she and it are used anaphorically to point to a noun with which these
pronouns agree in gender. He/she are used for human beings while it is
used for objects. Consider:
(7) a. The bride was not pretty nor was she very young.
b. Dan must demonstrate what he really wanted.
c. A ship is classed according to its tonnage.
The normative pattern of the third person singular pronouns used
anaphorically is often disregarded in colloquial, informal speech and
literary style.
3.2. In colloquial, informal speech people may either upgrade or
downgrade a non-human entity.
3.2.1. A non-human entity is upgraded when it is referred to either
as he or as she instead of the normative it. Upgrading indicates various
degrees of positive involvement on the part of the speaker. Almost
everything can be upgraded: ships, boats, steamers, balloons, airplanes,
animals, football teams, pieces of furniture, watches, pipes, kites, etc.
(8) a. A mare with her young
b. The fly…there was no courtesy in him.
c. The dogs of the house take no notice of this deer, being used
to her.
3.2.2. When people colloquially make use of it to refer either to
human beings or to objects that have previously been upgraded we have
to do with downgrading. Downgrading indicates various degrees of
negative involvement on the part of the speaker. Consider:
(9) a. “Dear me”! said Bengsington “that’s a cousin of the Prime
Minister, isn’t it”?
b. I can understand why they took the silverware. But why did
it take my piggy bank?
c. “Ay, ay! as large as life; and missy played the hostess. What
a conceited doll it is”!
70
3.3. In literary English, names of abstract nouns are generally
referred to as he or she (i.e., they are personified).
Nouns such as wisdom, crime, science, life, nature, fate, liberty,
church, music are feminine:
(10) a. I love wisdom more than she loves me.
b. Crime…she is not the child of solitude.
c. Music with her silver sound.
Names of towns are also feminine when used in literary style:
(11) Paris was herself again.
Oxford taught me as much Greek and Latin as she could.
Names of celestial bodies are masculine or feminine: Mars is
masculine, Jupiter is masculine while Venus is feminine. Sun is
masculine like time and year; moon is feminine like the names of the
seasons. Consider some illustrations below:
(12) a. I haven’t seen such a moon for years, she is like a great,
great lamp.
b. The world was opening her arms and calling to Michael .
c. The wind ceased from her whirling about continually.
In many cases the gender of nouns when used in literary style

depends on the nouns’ corresponding gender in Latin (Kruisinga 1931).
71
5. THE CATEGORY OF CASE

The Category of Case has a long history in the grammatical study
of human languages. It stems from the Classical Greek word that means
declension or modification of nominal categories (nouns, pronouns and
adjectives). Traditional grammar was primarily concerned with providing
answers to questions such as: “how does the variance of the
morphological forms of case yield a difference in meaning?” In modern
linguistics (e.g., Generative Grammar) it came to refer to the expression
of relationships between morphological forms of nominals and the
interpretational relations they bear in a sentence. Since Chomsky (1965),
grammatical relations have been taken to be structurally determined; so,
the morphological shape of a given nominal is also determined according
to the structural position of a nominal in the sentence.
It is well known that languages differ in terms of their way of
expressing morphological case on nominals. For example, in Latin every
noun has six forms of case, each of which is expressed by declining its
stem. While, like Latin, many languages of the world (Russian, Finnish,
Georgian, Romanian, etc.) have morphologically distinct forms of case
for all nominals, it is only for pronouns that English makes the distinction
in terms of their morphological shapes, and the distinction is only binary:
nominative (I, you, he, she, it, etc.) vs. non-nominative (me/my, you/your,
him/his, it/its, etc.). Still more curiously, in languages like Chinese or
Thai, there is simply no morphological distinction of case. However,
Generative Grammar takes the aforementioned differences among
languages with respect to the overt / covert marking of case to be
superficial and attributed to some parametric variations in morphology.
72
The important point is that, whether it is overtly displayed or not,
case should be present in all nominals at a more abstract level. This
abstract notion of case as a theoretical construct is called “abstract Case”
(with capital C) to contrast it with the morphological forms of case (with
small letter c). Under this view of Case and case, the morphological shape
of a given nominal is regarded as the morphological realization of Case,
an abstract feature assigned to that nominal by some rule. The point we
have to retain is that all nominals that appear in a sentence must bear
Case in all languages.
Cases appear on a nominal because they are assigned to it by
another element in the sentence and this assignment of Case happens due
to some specific configuration the nominal and the other element occur
in. Cases that are assigned under specific structural configurations are
called structural Cases (i.e., Nominative Case, Accusative Case, Dative
Case and Genitive Case) while Cases that are due to the contribution of
certain lexical items are called inherent cases or lexical cases (e.g.,
Ablative Case, Instrumental Case, Dative Case).
The cases of English nominals are Nominative Case, Accusative
Case, Genitive Case and Dative Case and they are all structural cases.
2.0. Formal Configurations: Government and Agreement.

Structural Cases
The basic formal configurations that relate the terms in structural
Case assigning process are government and agreement (concord)
(cf. “Government and Binding” 1981, “The Minimalist Program”
Chomsky, 1995). Government is a notion inherited from traditional
grammar. For instance, William Cobbett (1819) stated that “nouns are
governed, as it is called, by verbs and prepositions; that is to say, these
latter sorts of words cause nouns to be in such or such a case”. The two
elements involved in a government relation are a head and a dependent
term (in our case, the noun that is assigned Case) in a fairly local relation
called government relation. For instance, a prototypical example of
government is government of a nominal by a head verb in a structure
such as: see him (Accusative Case):
(1) I saw him.
73
Since Acc Case is assigned under government, it is a structural
case. Thus, the transitive verb see in (1) assigns Acc case only because it
enters the structural configuration of government with a subcategorized
constituent; otherwise, see can also be an intransitive verb meaning ‘to
understand’ (e.g. I see!).
Dative Case (when preceded by the prepositions to or for) is also
assigned under government:
(2) a. Sue gave the album to him.
b. John did it for her.
In (2a,b) the preposition to or for assigns Dative Case to the pronoun.
As far as Genitive Case is concerned we assume following Abney
(1987) that nominals are headed not by the noun itself (N) but by D
(determiner), which is the functional category responsible for the
inflection within nominals:
(3) John’s strong belief that Mary loves him
In the nominal John’s strong belief the possessive determiner ’s has
the ability to assign Genitive Case to John under government.
Nominative Case is assigned under agreement relation of the subject
and the agreement features on the verb. Take the following example:
(4) He loves her.
In sentence (4) the verb love assigns Acc Case to her under
government. What element in the sentence assigns Nominative Case to
the subject he? Notice that the inflection –s on the verb signals not only
the Present Tense but also agreement with the subject he in the singular.
If we say They love her, the verb agrees with the subject in the plural.
In English, the process of subject agreement / concord is also a
process of Nominative assignment.
3.0. Remarks on the English Cases in Traditional Grammar
3.1. In traditional grammar, linguists tended to associate case
inflections with certain semantic values. For example, Nominative Case
inflection (if any) was attached the sense of designating the actor of an
action:
74
(5) a. John is running now.
b. The girl wrote an essay.
c. Adrian left.
The actor – action characterization of the Nominative case was also
associated with the feature of animacy of the subject. However, there are
hosts of examples that do not conform to these characterizations:
(6) a. The envelope contains the bill.
b. I owe you 10 pounds.
c. John had a shock.
Since actor – action generalization failed, grammarians described
Nominative Case inflection in terms of the syntactic functions of noun
phrases in order to attain some generalization: Nominative Case is
characteristic of the subject of a sentence, of a predicative noun phrase as
well as of appositive noun phrases:
(7) a. The Reformation affected every country in Europe.
b. John is a teacher.
c. This is a book by Chomsky, the linguist.
3.2. Accusative Case inflection was described as designating the
person or thing on which the action of the verb is performed; the noun in
the Accusative case has the syntactic function of a direct object and
occurs after a transitive verb:
(8) a. John beats Paul.
b. John burnt the papers.
However, the semantic characterization fails again to acquire
generality. In a sentence such as: John fears the man, John, in the
Nominative Case is the affected entity while the man in the grammatical
Accusative is the cause of his fear and not the person on which the action
of the verb is performed. Grammarians reached the conclusion that “the
accusative has no meaning at all…but varies according to the infinitely
varying meanings of the verbs themselves” (Sweet, 1891): to kill the
insect / to kill a bottle / to kill the conversation / to kill the evening; to run
a risk / to run a business; to miss the train / to miss somebody; to throw a
ball / to throw a fit / to throw a party.
75
3.3. In Modern English Dative Case is marked by the prepositions
to and for and by word order:
(9) a. I gave flowers to Mary.
b. I gave Mary flowers.
c. I bought a book for John.
d. I bought John a book.
Nouns in the Dative Case function as indirect objects.
In Middle English, the Dative inflection disappeared and it was
replaced by the directional preposition to, which began to acquire an abstract
locative sense also associated with a change of possession meaning:
(10) a. We sent a book parcel to London.
b. We sent a book parcel to our aunt.
In (10b) the animate object designates the person who will come
into possession of the parcel, while London in (10a) simply represents a
destination point. This semantic difference has a syntactic correlate: only
in (10b) can the indirect object invert with the direct object in a well
formed sentence while in (10a) it can not:
(11) a. We sent our aunt a parcel.
b. *We sent London a parcel.
Dative Case is assigned to noun phrases by verbs alone in a very
idiosyncratic way or by the prepositions to or for. Not all verbs that assign
Dative Case occur in both constructions illustrated in (9). There are verbs
that occur only in prepositional dative constructions while other verbs
occur only in prepositionless dative constructions. Thus, verbs like
donate, transfer, select, mention, describe, explain, propose occur only in
constructions where the preposition to is present:
(12) a. We donated $10 to UNICEF.
a’. *We donated UNICEF $10.
(13) a. We transferred some money to Bill.
a’. *We transferred Bill some money.
(14) a. The waiter selected a French wine for us.
a’. *The waiter selected us a French wine.
(15) a. He mentioned the secret to Mary.
a’. *He mentioned Mary the secret.
76
As already said, these verbs’ near synonyms give, send or choose
do permit both constructions:
(16) a. We gave $10 to UNICEF.
a’. We gave UNICEF $10.
(17) a. We sent some money to Bill.
a’. We sent Bill some money.
(18) a. The waiter chose a French wine for us.
a’. The waiter chose us a French wine.
Furthermore, there are verbs in expressions such as allow somebody
a peek, give someone a punch in the nose, give someone a pain in the
neck, give somebody a cold, give somebody a kiss etc., which never occur
with prepositional indirect object:
(19) a. John allowed his sister a peek.
a’. *John allowed a peek to his sister.
(20) a. Measles germs give you measles.
a’. *Measles germs give measles to you.
(21) a. Bill gave Sue a kiss / a little pinch.
a’. *Bill gave a kiss / a little pinch to Sue.
These examples show that Dative Case is sometimes a structural
case (when it is prepositional) and sometimes a lexical case (when it is
assigned by certain lexical verbs). Lexical Dative case idiosyncrasies are
also manifest in many other languages (e.g., German and Romanian). For
example, it is difficult to predict that a Romanian verb such as a crede
assigns not only the structural Accusative Case (a crede pe cineva) but
also a lexical case, namely Dative: cel bogat nu crede celui sărac.
3.4. The Genitive Case was roughly paraphrased as meaning
belonging to, connected with or associated with. Genitive Case in English
is realized as inflectional genitive / synthetic genitive (‘s) and as
prepositional genitive (the construction with the preposition of). The
construction with the preposition of is sometimes used as equivalent to
inflectional genitive. What is of interest to us is under what conditions
one form is preferred and if there are cases in which only one of them is
admitted.
77
(i) Inflectional genitive is preferred with proper names: John’s
father, Shakespeare’s plays. However, when proper names are
coordinated or the noun phrase is complex the of– construction is the rule:
the father of John and Mary, the reign of James the Second, the works of
William Shakespeare.
(ii) When the nominal is preceded by the indefinite article the only
possibility is the construction with of: He was a great admirer of
Shakespeare. When the noun phrase is preceded by the definite article the
two constructions are possible but the construction with of is perceived as
more emphatic: Illness prevented him from attending the bishop’s funeral
/ The death of his uncle, coming almost immediately afterwards, was a
new shock to him.
(iii) With nouns that express characteristics and part-whole
relationships inflectional genitive is the usual, unmarked form: Mary’s
arm, John’s blue eyes. With large and familiar animals inflectional
genitive is preferred but with smaller ones or less known species of
animals prepositional genitive is preferred: The elephant’s trunk / The
trunk of the elephant; His horse’s tail / The tail of his horse; The wings of
a butterfly; The egg of a robin.
(IV) When nouns such as church, cathedral, shop, museum, house
are deleted the of construction cannot be used: Old St. Paul’s burnt down
in 1666; Would you like to visit Madam Tussaud’s?; There was a
tobacconist’s at the corner; I am dining at my uncle’s tonight.
(V) Nouns such as Lady Chapel, student hostel, Israel air force,
child study, doctor degree and afternoon tea are instances of an implicit
genitive construction. After the analogy with these implicit genitives
(without a distinctive genitive ending) many nouns dropped their genitive
ending; the process is extremely productive especially in journalese. The
implicit genitive construction favored the development of noun phrases in
which the inflectional genitive is deleted together with nouns such as:
shop, store and town. For example, Harrods, Longmans, Cooks, St.
Albans, St. Ives are derived from: Harrod’s shop, Longman’s store,
Cook’s shop, St. Alban’s town, St. Ive’s town. Nowadays they look like
plural nouns due to the dropping of the apostrophe in spelling such that
the s inflection is directly attached to the noun. Moreover, they make the
agreement with the verb in the plural: Harrods are offering bargains
these days.
78
(VI) Furthermore, measure phrases and calendar partitions occur
with the inflectional / synthetic genitive construction: today’s paper, in
two year’s time, a five miles’ distance, a three days’ trip, a moment’s
regret, a pound’s weight, a twenty minutes’ delay, yesterday’s deadlock,
last year’s prize, this month’s edition.
4.0. Case as a Conceptual Notion: Case Grammar

We have already seen that Case is related to the morphological
component of grammar (i.e., case inflections that indicate Nominative
Case, Accusative Case, Dative Case) and it is also related to the syntactic
component of grammar (i.e., the syntactic function of subject, direct
object, indirect object of noun phrases). It is clear that these levels of
grammatical analysis are interconnected and information from the
morphological component can be used in the description of the syntactic
component and vice versa.
The question is whether information provided by the morphological
and syntactic components is related to the semantic level of analysis (i.e.,
meaning) as well. In what follows we try to show that all the three levels
of analysis are indeed connected and grammatical levels of analysis are
not autonomous components but highly condition each other.
4.2. Predicates and their Argument Structure

“Case Grammar”, as initially proposed by Gruber (1965), Fillmore
(1971, 1977) and later on developed by Anderson (1971), Givon (1976),
Emonds (1989), Jackendoff (1987, 1990), Levin and Rappaport (1988)
among others, is a semantic theory concerned with the structure of events
function of their participants. Therefore, Case Grammar deals with
predicates and their participant structure.
A sentence always contains a predicate (an expression denoting an
activity or an event) and at least one participant in that action or event.
An expression denoting the participant in the activity or the event is
called the argument of that predicate. Arguments are subcategorized
constituents. For example, in sentences such as the following:
79
(22) a. [The dog] died
b. [Everybody] laughed
c. [The policeman] arrested [the suspect]
the italicized verbs are predicates and the bracketed expressions are their
arguments. All the sentences in (22) are propositions (i.e., they describe
the semantic content of the clause) and the participants in the events of
dying, laughing and arresting are the predicates’ arguments (i.e., the dog,
everybody, the policeman, the suspect). Even if the verbs die in (22a) and
laugh in (22b) are both intransitive verbs and the dog / everybody are
their subjects, there is a difference between these verbs. In (22a) we
semantically interpret the argument the dog as the Patient that undergoes
a change (i.e., from being alive it became dead) while the argument
everybody in (22b) is semantically interpreted as the Agent of the activity
who is volitionally doing the act of laughing and who does not suffer any
change. In (22c) the policeman is the Agent of the predicate as he is the
person who performs the act of arresting while the suspect is the Patient
of the predicate as he is the person who suffers the consequences of the
act.
Semantic cases such as Agent, Patient, etc. are not to be confused
with morphological case or with the syntactic function of a noun phrase
(i.e., subject, object, etc.). Notice that the Patient in (22a) bears
Nominative Case and is the subject of the sentence while the Patient in
(22c) bears Accusative Case and is the direct object of the sentence.
In more recent models of grammar (Government and Binding,
Chomsky, 1981, The Minimalist Program, Chomsky, 1995), semantic
cases are called roles (e.g., the dog in (22a) plays the role of a Patient) or
better thematic roles or theta roles (from the Greek letter theta: θ-role).
The combination of cases that may be associated with a given predicate is
called the role-structure of the predicate or argument structure of the
predicate.
The theory of thematic relations is concerned with the description
of the lexical structure of a predicate, function of the semantic
interpretation of its argument noun phrases. The speaker’s knowledge of
lexical concepts includes knowledge of the semantic interpretation that
noun phrases have in constructions with a verb. This type of knowledge is
stored in our mental lexicon. Particular theta roles (such as Agent,
80
Experience, Patient, Theme, Goal, Beneficiary, Source, Instrument,
Location etc.) are read off from the meaning of verbs.
Yet, there is no agreement about how many such specific thematic
roles there are and what their labels are. Some types are, however,
generally distinguished:
(23) a. AGENT/ACTOR: the one who intentionally initiates the
action expressed by the predicate; it is always animate.
(23) b. PATIENT: the person or thing undergoing the action
expressed by the predicate.
(23) c. THEME: the person or thing moved by the action expressed
by the predicate.
(23) d. EXPERIENCER: the entity that experiences some
psychological state expressed by the predicate.
(23) e. BENEFECTIVE/BENEFICIARY: the entity that benefits
from the action expressed by the predicate.
(23) f. GOAL: the entity towards which the entity expressed by the
predicate is directed.
(23) g. SOURCE: the entity from which something is moved as a
result of the activity expressed by the predicate.
(23) h. LOCATION: the place in which the action or state
expressed by the predicate is situated.
Other authors amalgamate the roles PATIENT and THEME under
the unique role of THEME.
(24) THEME2 : the entity affected by the action or state expressed
by the predicate.
Several thematic roles are illustrated in (25) below:
(25) a. John gave the detective story to Jane
AGENT THEME BENEFECTIVE/GOAL
(25) b. John rolled the ball towards Jane
AGENT THEME GOAL
(25) c. The ball rolled towards the foot of the hill
THEME GOAL
(25) d. Jane had been cold all day
EXPERIENCER
81
(25) e. John likes love stories
EXPERIENCER THEME
(25) f. Love stories please John
THEME EXPERIENCER
(25) g. John bought the book from Jane
AGENT THEME SOURCE
(25) h. John is in London
THEME LOCATION
As already said above, there is no systematic correspondence
between theta roles and their morphological cases or syntactic functions.
In fact, theta roles acquire substance only in relation to the predicate that
requires them, as they depend on the meaning of the respective verb.
Arguments are identified according to the semantic, rather than the
syntactic relation they bear to the predicate.
In what follows we try to provide answers to two questions:
(i) Why are theta roles important to grammar (see section 4.3.) and
(ii) How is the semantic interpretation of argument nouns related to
the morphological component and to the syntactic component
of grammar (see section 4.4.)
4.3. Semantic Fields of Several Verbs of Motion and Location

With the help of theta roles linguists managed to identify several
semantic fields of a great number of verbs and their arguments with a
high degree of generality. For example, the field of movement and
location (which involve Theme, Source, Goal, Location) and the field of
human action and causation (which involve Agent, Theme, Cause) are
just two of the many semantic fields that verbs and their arguments
generally form. These semantic fields constitute conceptual
configurations and show the way concepts are stored in the mental
lexicon. Let us illustrate and briefly discuss the field of movement and
location. Consider:
(26) a. The letter went from New York to Philadelphia.
Theme Source Goal
82
b. The message traveled from Bill to Alice.
Theme Source Goal
Intransitive verbs such as go and travel imply the presence of a
Theme, which changes location from a Source constituent towards a Goal
constituent. The general prepositions are FROM for Source and TO for
Goal. For intransitive verbs of movement, the Theme is selected as
subject. The generality we obtain is that the vast majority of intransitive
verbs of movement (e.g., move, travel, come, go, run, walk, fly, swim,
float, roll, etc.) form the concept of movement and location by using
precisely these thematic roles and not others.
With transitive verbs of movement and location it appears that
movement of the direct object Theme is induced by an Agent / Instrument
/ Cause. There are dozens of non-causative (intransitive)/causative
(transitive) pairs of verbs of movement and thus, other generalizations
may be set up. Consider:
(27) a. The ball rolled down the hill.
Theme
b. He rolled the ball down the hill.
Agent Theme
c. The blast / the stick rolled the ball down the hill.
Cause Inst Theme
(28) a. His voice lowered to a whisper.
b. He lowered his voice to a whisper.
(29) a. The plane flew.
b. She was flying a kite.
4.4. On the Relevance of Theta-Roles in Syntax and Morphology

To see the relevance of semantic/theta roles to syntax and
morphology consider the examples below that are near paraphrases; the
verb empty evinces an alternation in its argument structure:
(30) a. We emptied water from the tank.
Th Source
b. We emptied the tank of water.
Source Th
83
The sentences are subtly distinct: in (30a) we understand that some
water was removed from the tank (change of location of water = change
of location variant) while in (30b) we understand that all water was
removed from the tank (change of state of the tank: from full of water it
became empty of water = change of state variant). However,
morphologically and syntactically, the two sentences are not distinct. In
both sentences the verb empty is followed by a direct object and by a
prepositional object: in (30a) water is the direct object while from the
tank is the prepositional object; in (30b) the tank is the direct object while
of water is the prepositional object. Syntactically, both in (30a) and in
(30b) the verb empty is followed by a nominal (water in (30a) and the
tank in (30b)) that is followed by a prepositional group (from the tank in
(30a) and of water in (30b)). Thus, syntactically, the two sentences are
not distinct either. However, the near paraphrase relation of the two
sentences in (30) is captured only if we notice that the arguments that
follow the verb bear the same semantic relations to the verb empty in both
constructions (as indicated in 30) and that the meaning of (30b) (the
change of state variant) includes the meaning of (30a) (the change of
location variant).
Thus, a representation that allows arguments to be identified in
terms of the semantic relations they hold to the verb is preferable for
capturing the near paraphrase relation. Moreover, other verbs that
describe the same process of removing some substance from a location
evince the same alternation and the same argument structures as the verb
empty (e.g., clean, cleanse, drain).
Another piece of evidence that shows the relevance of
semantic/thematic roles to syntax is provided by the various argument
structures evinced by verbs such as: break, bend, shutter, crack, fold, melt
and many others. Consider:
(31) a. John opened the door.
Ag Th
b. The door was opened.
Th
c. The key opened the door.
Instr Th
84
The examples in (31) show two important things: firstly, all
sentences must have subjects and secondly, some theta roles are more
prominent than others are and grammatical processes (such as
passivization) are sensitive to the relative degree of prominence of roles.
Fillmore (1968) noticed that for each class of verbs there is a
preferred or ‘unmarked’ subject choice: “if there is an Agent, it becomes
the subject; otherwise, if there is an Instrument, it becomes the subject;
otherwise, the subject is the Theme.”
Moreover, we notice that once theta roles are assigned to
arguments by the verb they are preserved, irrespective of the syntactic
configurations/positions in which arguments occur (compare 31a with
31b).
85
6. THE CATEGORY OF ASPECT

The functional category of aspect and the functional category of
tense are tightly related as they both pertain to the domain of time. They
are exclusively verbal categories.
Tense (the grammaticalized form of time, roughly the present tense,
the past tense and the future tense) locates events in time with respect to
the moment of speech. This means that we cannot conceive of a past or
future event unless we have a present moment of time in mind. For
example a sentence such as George left yesterday cannot be interpreted
unless the hearer has a ‘today’, a present moment of time in mind with
respect to which George’s leaving can be located.
Tense is a deictic category (it is oriented towards the time of the
speaking ego): it relates different kinds of events to the speech time and
structures them by the relations of simultaneity and sequence (see the
Category of Tense).
Aspect is not a deictic category. Let us consider the following pair
of sentences:
(1) a. John read a book.
b. John was reading a book (when the phone rang / at 3 o’clock).
The difference between the sentences in (1) is not in terms of tense
(both are in the past tense) but in terms of aspect.
The sentence in (1a) presents the situation in its totality, as a whole,
as completed, while the sentence in (1b) presents only some internal
phases/stages in its development. We do not know when John began
reading the book or whether he finished reading it – we only know that
his reading was unfolding in time when the phone rang/at 3 o’clock.
86
Intuitively, aspect predicates about the size of a situation (the whole
of it or only parts of it) while the contribution of tense is to locate that
situation in time. Both tense and aspect pertain to the domain of time as
situations, irrespective of their size, occur in time. Below we shall see that
the aspect of a situation, just like its tense, is conspicuously defined in
terms of its temporal structure.
1.1. The Perfective – Imperfective Grammaticalized Aspectual

Opposition
In traditional grammars, the notion “aspect” was used with respect
to the perfective-imperfective opposition expressed by inflectional
morphemes on the verb (as illustrated in (1) above).
The perfective aspect in (1a) provides a holistic, summarizing or
unifying view upon the event. The imperfective aspect (instantiated in
English as the progressive aspect and illustrated in (1b)) is concerned
with presenting the event as divided up into internal phases, there being
no concern for the whole situation.
1.1.2. Grammaticalizations of the Perfective – Imperfective

Aspectual Opposition
There are various ways in which languages grammaticalize the
perfective / imperfective aspectual opposition. For instance, Russian and
Chinese use different affixes to distinguish between the two aspects.
English and Dutch avail themselves of syntactic means to signal the
opposition: for instance, the contemporary English form He is working
(be + V-ing) developed historically from He is on/at working (in time, the
prepositions, reduced to a or o, disappeared). Languages like Romanian,
French or Old Greek make use of syncretic means to signal the opposition
(i.e., the grammatical markers of aspect have fused with those of tense).
For instance, the Romanian present tense and imperfect tense signal
imperfective aspect and present and past tense, respectively while tenses
such as the perfect compus, the perfect simplu, the mai mult ca perfect
signal perfective aspect and various sub-species of past tense.
In English, the opposition perfective / imperfective has not been
fully grammaticalized but the opposition non-progressive – progressive is
compatible with it. Progressive aspect is signalled by distinct
87
morphological marking: be – ing (e.g., He is/was singing). Perfective
aspect (also called “simple / indefinite aspect”) is rendered by the simple
temporal form of the verb with no distinct morphological marking (e.g.,
He sang).
The perfective / imperfective aspectual opposition instantiates
grammatical aspect.
2.1. Situation-Type Aspect

However, the category of aspect does not reduce to grammatical
aspect only.
It was long ago noticed that verbs themselves, as lexical items,
contain aspectual information.
For instance, the verb drink designates the activity/process of
drinking but drink a cup of coffee necessarily contains a processual part
(that of drinking coffee) and an endpoint (the cup of coffee is empty at
the end of drinking). Thus the aspectual interpretation of a verb is
modulated by the contribution of its arguments (in our example that of the
direct object argument of the verb drink). Below we shall see in detail that
the aspectual interpretation of a sentence is not decided by the meaning of
the verb alone.
This inherent aspectual dimension of the whole predication is
called in modern linguistics situation-type aspect (Carlota Smith 1991).
3.1. Situation-Type Aspect versus Grammatical Aspect

Situation-type aspect is “covert” as it lacks explicit morphological
marking and it is semantic in nature since its detection is based on the
semantics of various verbs and their argument structure. In contrast,
grammatical aspect is “overt” as explicit morphological markers signal it.
Human beings decide on the situation-type aspect of various
constructions on perceptual and cognitive grounds. However, intuition
fails us quite often and that is why we need principled grounds on which
to identify the aspectual import of situations. We shall see below that both
grammatical aspect (the perfective / imperfective aspectual opposition)
and situation type aspect are defined and identified by using the same
means of characterization: their temporal structure (C. Smith 1991).
88
In sum, the aspectual system of languages is made up of two
components: the aspectual situation type component and the aspectual
opposition perfective / imperfective (C. Smith 1991).
The aspectual situation type component is made up of three
aspectual situations: state aspectual situations, process (or activity)
aspectual situations and event aspectual situations. In the current
literature, state, process/activity and event aspectual situations are also
referred to as eventualities (Bach 1981)
4.1. Temporal Structure and Aspectual Situation Types

Language philosophers (Ryle 1949, Kenny 1963, Vendler
1957/1967 among many others) classified verbs in terms of their
aspectual import and provided grammatical and logical criteria to
distinguish among them. They intended to devise a classification of verbs
alone in various aspectual classes. However, they were highly aware of
the obvious contribution of the verb’s arguments (subject and object) in
aspectually classifying verbs.
We adopt Zeno Vendler’s (1967) aspectual classification of verbs
into states, activities and events (which, in their turn split into
accomplishments and achievements).
Examples illustrating Vendler’s categories are given below:
(2) States: believe, desire, have, own, resemble, love, live in
London, be tall
Activities: swim, walk, push a cart, breathe
Accomplishments: draw a circle, make a chair, deliver a
sermon, recover from illness
Achievements: realize, recognize, spot, lose, find, reach
The criterion that lies behind this aspectual classification of verbs is

their temporal structure: their duration (i.e., whether or not they have
internal stages/phases) and their endpoints (i.e., whether or not they have
endpoints) (C. Smith 1991).
89
4.1.1. States
At an intuitive level, state verbs (such as be tall, believe, love,
desire, live in London) hold over an undifferentiated period of time and
they do not contain endpoints. They are homogeneous eventualities. The
change into or out of a state is determined by an external agent and these
changes do not pertain to the state itself.
For instance, if John is tall he is tall over his adult lifetime and
irrespective of whether he stands up or sits down. In the same line, if John
lives in London his living in London holds over an undifferentiated period
of time (it may be up to the end of his life) but this state changes if he
decides to move over to Manchester.
4.1.2. Processes/Activities
Again at an intuitive level process/activity verbs (such as swim,
drink, write, dig, talk, run, or walk) consist of successive stages that
unfold in time over an interval. Activities contain arbitrary endpoints.
They are also homogeneous eventualities.
For example, if John is swimming his swimming is made up of
successive strokes in time and he may arbitrarily end his swimming when
he is tired.
4.1.3. Events
Events are of several subspecies, mainly accomplishments and
achievements.
4.1.3.1. Again at an intuitive level, accomplishments (such as build a
house, drink a cup of coffee, run a mile, dig a hole, kill) are bipartite events:
they consist of a processual part, with successive stages and a natural
endpoint (or outcome, result, upshot), which constitutes a change of state.
For instance, if John built a house his building of the house consists
of the proper activity/process of building and we can truthfully say that he
built the house only when the result of building is attained: the house
stands erected. The result/outcome signals a change of state as the house
changes from being under construction into a finished house.
4.1.3.2. Achievement verbs (such as recognize, leave, lose, find) are
instantaneous events that consist of a single stage, which results in a
change of state.
90
For example, when we say a sentence such as John found a penny
in the street, it aspectually means that John has the penny in his pocket
the instant he finds it but not before. Such an instantaneous event is a
change of state, as John did not have the respective coin before he found
it but he had it after he found it.
Events (both accomplishments and achievements) are non-
homogeneous or quantized eventualities.
At the most general level of classification, three main classes of
verbal predicates and sentences are distinguished: events (i.e.
accomplishment and achievements), states and processes, characterized
by de Swart (1998) as in the chart below:
HOMOGENEOUS NON– HOMOGENEOUS / QUANTIZED

state process event
STATIVE DYNAMIC
5.0. The Temporal Structure of the Perfective – Imperfective

Aspectual Opposition
Just like aspectual situation types (states, processes/activities and
events), the perfective / imperfective aspectual opposition can also be
characterized in terms of temporal structure (i.e., whether or not the two
aspects contain internal stages and endpoints) (C. Smith 1991).
Grammatical aspect, embodied by the perfective / imperfective
aspectual opposition, focuses either the entire situation (the perfective
aspect) or only parts/stages of it (the imperfective aspect). Thus,
grammatical aspect contributes to the visibility of the whole or only part
of a situation/predication.
In a sentence such as Mary wrote a novel the perfective aspect
presents the situation as completed, as a whole, including both its initial
and final endpoints. The entire situation is predicated of, not only some of
its internal stages.
In contrast, the imperfective/progressive aspect in a sentence such
as Mary was writing a novel focuses on some internal stages of the event
of writing that includes neither its beginning nor its end.
Grammatical aspect tells us about how much we see of a situation.
That is why C. Smith (1991) also uses the term “viewpoint” for
91
grammatical aspect: in her system, perfective aspect is called perfective
viewpoint while imperfective aspect is called imperfective viewpoint.
In Smith’s words “aspectual viewpoints function like the lens of a
camera, making objects visible to the receiver. Situations are the objects
on which viewpoint lenses are trained. And just as the camera lens is
necessary to make the object available for a picture, so viewpoints are
necessary to make visible the situation talked about in a sentence”
(C. Smith 1991:91). “Continuing the analogy of a viewpoint with the lens
of a camera, we shall say that the part focused by a viewpoint is visible to
semantic interpretation. What is focused has a special status, which I shall
call visibility. Only what is visible is asserted….” (Smith 1991:99)
It is essential to understand that the two components of aspect,
i.e., aspectual situation types and grammatical aspects / viewpoints,
although independent, they interact in language.
For instance, the temporal schema of process/activity situation type
(which presupposes arbitrary endpoints) is compatible with the closed
interpretation conveyed by the perfective aspect/viewpoint, e.g., Mary
swam yesterday. On the other hand, processes contain successive stages
and so, they can select for the progressive aspect/viewpoint, e.g., Mary
was swimming. However, state predications, which are temporally
unbound (with no endpoints) cannot occur in the progressive as in
English the progressive aspect/viewpoint shows limited duration, e.g.,
*John is knowing Latin. When the duration of a state situation is
perceived as limited, the state can occur in the progressive; compare John
is silly with John is being silly. In the sections below we shall analyze in
detail the compatibility or lack of compatibility between each aspectual
situation type with the two grammatical aspects/viewpoints.
6.0. Conceptual Features of Situation Types and Grammatical /

Viewpoint Aspects
The following semantic features are assumed to define and
distinguish among situation types (states, processes and events) and
between the grammaticalized aspectual opposition (perfective /
imperfective).
92
6.1. [± Stativity]
The feature that is crucial in the characterization of situation types
is the feature [±stative]. Cognitively this distinction between ‘stasis’ and
‘motion’ (change) is fundamental. The feature of stativity divides
situation types into the classes of states and non-states (Parsons 1990).
States are the simplest of the situation types. From a temporal point
of view, they consist only of an undifferentiated period of time, without
endpoints.
Non-stative situations form the natural class of ‘events’ (activities
and events proper). As shown by Ross (1972) non-statives are ‘doings’;
they are dynamic, involving causation (which includes both agentive and
non-agentive subjects), activity and change. They consist of stages/phases
rather than undifferentiated moments. The successive changes of
activities and accomplishments over time reflect dynamism, as do the
single stage of achievements.
6.2. [± Telicity]
Situation types are also characterized as [±telic].
Telic eventualities are directed towards a goal / outcome, i.e. they
have an inherent culmination point. When the goal is reached a definite
change of state occurs and the event is complete (Garey 1967) i.e. it
attains a final/resultant state. The goal may be intrinsic to the event and in
this case the attainment of the goal constitutes the natural endpoint of the
event. A good example is the verb break, which is an inherently telic verb
(e.g., John broke the stick in a second) (Ramchand 1999). It follows that
telic events are bounded events, i.e. the final point must be specific.
Atelic eventualities are simply processes, which are realized as soon
as they begin. Atelic eventualities have no (inherent) endpoint, but rather
an arbitrary final point: they can stop or be terminated at any time. For
example, if one doesn’t continue running, one automatically ceases
running. Activities / processes and states are atelic situations.
6.3. [± Duration]
The feature [±durative] also categorizes aspectual situations: some
take time (activities, states, accomplishments are durative) others are
instantaneous (achievements).
93
6.4. Compositionality
Another crucial aspectual property of all sentences is that the
aspectual interpretation of a sentence is established compositionally
(cf. Smith, 1991).
The aspectual center of a sentence is the verb but it is not the only
factor of importance, since situation types are associated with verb
constellations (i.e., the verb and all the other elements present in the
sentence).
The suggestion put forth by C. Smith (1991) is that the aspectual
feature of the verb may be overridden when combined with other
linguistic forms in a sentence.
The examples in (3), taken from Smith (1991:73), illustrate the
interaction between the inherent aspectual values of a few verbs with the
contribution brought in by its arguments; the resultant aspectuality of the
predications is calculated compositionally:
(3) Mary walked. = walk [+activity, -telic]
walk the dog: V [-telic] + Nom [count] = VP[+activity, -telic]
walk to the park: V [-telic] + PP [directional] = VP[+event, +telic]
John built a house. = build [+event, +telic]

build the house: V [+telic] + Nom [count] = VP [+event, +telic]
build houses: V [+telic] + Nom [mass] = VP [+activity, -telic]
7.0. Aspectual Recategorization / Shift

Aspectual recategorization or aspect shift refers to the process by
means of which the aspectual interpretation of a sentence may change and
be different from the aspectual class of the verb in the sentence. This is
due to the fact that the aspectual meaning of a predication is calculated
compositionally: it is given by the aspectual value of the verb in
conjunction with the aspectual contribution brought in by the other lexical
items present in the sentence. In what follows we analyze several such
recategorization processes or aspect shifts brought about by the type of
arguments of the verb (i.e., subject and object) and adjuncts (elements
that are not either subject or object) (Dowty 1979, Filip 1999, Rothstein
2004).
94
7.1. Consider the aspectual information conveyed by the following
predications:
(4) a. Mary walked to school. (perfective / event)
b. Mary was walking to school. (imperfective / activity)
c. Mary walked in the park. (perfective / activity)
Example (4a) presents an event that has a ‘goal’, a natural endpoint
(signaled by the expression to school). The situation is described as
closed, completed at a time prior to ‘now’. The viewpoint of the situation
is perfective. Example (4b) presents the same situation but it does not
convey whether the goal was reached. The progressive form of the verb
indicates that only a partial view of the situation is perceived. The
presence of the progressive marker changes the aspectual interpretation of
the verb + to school – the predication becomes an activity predication
(i.e., he was in the process of going to school). The viewpoint of the
situation is imperfective. Example (4c) presents a completed, closed
situation that does not involve a goal (i.e., in the park does not express
telicity). The viewpoint of the situation is perfective.
As can be noticed aspectual information is given by the linguistic
forms of the sentence: situation type is signaled by the verb and other
items (i.e., in (4) by the verb and the prepositional groups) while
grammatical viewpoint/aspect is signaled by grammatical morphemes.
Only situation type aspect can shift but not grammatical aspect because
only the former is not grammatically encoded.
7.2. Consider the following examples that contain the activity verb
run:
(5) a. John ran last Sunday. (activity, perfective)
b. John ran to the park. (accomplishment, perfective)
c. John ran a mile. (accomplishment, perfective)
Sentence (5a) illustrates the basic aspectual feature if run: it is an

activity verb of motion. The whole predication is an activity predication
and due to the presence of the simple past tense it is also perfective.
However, when this activity verb of motion occurs with a locative of
destination (as in (5b)) or an adverb of extent (as in (5c)) the whole
95
predication is interpreted aspectually as an event due to the semantic
contribution of the prepositions (i.e., accomplishments).
7.3. One of the tests that distinguish between activities and events is
their restriction of occurrence with time-span adverbs such as ‘for an
hour’ and ‘in an hour’. Activities select ‘for x time’ adverbs while events
select ‘in x time’ adverbs:
(6) a. John swam for an hour.
b. *John swam in an hour.
(7) a. John built the kite in an hour.
b. *John built the kite for an hour.
However, when a sentence contains an event (accomplishment)
verb such as eat something or build something and a time-span adverb,
the aspectual interpretation of the whole sentence depends on the
grammatical type of the direct object of the verb (Dawty 1979). Consider
the contrasts below:
(8) a. John ate a bag of popcorn in an hour. (accomplishment, perfective)
b. John ate popcorn for (*in) an hour. (activity, perfective)
(9) a. John built the kite in an hour. (accomplishment, perfective)
b. John built kites for (*in) an hour. (activity, perfective)
In (8b) the direct object of the verb eat is a mass noun (i.e.,
popcorn) and in (9b) the direct object of the verb build is an indefinite
plural (i.e., kites). These features of the direct objects shift the aspectual
interpretations of the sentences in (8) and (9): from basic
accomplishments as in (8a) and (9a) into activities as in (8b) and (9b).
Similarly, sentences that contain event verbs of the achievement
type may shift in interpretation function of the type of direct object or
subject these achievement verbs occur with. Consider the following
examples:
(10) a. John discovered the buried treasure in his yard in two days.
(achievement)
b. John discovered fleas on his dog for (*in) six weeks.
(activity)
(11) a. Tourists discovered that quaint little village for years.
(activity)
b. *John discovered that quaint little village for years.
96
Sentence (10b) recategorizes as activity predication and it is well
formed because John discovered different fleas (not the same flea) for six
weeks in a row. John’s perpetual discovery of fleas qualifies as an activity
not instantaneous achievement. Similarly, in (11a) the indefinite plural
subject (i.e., tourists) of the achievement verb triggers aspectual
recategorization of the sentence: since various tourists discovered the
little quaint village for years on end these ‘multiple instantaneous
discoveries’ shift the interpretation of the sentence into an activity.
7.4. There are also cases when intransitive activity verbs can occur
in transitive resultative constructions, which are aspectually derived
accomplishments:
(12) a. At the opening of the new parliament building, the crowd
cheered the huge gates open.
b. Mary drank John under the table / sick / dizzy.
However, aspectual category shifts are not always available. For
instance, the verbs wheeze and croak are both characterized as ‘sound
emission verbs’ which denote activities (Levin and Rappaport 1995, Filip
1999). Yet, it is only one of them that can be ‘fitted’ into the direction-
motion construction, qualifying as accomplishment:
(13) a. The elevator wheezed to the seventh floor.
b. *The frogs croaked to the pond.
In sum, the conclusion we draw is that the aspectual properties of a
predication are decided at the level of the whole sentence and does not
depend exclusively on the aspectual class of the verb, although its
aspectual contribution is central.
As Verkuyl remarks “aspect is not a matter settled at the verbal
level. I propose that aspect be ‘taken away’ from the verb and be assigned
to higher levels of sentential structure: first of all to the VP [verbal
phrase/group] because this node dominates the verb and the object, and
subsequently to the sentence because the nature of the subject appears to
be a determinant of aspect as well. So, the basic idea is that the verb needs
to be specified as to its having a specific meaning element engaged in the
composition of aspect, but this feature cannot be identified with aspect
itself, because aspect is to be considered a complex sentential property”
(Verkuyl 1989:40).
97
All the examples above show that shifts in aspectual situation types
are to a large extent systematic and predictable as they are based on the
inherent lexical meaning of the verbs and on the properties of the
linguistic contextual elements that induce the shift.
As noticed above, in language, aspectual situation types do not
occur independently from grammatical (viewpoint) aspect. That is, state,
activity and event situation types conveyed by the constellation of the
lexical elements present in the predication co-exist with the grammatical /
viewpoint aspect (i.e., the perfective / imperfective aspectual opposition)
conveyed by grammatical morphemes. These pairings are not always
straightforward and may acquire a variety of interpretations. The sections
that follow present a detailed analysis of these pairings.
8.0. The Perfective Viewpoint and Aspectual Situation Types

Before we proceed with the analysis of aspectual meaning shades
of situation types in the perfective aspect we briefly resume the main
properties of the perfective aspect.
8.1. General Characteristics of the Perfective Grammatical /

Viewpoint Aspect
The presence or absence of grammatical morphemes overtly signals
grammatical / viewpoint aspect.
In English, perfective aspect is identical in form with the past tense
of the verb:
(14) Mary talked with her father.
The perfective viewpoint is the dominant viewpoint since it is

available with the entire range of situation types (states, activities and
events) as we are going to see below.
Semantically, its temporal structure is characterized by two
properties: (i) the situation made visible by the perfective aspect includes
both endpoints of the situation and (ii) the situation is not presented as
durative.
The temporal schema of the perfective viewpoint is presented in
(15) (C. Smith 1991):
98
(15) I………………F (perfective viewpoint)
///////////////////////
The slashes indicate the part of the situation that is focused by the
perfective viewpoint.
The perfective includes both the initial point (I) and the final point
(F) of the situation. Informationally, perfective viewpoint presents
situations as ‘closed’ or ‘completed’.
8.2. States, Activities and Events in the Perfective Aspect

In English the perfective viewpoint interacts with all situation types
but its span depends on the endpoint properties of situation types; in
principle, the situation cannot continue after its endpoint. Compare:
(16) a. activity Pluto chased a car (*and is still chasing it)

b. accomplishment Susan wrote the report (*and is still writing it)
c. achievement The plane landed (*and is still landing)
d. state Sam owned several apple orchards (and he still
owns them)
Sam owned several apple orchards (but he no
longer owns them)
In the examples (16a,b,c) the situations are presented as closed.

There are nevertheless slight differences among these situations function
of the endpoint properties of each situation type. The sentence in (16a)
presents a terminated eventuality since the activity described qualifies as
atelic. The sentences in (16b,c) present intrinsically completed situations
since these events describe telic situation types.
Moreover, the ‘stop’ and ‘finish’ tests distinguish between the two
types of endpoints. Atelic predications felicitously occur with ‘stop’ (stop
chasing the car; stop kicking the ball) while telic events felicitously occur
with ‘finish’ (finish writing the report).
In contrast, stative sentences with a perfective viewpoint, illustrated
in (16d) are flexible in interpretation (as known, states do not standardly
occur in the progressive viewpoint). Since stative situation types do not
conceptualize endpoints in their temporal schema, statives in English are
compatible with both a closed and an open interpretation, depending on
99
the context. On the open reading the state continues into the present and
such sentences naturally conjoin with present tense affirmative sentences
as the example above indicates. Statives also allow for a closed
interpretation i.e. the state has ended; this reading can be conveyed by
conjoining stative sentences with negative present tense sentences.
In sum, the perfective viewpoint with non-stative situation types
makes visible for semantic interpretation the whole situation.
9.0. The Imperfective Viewpoint and Aspectual Situation Types

As above, before analyzing the shades of meaning of aspectual
predication types in the imperfective aspect we first discuss the general
characteristics of the English imperfective.
9.1. General Characteristics of the Imperfective Grammatical /

Viewpoint Aspect
The English progressive is the imperfective viewpoint pendant
found in Romance and other languages.
The English progressive is independently realized as the auxiliary
BE plus the present participle of the verb:
(17) Mary was talking with her father (when I entered).
The progressive viewpoint makes visible only part of the situation,
with no information about its endpoints.
Informationally, sentences in the progressive form are open i.e., the
progressive viewpoint does not linguistically present closed situations.
The progressive viewpoint is felicitous with situation types that are
temporally characterized as having internal stages, namely
activities/processes and accomplishments:
(18) a. Mary was sleeping/running/walking (when I arrived)
b. Susan was writing a report/eating an apple/drawing a
triangle (when I arrived)
c. *Susan was loving me (when I asked her)
Basically, the progressive presents eventualities from an internal
perspective, focusing on the internal stages of non-stative predications.
100
The temporal schema of the progressive proposed by Smith (1991)
is offered below:
(19) I …… //////////….. F
[+stage]
States are disallowed from occurring in the progressive unless they
acquire a marked interpretation of dynamism and volition (see the
sections below) (Comrie 1976, Dahl 1985, C. Smith 1991).
Again in principle, instantaneous achievement predicates cannot
occur in the progressive because they do not contain internal stages as
illustrated in (20a). However, some achievements can occur in the
progressive when the progressive focuses on the preliminary stages of the
event as illustrated in (20b,c).
(20) a. *John was finding a penny in the street (when I saw him)
b. The plane was landing
c. John was dying
The progressive is viewed as an operator: it operates on non-
progressive sentences that denote a non-stative situation (i.e., an activity
or an event) and changes it into a homogeneous situation (i.e., a state).
The English progressive is not the exact counterpart of the
Romance imperfective. First of all, in English the progressive cannot
standardly occur with states (e.g., *She was being beautiful) but it does in
Romance (e.g., Era frumoasă). As we are going to see below, the
Romance imperfective includes the meaning of the English progressive
but it spans much larger intervals.
9.2. Jespersen’s (1933) View on the Properties of the English

Progressive
Jespersen (1933) detected the essential properties of the English
progressive by means of analyzing the sentence: He was hunting.
“The chief use of the expanded tenses [progressive aspect] is to
serve as a frame round something else, which may or may not be
expressly indicated. This is easily understood if we start from the old
phrase he was on hunting, which meant ‘he was in the course of hunting,
engaged in hunting, busy with hunting’; he was, as it were, in the middle
of something, some protracted action, denoted by the substantive
101
hunting. Here on became phonetically a ….and a was eventually
dropped, exactly as in other phrases: burst out on laughing, a-laughing,
laughing/ fall on thinking, a-thinking, thinking; set the clock on going, a-
going, going, etc. If we say he was (on) hunting, we mean that the
hunting (which may be completed now) had begun, but was not
completed at the time mentioned or implied in the sentence; this element
of relative incompletion is very important if we want to understand the
expanded tenses, even if it is not equally manifest in all cases. The action
or state denoted by the expanded tense is thought of as a temporal frame
encompassing something else which as often as not is to be understood
from the whole situation…”.
Thus, in Jespersen’s view the properties of the progressive are:
• it shows a durative, protracted action (i.e., it is durative);
• the progressive focuses a ‘framed interval’ (i.e., it is temporary);
• the action is incomplete.
Let us take a closer look at these properties.
The intuition developed by Jespersen is that the progressive
situation takes place at an extended “framing interval” (i.e., called Event
Time (ET)). The progressive itself focuses only on a “framed interval”
(i.e., called Reference Time (RT)). Thus, the meaning of the sentence He
was hinting is roughly He was hunting (when I saw him).
Essentially, this is also the meaning we attribute to a state sentence
such as: She was here (when he called me).
Other examples that illustrate the important role played by RT in
the understanding of the English progressive are offered below:
(21) Mary was making coffee [when John came home]RT
[at five o’clock]RT
The role of RT is to ‘anchor’ in time the situation described in the
progressive clause for a complete temporal understanding of the whole
sentence. Mary was making coffee, in isolation, is informationally
incomplete. From sentences such as those in (21) we understand that
Mary’s making coffee is true at least at the RT (when John came home,
at five o’clock), although her making coffee usually lasts longer.
From another point of view we can say that the main clause ‘Susan
was making coffee’ (the framing interval) establishes the interval against
102
which the event of ‘John’s coming home’ (the framed interval) is set (cf.
Kearns, 1991). As such, the main clause is the background against which
the foreground information is asserted (when John came home / at five
o’clock).
In sum, the progressive describes eventualities that are durative /
protracted and at the same time ‘temporary’ (i.e., they hold at least at the
RT).
Jespersen definition also includes another important intuition,
namely that the ‘protracted action’ denoted by the progressive sentence
had begun before the time stated or implied in the sentence (before RT).
9.3. Bennett and Partee’s (1972/1978) Formalization of the

Progressive Aspect
Bennett & Partee (1972 / 1978) formalize this intuition as in (22):
(22) The progressive sentence is true at an interval I just in case
there is an interval of time I’ that properly includes I.
The interval I’ stands for the extended Event Time (ET) interval,
while I stands for the interval focused on by the progressive, i.e., RT.
Proper inclusion refers to the fact that no element from the interval I
coincides with the endpoints of the interval I’. (For instance, if we
associate I’ with the string of numbers 1, 2, 3, 4, 5, 6, 7, properly included
strings of numbers in I’ are: 2, 3, 4, 5, 6 but not 1 or 7).
According to the formalization proposed by Bennett and Partee, the
progressive focuses on internal stages of the eventuality and it excludes initial
or final endpoints. This is the ‘incompletion’ property of the progressive.
However, the ‘incomplete’ property of the progressive (i.e., it does
not include endpoints) is more intricate than that.
It is surely uncontroversial that the ‘protracted action’ denoted by a
progressive sentence began before the time stated or implied by RT; it
does not contain an initial endpoint. There are, however, cases when a
progressive sentence does include a final endpoint and yet the sentence is
well formed:
(23) a. John was watching TV when he fell asleep
b. John was crossing the street when a truck hit him
103
It is clear that John stopped watching TV / stopped crossing the street
after he fell asleep or after the truck hit him, respectively (Dowty 1979).
9.4. The Imperfective Paradox (Dowty 1979)

This type of sentences are related to another problem raised by the
progressive, viz., “the imperfective paradox” (Dowty, 1979). Dowty
noticed that the truth conditions of the progressive are sensitive to the
aspectual situation type of the simple sentence (activity or event). Consider:
(24) Mary was dancing/swimming/sleeping/playing the violin
entail (activities)
Mary danced/ swam/slept/played the violin
(25) Mary was crossing the street/building a house/playing a
Mozart sonata (events)
do not entail
Mary crossed the street/built a house/played a Mozart sonata
The progressive activities in (24) entail (i.e., logically imply) the
truth of the simple sentences Mary danced / swam / slept / played the
violin (i.e., if Mary was dancing / swimming / sleeping / playing the
violin she also danced / swam / slept / played the violin). In contrast, the
progressive events in (25) do not entail the truth of the simple sentences
Mary crossed the street / built a house / played a Mozart sonata. As in
(23b) above, she may have been interrupted and may not have finished
crossing the street / building a house, etc. That is the “imperfective
paradox”: it arises mainly with progressive accomplishment situation
type. It looks as if the event/accomplishment described by the progressive
in (23b) can remain forever incomplete since the progressive focuses on
internal stages but tells us nothing about its endpoints (Portner, 2005).
Dowty’s solution to the “imperfective paradox” is to say that the
progressive is not just a temporal relation (i.e., ET includes RT) but also a
modal relation. Modal operators point to the existence of ‘alternative’
states of affairs. That is, modality involves paying attention to what
occurs in non-actual, possible circumstances (worlds). It means that we
have to formalize the intuition that even though John didn’t cross the
street in the actual world, he did cross it in some non-actual, possible
worlds. Possible worlds in which things go in the normal or expected way
104
are called “inertia worlds” (Dowty, 1979). This is useful as a theoretical
notion, since in (23b) we want to describe what came out of what John
meant to do – cross the street – rather than what he actually did – get
killed (cf. Portner, 2005).
Notice that the actual world is a world in which things may go in
the normal or expected way (as in the situation in which Mary may have
finished crossing the street/building a house/playing a Mozart sonata in
the actual world).
How does Dowty solve the “imperfective paradox” of both
activities and accomplishments in terms of the entailment relations stated
above? He does it in the following way. Progressive activities such as
‘Mary was dancing / swimming / sleeping / playing the violin’ entail the
truth of the simple sentences ‘Mary danced / swam / slept / played the
violin’ in the actual world. Progressive accomplishments such as ‘Mary
was crossing the street / building a house / playing a Mozart sonata’
entail the truth of the simple sentences ‘Mary crossed the street / built a
house / played a Mozart sonata’ in some possible world, out of which the
actual world is an option.
Since there are many possible worlds, in Dowty’s system, time is
no longer linear but branching. Thus, the truth conditions for both
progressive activities and progressive accomplishments formulated by
Dowty are:
(26) PROG φ is true at an interval I if there is an interval I’
including I [and I is not a final subinterval for I’]; I’ extends
into some ‘possible futures/worlds’ of I and φ is true at I’.
I’
⎯⎯⎯⎯⎯⎯⎯
[ [ ]
←⎯⎯⎯⎯→
I
In the truth conditions stated in (26), φ stands for the simple non-
progressive form of the sentence. PROG is viewed as an operator: it
operates on non-progressive sentences that denote a basic non-stative
situation types (activities and events) and changes them into
105
homogeneous eventualities (i.e., states); they may finish in the actual
world or in some possible world.
Dowty’s truth conditions captures all the properties of the
progressive noticed by Jespersen (1933).
Thus, the first part of the above definition in (26) states that at
interval I (a limited interval) the progressive sentence is true on condition
that φ (‘John watch TV’ / ‘John cross the street’) is true at I’. This means
that in order for a situation to be unfolding at interval I the situation must
have started before I (i.e. at I’).
The second part of the truth conditions roughly corresponds to
Jespersen’s notion of ‘relative incompletion’ and accounts especially for
the entailments/inferences of accomplishment predicates: the event
denoted by the progressive may ‘finish/culminate’ in one of the ‘possible
worlds/futures’, out of which the actual world is an option.
10.0. States in the Progressive Aspect

10.1. State predicates, which are characterized as [+stative] are
incompatible with the progressive. States are described as having an
abstract quality and an atemporal interpretation.
In the examples below the state describes a ‘characteristic
property’ of the subject John. This property holds independently of
whether at a given moment of time (and place) John does not prove to be
intelligent, a nice person or a hero:
(27) a. John is intelligent.
b. John is a nice person.
c. John is a hero.
Other examples of state predicates that describe character-
ristic/inherent properties of the subject are be tall, be blonde, be erudite,
be green, be widespread.
In general, state predicates of this type do not have the property of
‘agency’ (i.e., in the state predications in (27) John is not the doer of any
action) and do not occur in the progressive.
10.2. However, state predicates that denote transitory properties of
the subject can occur in the progressive (e.g. prototypical examples are be
scared, be afraid, be available). Consider the contrasts below:
106
(28) a. William is a hero. (characteristic property of the
individual William)
b. William is being a hero. (temporary property, valid only
over the interval stated by the
progressive)
a. Mary is a naughty child.
b. Mary is being naughty.
a. Harry is clumsy.
b. Harry is being clumsy.
All the sentences in (28a) describe a rather permanent property

ascribed to an individual.
In all the (b) sentences the interpretation is that the subject
deliberately or intentionally acts in the way described by the predicative
adjective or noun. In other words, the eventuality is viewed as containing
some processual stages/phases (Parsons 1990). This interpretation allows
the use of the progressive with these state predicates and they are
described as temporary.
10.3. Locative and position verbs are a sub-class of state predicates
that have been characterized as ‘interval statives’ as they hold for an
interval larger than a moment (Dowty 1979). Prototypical examples are
predicates like lie, perch, rest, remain, stand, sit, rest, sprawl, live etc.
They may occur with either agentive or non-agentive subjects.
Hence, depending on the context, they may occur in activity,
accomplishment/achievement or state predications:
(29) a. Mary sat down (suddenly) on the chair.(accomplishment or
achievement)
a’ Mary was sitting down when I came in. (process in progress)
(30) a. Mary stood in the doorway. (activity)
a’. Mary was standing in the doorway (process in progress)
when the phone rang.
(31) a. The statue stood in the corner. (state)
When interpreted as accomplishments or activities as illustrated in
(29) and (30), these predicates focus on the position or location of an
agent, hence the maintaining of the position requires will or dynamism.
107
Under this interpretation these verbs can occur in the progressive and
they focus on a process in progress at a given reference time.
As the example in (31) illustrates state verbs of position and
location can also occur with non-agentive/inanimate subjects. Consider
some further examples (Dowty 1979):
(32) a. The magazine lies on the table.
b. The magazine is lying on the table.
a. The rug covers the floor.
b. The rug is covering the floor.
a. The socks lie under the bed.
b. The socks are lying under the bed.
The examples in (32) show that state verbs of position and location
with inanimate subjects can appear either in the simple form of the tense
or in the progressive form of the tense. When these verbs describe
relatively permanent states of the subject they occur in the simple form of
the tense as illustrated in (32a) sentences; when they describe relatively
temporary states of the subjects they occur in the progressive as illustrated
in (32b) sentences.
Moreover, there is a further semantic restriction with respect to the
use of these verbs in the progressive: the subject of the sentence must not
designate a movable object, otherwise the progressive cannot be used. An
immovable object such as a city, a house, etc. cannot occupy a temporary
position or have a temporary location. Consider the following examples
and contrast them with those in (32):
(33) a. New Orleans lies / *is lying at the mouth of the Mississippi
River.
b. That argument rests / *is resting on an invalid assumption.
c. John’s house sits / *is sitting at the top of the hill.
Dowty (1979) also remarks that the acceptability of the progressive
with these verbs may also depend on the context. Compare:
(34) a. ??Two trees were standing in the field.
b. After the storm only two trees were still standing in the field.
108
Moreover, in narrative contexts the progressive of these verbs can
also be used to describe stationary objects that momentarily come into the
observer’s view. The moving observer perceives the stationary objects as
being ‘temporarily’ taken into sight:
(35) When you enter the gate to the park there will be a statue
standing on your right, and a small pond will be lying directly
in front of you.
Motion verbs like flow, run and enter can also be used in the
progressive aspect (but not entailing literal motion). They cannot be used
in the progressive when they describe a relatively permanent state or
property:
(36) a. The river flows through the center of the village.
(a fact of geography)
b. The river was flowing through the center of the village.
(a flood in progress)
11.0. Activities / Processes in the Progressive Aspect

11.1. Processes are homogeneous and atelic situations. Processes
have been described as being made up of internal stages. Their endpoints
are arbitrary.
They naturally occur in the progressive and the internal
stages/intervals of a process predication are always anchored
contextually. In the examples below the temporal anchorage is provided
by the when-clause or by pure temporal adverbs.
(37) a. It was raining heavily when she arrived home.
b. We are traveling now for amusement and instruction.
c. This time last year I was traveling through Europe.
In (37) the process is in progress and it is true at least at the
anchorage time (i.e., RT). The time of the progressive is said to be
definite (Kearns 1999).
11.2. There are cases when the interval of time denoted by a
durational adverb is not felt to be shorter than the full progressive
situation. Consider the following examples:
109
(38) a. All through dinner they were talking of nothing else but
the match.
b. I was knitting for two hours this morning.
c. The band was playing, the flags were fluttering and the
crowd was cheering as the players ran onto the field.
d. Mary was sleeping when I was working.
e. They were watching television while we were working.
The sentences above describe a situation in which the main clause
in the progressive is simultaneous with the event in the subordinate clause
(or with the time interval stated by the durational adverb).
Durational adverbs like for x time, all morning, etc. are not very
common with the past/present progressive, being a survival from an
earlier stage of the language. Mittwoch (1988) remarks that we can still
find the progressive used with vague or hyperbolic durationals as in (39):
(39) a. You were talking on the phone for hours.
b. They were working on that project for ages.
Leech (1971) described the examples in (39) as containing an
element of colloquial hyperbole or exaggeration, their tone being one of
amusement or irritation. They are similar in interpretation with the
sentences in (40) which contain hyperbolic durationals such as always,
forever or constantly:
(40) a. My father was forever getting into trouble with the law.
b. He is always complaining about money.
c. She is always breaking things.
d. My car is constantly breaking down.
12.0. Events in the Progressive Aspect

12.1. Event predications (i.e., accomplishments and achievements)
are described as telic eventualities. They involve a product, upshot or
outcome, which is a definite change of state.
With accomplishment predications the change of state is preceded
by some activity/process: it is the activity/process that can occur in the
progressive. Therefore, accomplishments are complex events. An
accomplishment hides a causal structure of type [e1 causes e2] where e1 is
the causing activity and e2 is the resulting change of state. In the
110
progressive, accomplishments are conceptualized are ‘durative’
eventualities. Consider the following examples:
(41) a. The river was cutting a new channel to the sea, but the men
with the sandbags stopped it from doing so.
b. Maybe she is making a Spanish omelet.
c. Within a stone’s throw of my house they are building
another house.
All the sentences above contain accomplishment predications that
undergo a change of aspectual class due to the presence of the progressive
operator: they shift into process predications that are unfolding in time at
the a given reference time, the reaching of their goal being ‘suspended’.
12.2. Achievement predications, the other subclass of events, denote
eventualities that take place at single moments of time. They are
instantaneous events that have no proper internal parts (phases); hence
the progressive cannot be applied to them.
Achievements focus on the change of state, simply leaving out the
causing activity / process and the causing factor:
(42) a. My father died.
b. My father died (from his wounds in the end).
There are however several classes of achievement verbs that can
occur in the progressive but they acquire distinct aspectual interpretations.
First, we distinguish between the two subtypes illustrated below:
(43) a. The plane was landing when the storm started.
b. John was dying when the doctors operated on him and saved
his life.
c. Mary was winning the race when she stumbled and fell.
d. He was falling asleep when he heard a noise in the garden.
(44) a. We are constantly receiving letters of appreciation.
b. New guests were continually arriving.
c. He’s been discovering fleas on his dog all morning /*John is
discovering a flea.
The examples in (43) have already been discussed: in this case, the
progressive focuses on the preliminary stages of the actual change of
state, rather than on the internal stages of the achievement predicate, as it
111
has none. This ‘process’ must be a kind of immediate ‘prelude’ to a
certain ‘outcome’ (Kearns 1991). Freed (1979) and Kearns (1991)
characterize these achievements in the progressive as ‘derived processes’.
In the second set of examples we have to do with achievement verbs
iterated in series. The progressive picks out only a subset of the series of
events. In the progressive, these achievement verbs are recategorized as
processes due also to the cumulative subjects and objects.
12.3. There is finally another group of achievement verbs that can
occur in the progressive. They form the so-called ‘degree-achievements’
(e.g., melt, widen, cool, age, sink, rise, fall) and belong to the class of
‘vague predicates’ (Dowty, 1979). Semantically these predicates express
changes of state just like other achievements do but different from
standard achievements they allow durational adverbs:
(45) a. The soup cooled for ten minutes / in ten minutes.
b. The ship sank for an hour (before going under completely) /
in an hour.
c. John aged forty years during that experience.
Such predicates are highly compatible with the progressive and in
this case they qualify as processes in progress at a given reference point
(Smith 1991).
(46) a. They are widening the road.
b. The tobacco leaves are drying in the sun.
These types of predicates refer to situations of gradual change
(Dowty 1979).
112
7. THE CATEGORY OF TENSE

The category of tense is a functional category of the lexical
category verb alongside the categories of aspect and mood. It is signaled
by inflectional markers either on the verb itself (e.g., walked) or on the
auxiliary verb (e.g., will walk).
As we are going to see, beside tense inflections marked on the verb,
temporal adverbs as well contribute to the temporal specification of
sentences.
Tense represents the chronological order of events as perceived by
the speaker at the moment of speaking. That is, tense is fundamentally a
deictic category as the chronology of events (i.e., their order in time) is
always established function of the moment ‘now’ of the speaking ego
(i.e., events are present, past or future).
1.1. The Notion of Axis of Orientation

At an intuitive level, the speaker is able to place a perceived event
in time because he sets it in relation with another event with respect to
which the perceived event occurs either before, after or is simultaneous
with the other event. An event that serves as orientation event for other
events is a ‘source event’ and it is said that it creates ‘an axis of
orientation’. There are many kinds of events that serve as ‘source events’
for human beings.
For instance, natural cosmic phenomena can serve as source events
for public axes of orientation (e.g., the sunrise, the sunset, the nightfall,
the stages of the moon). Perceived events can be ordered relative to them:
some events occur at sunrise, before sunset or after nightfall. In a similar
113
way, the 12 points on the face of the clock serve the same role: an event
may occur at 5 p.m., before or after 5 p.m.
Cultural events are also used to play a similar role: the birth of a
famous person, the outbreak of a war, an earthquake, a catastrophic flood,
etc., may serve as source events for public axes of orientation. For
example the birth of Christ is the source event for the Gregorian calendar,
the flight of Mohammed from Mecca (AD 622) is the source event for the
Moslem calendar, etc.
At the level of restricted groups, private events such as a marriage,
the birth of a child or the death of somebody dear can serve as source
events for axes of orientation recognized by the members of the group.
From this description of source events that create various axes of
orientation it follows that there are two order relationships between an
event and a source event on an axis of orientation: the event is either
simultaneous with the source event or the event is sequential to it.
The moment of speech of the speaking ego serves as source event in
grammar. The moment of speech ‘locates’ the speaker in time and other
events are present (i.e., they occur simultaneous with it), past (i.e., they
occur before it) or future (i.e., they occur after it) relative to the moment
‘now’ of the speaker. The present moment ‘now’, which is the moment of
initiating a discourse creates the present axis of orientation.
However, the present moment is a fleeting moment and as soon as
it has been established it inevitably undergoes the change into a past
moment: it becomes a past moment of time with respect to which other
events are simultaneous with, occur before it or after it. A past moment in
time is another source event and it creates the past axis of orientation
(e.g., left, had left, would leave).
By the same reasoning a future event in time can become a source
event and other events can be simultaneous with this future event, can
occur before it or after it. In this case, the future event creates the future
axis of orientation (e.g., will leave, will have left).
The present axis of orientation, the past axis of orientation and the
future axis of orientation are the basic axes of orientation in grammar.
The tenses that serve as source event form the class of absolute tenses (in
English, the present tense, the past tense and the future tense) while all
the other tenses are relative tenses.
114
However, languages do not grammaticalize time in the same way:
for one thing, the number of tenses in natural languages varies and the
grammaticalized tenses do not have exactly the same values in all
languages. On the other hand, we have seen that Aspect is tightly related
to Tense as Aspect is also defined in terms of temporal structure.
The question is what primitive/basic elements make up a tense and
how a grammar of tense can be devised.
In what follows we describe the tense theory proposed by the
logician Hans Reichenbach in 1947, which with minor modifications is
the most widely accepted theory of tense in present day linguistics.
2.0. The Notions of Speech Time, Reference Time and Event Time
Reichenbach’s tense theory characterizes all tenses in terms of three
basic constitutive temporal entities and as we are going to see below the
relations between them define both tense and aspect.
Reichenbach’s tense theory takes three temporal entities as basic
viz., speech time (ST), reference time (RT) and event time (ET). These
primitive temporal notions are characterized as following. Speech Time is
the time at which a certain sentence is uttered, i.e., the moment of
utterance/speech. Event Time is the moment at which the relevant event
or state occurs.
Let us calculate the temporal interpretation of the following
sentence, taking into account two of the temporal notions defined above
(namely ST and ET):
(1) Mary won the prize last week.
ST is now. ET is calculated by taking into account the information
brought in by the past tense of the sentence in conjunction with the past
temporal value of the adverb last week. The relation between these two
temporal notions is written as ET < ST (the sign < indicates that ET
occurs before ST and the point of this sign indicates the temporal entity
that temporally precedes the other temporal notion).
Reichenbach does not explicitly define RT. However, it can be
intuitively characterized as standing for the temporal axis (present, past or
future) specified in the sentence. It is calculated by taking into account the
115
temporal information brought in by the tense of the sentence and by the
information of the temporal adverb.
RT is the key notion of Reichenbach’s system. Indeed, without it
the temporal specification of perfect sentences cannot be calculated.
Consider the following example:
(2) Mary had already won the prize last week.
The event predicated of in sentence (2) is past as it was in sentence
(1) but this time, we understand that the past event of Mary’s having
already won the prize (ET) occurred in the past prior to something else,
which is also past. RT signals this ‘something else’, without which we do
not understand the sentence from a temporal point of view. Thus, the
temporal specification of the sentence is:
ST = now
RT = past, RT < ST
ET = had, already with respect to RT, ET < RT
It is important to emphasize that the temporal notion of RT has
been generalized and enters the calculation of the all sentences. For
instance, in sentence (1) above, ET is taken as unspecified as we do not
know precisely when it took place in the period of last week. Thus, ET
and RT are taken to be simultaneous (i.e., RT = ET). The overall
interpretation of sentence (1) above is:
(1’) ST = now
RT < ST
ET = RT
It can be noticed that Reichenbach’s tense grammar takes ST, RT
and ET to be time moments or time intervals on the time line and tenses
are viewed as means of representing events vis-à-vis the moment of
speech (Hornstein, 1982). Each tense appears represented as a complex
configuration with a characteristic structure whose elements are ST, RT
and ET concatenated by the relations of simultaneity and sequence.
Let us take some more examples and see how the proposed theory
handles them:
(3) Mary is leaving tomorrow.
116
ST is now; the tense form and the temporal adverb point to a time
after ST that is, on the future axis of orientation (ST < RT). The ET is left
unspecified, i.e., we do not know when the event is scheduled to take
place (represented by the formula ET = RT). The overall interpretation of
sentence (3) is:
ST = now
ST < RT
ET = RT
Notice that the tree temporal notions ST, RT and ET enter temporal
calculation in this precise order. However, since RT is the key notion in
Reichenbach’s tense theory it is RT not ST that is first written in a
formula. In fact, in his system the formula of the future tense is RT > ST.
This formula, although counterintuitive, is equivalent to ST < RT. (In
present day tense theories the formula ST < RT is in use (cf. Demirdache
and Etxebarria (2002, 2004)).
Consider now the analysis of sentence (4):
(4) Joan decorated the cake before midnight.
ST is now. In sentence (4), RT is specified by the tense and by the
temporal adverb midnight, which relative to now is past (RT < ST). The
ET, the time when the cake was decorated, is specified by the preposition
before; ET is situated before RT. The overall representation of sentence
(4) is:
ST = now
RT = past, RT < ST
ET = before RT, ET < RT
Consider also sentence (5) below where the speaker at now
anticipates two events in the future:
(5) John will have eaten the cake (by the time you come).
ST is now, RT is future, that is, after now (RT > ST, cf. in
Reichenbach’s system) while ET is before the future RT (ET < RT). The
overall interpretation of sentence (5) is:
117
ST = now
RT > ST
ET = before RT, ET < RT
Consider finally the temporal interpretation of the sentence in (6):
(6) Bill has met her before.
ST is now. RT is present (i.e., in English, an implied now), while
ET is anterior to now as indicated by have. The overall interpretation of
the sentence in (6) is:
ST = now
RT = present, RT = ST
ET < RT
The above examples point to the very important role ST, RT and
ET play in the interpretation of both tense and aspect in a sentence.
In fact, the relations that hold between ST and RT instantiate tense
(present: ST simultaneous with RT, past: ST after RT and future: ST
before RT). On the other hand, the relations that hold between RT and ET
instantiate aspect (perfective aspect: RT coincides with ET, progressive
aspect: RT is included in ET (is ‘within’ ET).
In sum, the relations holding between ST, RT and ET render both
tense and aspect as both tense and aspect are temporal notions. This is a
very welcome result of Reichenbach’s theory as it is clear that tense and
aspect are both present and interfere in a sentence. The temporal and
aspectual formulae presented above form what we call the Abstract
Temporal Representation of a sentence (ATR).
The above examples also clearly illustrate the importance of the
temporal values of temporal adverbs in the correct interpretation of a
sentence.
However, it is known that not all sentences occur with temporal
adverbs in everyday speech. We expect such sentences to be temporally
vague and liable to many interpretations. Yet, this expectation is not
borne out and “speakers deal with incomplete sentences in a strikingly
consistent manner” (Smith, 1977:270). Consider the following examples
whose interpretation in isolation is given by the ATRs to the right:
118
(7) John played the piano. [RT < ST, ET = RT]
(8) Cousin Judith fed the cat. [RT < ST, ET = RT]
Although in sentences (7) and (8) the relation between ET and RT
is not given they are readily interpreted as past tense sentences rather than
past habituals, a reading also available in these sentences. This is so
because, in general, we maximize the amount of information present at
our disposal and a specific reading is usually chosen if it is plausible (cf.
Smith, 1977).
3.0. The General Classification of Temporal Adverbs and Their

Temporal Specification
3.1. Together with Tense, temporal adverbs help us locate in time
the situations talked about. Importantly, temporal adverbs also contribute
to the aspectual interpretation of sentences. The classification we adopt
has been standardly recognized since Bennett and Hall-Partee
(1972,1978) and Smith (1978).
Temporal adverbs fall into the following classes:
a. locating adverbs or frame adverbs
b. duration adverbs
c. completive adverbs or containers
d. frequency adverbs.
First, we discuss the temporal adverbs of duration and completive
adverbs because these types of adverbs also have an aspectual value,
requiring compatibility with the situation type.
3.1. Duration Adverbs
Duration adverbs include expressions like for three weeks/a
month/a day, for a while, since the war/Christmas, at night, all the
afternoon, half the afternoon, for hours, all the time, over the weekend,
through August, a few days, during the war, always, permanently, all day
long, throughout, from June to/till October, all day/night long, etc.
Duration adverbs have been defined as:
– they indicate the duration of the described event by specifying
the length of time that is asserted to take (Bennett & Hall Partee, 1978);
119
– they express measures of time that are not specifically confined
to future or past (Quirk, 1985)
– they contribute to the location of a situation in time (Smith, 1991)
The definitions above suggest that duration adverbs have aspectual

value: they are compatible with atelic sentences and odd with telics. That
is, duration adverbs are sensitive to the aspectual character of the
eventuality description they combine with. They are restricted to
occurring with homogeneous eventualities / situations (processes and
states) as the examples below indicate:
(9) a. Susan was asleep for two hours. (atelic)
b. Andrew swam for three hours. (atelic)
c. (?) John wrote a/the report for two hours. (telic)
d. *The train arrived late for 2 hours. (telic)
The role of a single duration adverb with atelic situation types (as
indicated in (9a,b)) is to locate an eventuality within the stated interval.
The interpretation of the sentences above is that the situation denoted by
the verb lasts at least as long as the denotation of the durative adverb.
Whenever the situation type features and the adverbial features are
compatible, the standard interpretation of the adverb is to locate the
situation within the stated interval.
Whenever telic events occur in the context of duration adverbs (as
in (9c,d)) there is a clash between the aspectual properties of the situation
type and the aspectual properties of the adverbs.
However, there are cases when telic events can occur with duration
adverbs. Such clashes are possibly resolved by a shift in the value of the
verb constellation, which receives a marked interpretation. De Swart
(1998) building on ideas developed by Moens (1987) assumes that the
contextual reinterpretation is made possible by the process called
aspectual coercion. Instantaneous telic eventualities with durative adverbs
(as in (10c)) and durative telic verb eventualities (as in (10a,b)) are
reinterpreted as atelic / durative in the context of duration adverbs:
(10) a. I read a book for a few minutes.
b. Jerry wrote a report for two hours.
c. John knocked on the door for two hours.
120
In (10a,b) the events of book reading and report writing are coerced
into processes (paraphrased as ‘I read from a book for a few minutes’ and
‘Jerry wrote at the report for two hours without finishing it). The sentence
in (10c) receives an iterative reading (paraphrased as ‘John knocked
repetitively on the door for two hours’ such that the instantaneous telic
eventuality is interpreted as ‘durative’).
The other duration adverb in English that evinces sensitivity to the
aspectual make-up of the situation type is the since adverb. As its name
shows, since measures the entire time span of both homogeneous (states
and activities) and non-homogeneous (events) eventualities and it occurs
exclusively with the present perfect tense:
(11) a. John has been here since 3 o’clock.
b. John has sung since 3 o’clock.
c. John has moved to Paris since summer.
Both sentences (11a) and (11b) mean that John arrived sometime
around 3 o’clock and has been here ever since or that he began singing
around 3 o’clock and has been singing ever since. Sentence (11c) means
that John’s moving to Paris lies somewhere within the period from last
summer to the time of utterance.
Remark: The English since adverb is an indefinite one, distinct from
the Romanian de-phrase by which it is translated. The Romanian
equivalent of (11c) is Ion s-a mutat la Paris de vara trecută, and has a
different interpretation. It means that John moved to Paris last summer
and he has been there ever since. Thus, different from since, the
Romanian de-phrase is a dating/definite adverbial as it establishes the ET
of the predication and measures its result state. To date the event in the
present perfect, English makes use of another adverb as in (12):
(12) Since 1991 I have been to Cape Cod only once, namely in the
fall of 1993. (Anagnastopoulou, Iatridou and Izvorski 1998)
The other duration adverbs mentioned in the list above are not
sensitive to the aspectual feature of the eventuality and occur with both
homogeneous and non-homogeneous eventualities.
121
3.2. Completive Adverbs
Completive adverbs are also known as containers (or adverbs of the
interval (Smith 1991) and include expressions like in 2 hours, in five
minutes, within two months.
Their role is to locate a situation/eventuality at an interval during
which the event is completed/culminates. Aspectually, completive
adverbs are telic. The assumption then, is that they are compatible with
telic eventualities and odd with atelics. The examples below confirm this
assumption (Smith 1991):
(13) a. John drew a circle in five seconds.
b. Mary wrote a sonnet in ten minutes.
c. *Bill swam laps in an hour.
d. *Mary believed in ghosts in an hour.
Completives denote an interval within which the situation occurred
or took place and they are well formed with events. The atelic situations
in (13c,d) are ill-formed.
3.3. Frequency Adverbs

Frequency adverbs also give information that contributes to the
temporal location of a situation (Smith 1991). Specifically, they indicate the
recurrent pattern of situations within the reference interval. The adverbial
expression of frequency reinforces the notion of repetition, iteration:
(14) a. Samuel cycles to work most days/every day.
b. We always/often went to the mountains in wintertime.
Such sentences express a series of individual events, which as a
whole make a state of the habitual type. Examples of frequency
adverbials are frequently, on Sundays, never, sometimes, often, whenever,
monthly, daily, once a week, every week/month/year, usually, seldom, etc.
3.4. Locating Adverbs (or Frame Adverbs)

This type of adverbs contributes to the specification of RT and ET.
As the name ‘frame adverb’ indicates they refer to ‘an interval of
time within which the described action is asserted to have taken place’
(Bennett & Hall Partee, 1978). The situation talked about in the sentence
122
fills all or part of the time specified by the adverb (Smith, 1991). Just like
Tense, frame adverbs require an orientation point, and just like tense they
mirror the three possible temporal relations: simultaneity, anteriority and
posteriority. Frame adverbs have the role ‘to locate situations in time by
relating them to other times or to other situations’ (Smith, 1991).
According to the time of orientation they indicate we can distinguish
three subclasses:
(i) Deictic adverbs: which are oriented to the time of utterance.
Such adverbs are represented by the following expressions: now, today,
last Sunday, last week, this week / year, tomorrow, next week, the day
after tomorrow, tonight, a week ago, etc. As can be noticed, all adverbs in
this class refer to some specific time (span) that is related to some other
specific time, which is ST but most of them give only the ‘maximal
boundaries’ of the time span(s) in question (Klein, 1992).
(ii) Anaphoric adverbs include time expressions that ‘relate to a
previously established time’ (Smith, 1978) such as until, till, in the evening,
on Sunday, at night, early, before, in three days, on Christmas, at
lunchtime, two years later, in March, already, etc. In this case too, we have
to do only with the ‘maximal boundary’ of the time span in question.
(iii) Referential adverbs which refer to a time established by clock
or calendar (Smith, 1978), such as at six, August 19, in 1997, etc.
Time adverbs that are explicitly related to the time of utterance are
known as ‘anchored’ adverbs. Deictic adverbs are ‘anchored’ adverbs.
The last two classes are known as being ‘unanchored’ i.e. they are not
anchored to the utterance time and their interpretation is made possible by
an orientation point other than the time of utterance.
4.0. Temporal Values of the Main Tenses in English

4.1. General Properties of the English Simple Present Tense
The Present Tense is essentially a deictic tense (i.e., it is always
oriented to the moment of speech and the speaking ego). It enjoys both
psychological being at the present moment (Leech, 1971) and actual
being at now.
123
Informationally, a present tense sentence can not include the
endpoints of the situation. The present tense is incompatible with
perfectivity. In Reichenbachian terms, its ATR is [RT = ST, ET = RT].
Thus, present tense sentences, irrespective of the situation type
(state, activity or event), must be states (Smith, 1991).
4.1.1. From the very start it is important to notice a peculiarity of
the English present tense: predicates belonging to the accomplishment
and activity classes do not allow a continuous, ongoing / imperfective
interpretation of the present tense (Giorgi and Pianesi 1997). This is in
sharp contrast with the other Germanic languages or Romance languages.
Consider the following examples that contain a state, an
accomplishment, an activity and an achievement predicate in the simple
present tense in English and Romanian:
(15) John loves Mary. (state)
(16) John eats an apple. (accomplishment)
(17) John runs. (activity)
(18) *John finds a book. (achievement)
(15’) Ion o iubeşte pe Maria. (state)
(16’) Ion mănâncă un măr. (accomplishment)
(17’) Ion aleargă. (activity)
(18’) *Ion găseşte o carte. (achievement)
The state predications in (15, 15’) in the simple present tense mean
that a certain state holds of the subject at ST. There are no interpretative
differences between English and Romanian.
The achievement predications in (18, 18’) are ill formed in both
languages.
The accomplishment predications in (16, 16’) and the activity
predications in (17, 17’) receive sharply distinct interpretations in the two
languages under study. The English accomplishment predication in (16)
does not mean that John is presently engaged in an ongoing activity of
eating an apple as it does in Romanian (see the accomplishment
predication in (16’)). In English, it means that John is an apple-eater.
Again, the English activity predication in (17) does not mean that John is
presently engaged in the activity of running as it does in Romanian (see
124
the activity predication in (17’)). In English, it means that John is a
runner.
In order to obtain the continuous reading in English the progressive
must be used:
(19) John is eating an apple.
(20) John is running.
Interestingly, English is not different from other Germanic or
Romance languages as far as the other possible interpretations of
accomplishment and activity predications are concerned.
a. German, Italian, Romanian and other languages admit the
so-called instantaneous or reportive reading of accomplishments and
activities in the present tense. English too admits this interpretation of the
simple present tense. The sentences in (16, 16’) and (17, 17’) can also be
used as commentary on a picture or a movie or when uttered by a radio
commentator (see below for more details). In German, Italian, Romanian
and English the eventualities are described as perfective (i.e., they mean
that at ST a complete event of eating an apple or a complete action of
running have been performed) and their time is not directly related to the
ST of the speaker.
b. In English and other languages, present tense accomplishment
and activity predications can express habituality and genericity (see
below for more details):
(21) John eats an apple every day. (habitual reading)
(22) Dogs bark. (generic reading)
In what follows we try to find a principled explanation for the sharp
contrast between the English present tense and the present tense in other
languages as evinced by accomplishment sentences (16 and 16’) and
activity sentences (17 and 17’).
4.1.2. As already argued, the temporal interpretation of a sentence
involves the anchoring of the event denoted by the verb to the ST.
Intuitively, ST, functioning as an anchoring point for the sentence, can be
conceptualized as punctual. The punctuality of the ST amounts to its
being devoid of internal structure.
Giorgi and Pianesi (1997) state the following universal principle
valid for ST:
125
(23) The anchoring event ST is punctual.
In view of this universal principle that characterizes ST, let us
consider in more details the accomplishment and activity predicates.
It is known that accomplishment and activity predicates do have
internal structure, which should be compatible with the punctuality of the
ST (the internal structure of accomplishment and activity predicates could
include the punctual ST, which has no internal structure). The most
common test to detect whether an eventuality has internal structure (i.e.,
is a process or has a processual part) is ‘the progressive criterion’
(Mourelatos (1978), Galton (1984), Haeksema (1984)). All eventualities
that can occur in the progressive evolve in time and have internal
structure; those that do not occur in the progressive do not have an
internal structure. Consider:
(24) *John is loving Mary (state)
* Ion o iubeşte pe Maria acum (on the intended reading)
(25) John is running (activity)
Ion aleargă
(26) John is eating an apple (accomplishment)
Ion mănâncă un măr
(27) *John is winning a race (achievement)
*Ion câştigă o cursă
As can be noticed, in English (and Romanian) only activities and
accomplishments undergo the progressive test and prove to have internal
structure.
States have internal stages but they are temporally undifferentiated,
they are all the same and states can’t occur in the progressive. (In
Romanian the sentence is still ill formed but Romanian has no
morphologically marked progressive aspect; the incompatibility is
signaled by the presence of the deictic acum). Achievements have only
one internal stage and, in principle, they can’t occur in the progressive (in
neither language).
We conclude that only eventualities that are either processes or
contain a processual part have internal structure (and occur in the English
progressive).
126
The above facts show that Romanian eventualities that have
internal structure (activities, accomplishments) can be simultaneous with
ST (which has no internal structure/is punctual but is included in their
time span) and an imperfective / ongoing interpretation is obtained. In
English, the ongoing interpretation of activity and accomplishment
predicates is obtained only when these predicates occur in the present
progressive form.
4.3. On the [+Perfective] Feature of English Simple Present Tense

In this section we consider a possible explanation of the
impossibility of English simple present tense eventive verbs to have a
continuous/on-going interpretation (which contrasts with Romanian and
other Germanic languages).
Giorgi and Pianesi (1997) hypothesize that English associates the
aspectual value [+perfective] to all eventive predicates. The [+perfective]
feature on a verbal form entails perfectivity. In other words, all activity /
accomplishment predicates in English are closed / perfective
eventualities. If these predicates are closed they cannot be interpreted as
simultaneous with the punctual ST. This is not the case in Romanian (and
other Germanic and Romance languages) where eventive predicates are
not closed and hence can be mapped on the punctual ST.
The big difference between the English simple present tense and
the simple present tense in other languages lies in the morphological
properties of the English verb. Giorgi and Pianesi (1997) notice that
English verbs are well-formed words even without the addition of any
inflectional morphology (see also Roberts, 1993). For instance, a word
such as eat is ‘a naked’ form and can express one of the following verbal
values: it can be a short infinitive; it can be the 1st, the 2nd person singular;
it can be the 1st, the 2nd or the 3rd person plural. Many English words are
even categorially ambiguous in that they can either identify ‘an object’ or
‘an action’, such as dream, dance, dress, want, fall, etc. In Romanian and
other languages, verbs are always complex words, consisting of a lexical
morpheme plus inflection and the verbal inflection is distinct from the
nominal one (e.g., a visa / vis, a dansa / dans).
Giorgi and Pianesi (1997) argue that the [+perf] feature on the
English verb helps to identify the verb as a lexical category (i.e., verb).
127
Additional evidence in favour of the idea that English verbs denote
closed events comes from perceptual reports (Giorgi and Pianesi, 1997).
Consider:
(28) a. John saw Mary eat an apple.
b. John saw Mary eating an apple.
In English, perception verbs take as their complements either the
‘naked’ form (Acc + short infinitive) or the Acc + ing. The ‘naked’ form
allows only the perfective reading (we infer that Mary ate the whole
apple). The complement verb in (28b) refers to a non-closed event – we
do not infer that the apple was eventually eaten.
Note that in Romanian, the only corresponding form to the two
possibilities in (28) is with an embedded ‘gerunziu’, which invites to a
non-closed/ongoing interpretation:
(29) Ion a văzut-o pe Maria mâncând/*a mânca un măr
In sum, present tense English activity and accomplishment
predicates denote closed / perfective events that cannot be simultaneous
with the punctual ST. This property of the English eventive predicates is
captured by what Giorgi and Pianesi call the ‘Punctuality Constraint’:
(30) A closed event cannot be simultaneous with a punctual event
4.4. Other Values/Uses of the Simple Present Tense

4.4.1. Generic and Habitual Sentences in the Simple Present Tense
As mentioned above, in English and other languages, present tense
accomplishment and activity predications can express habituality and
genericity:
(31) John eats an apple every day. (habitual reading)
(32) John goes to the cinema twice a week.
(33) John smocks cigars. (generic reading)
(34) Dogs bark.
Habitual sentences explicitly contain a frequency adverb that
conveys the meaning of repetitivity as in (31) and (32). Generic sentences
do not occur with frequency adverbs but implicitly contain the adverb
always as in (33) and (34).
128
Both habitual sentences and generic sentences predicate of states of
affairs that hold true at the moment of speech and aspectually they are
states.
The question that arises now is why present tense activities and
accomplishments, which we have argued, are perfective/closed can be
states in habituals and generics.
As a matter of fact, the acceptability of habituals and generics in the
present tense is due to the fact that the conflict between the punctuality of
the ST and the closure of the event denoted by the predicate does not
arise. This is because habituals and generics are understood as asserting
the occurrence of a series of events that includes the ST. According to
Giorgi and Pianesi (1997), a habitual/generic sentence only requires that
ST be a temporal part of the interval where the habitual/generic sentence
holds. Thus, habitual/generic predicates denote closed/perfective events,
which by means of iterativity, along the present axis of orientation,
become open informationally and are interpreted as states.
4.4.2. The Instantaneous Use of the Simple Present Tense

This use of the present tense portrays an event that looks simultaneous
with the present moment now although the event (which has intrinsic
duration) is treated as perfective. More often than not, this use of the present
tense acquires a dramatic value and its duration is ‘telescoped’ to a point. The
instantaneous use of the simple present tense occurs only with events (not
with states) in certain easily definable contexts:
---in sports commentaries on the radio where the commentator is
reporting something that the listeners cannot see:
(35) Napier takes the ball and runs down the wing. He passes the
ball to Attwater. Attwater beats two men, he shoots. It’s a goal!
---it is used in demonstrations, when the audience can see what is
happening. Once again, the demonstrator is reporting the activity
performed as perfected at the moment now. The use of the simple present
tense is the only appropriate form in this context:
(36) Look, I take this cart from the pack and place it under the
handkerchief – so!
---it is used in stage directions:
129
(37) He yields. The specter vanishes!
---it is also used with performative verbs: in this case, the event
announced and the act of announcement (ST) are the same:
(38) I beg your pardon / I accept your offer / I promise to be here at
5 / I declare war / I pronounce you man and wife / I declare
the meeting open
4.4.3. The Historical Value of the Simple Present Tense

The simple present tense may also be used with reference to the
past. This use is best treated as a story-teller’s licence, whereby past
happenings are imagined as if they were going on at the present time. Its
contexts of occurrence are:
---it is typical of a highly colored popular style of oral narrative and
can be accompanied, with apparent incongruity, by an adverbial
expression indicating past time:
(39) Last week I am in the sitting room with my wife, when this
chap next door staggers past and in a drunken fit throws a
brick through the window.
---the simple present tense can also be used in fictional prose where
we expect the use of the past tense to describe imaginary happenings.
With some writers, transposition into the fictional present is a device of
dramatic heightening; it puts the reader in the place of someone actually
witnessing the events as they are described:
(40) Mr. Tulkinghorn takes out his papers, asks permission to place
them on a golden talisman of a table at my Lady’s elbow,
puts on his spectacles, and begins to read by the light of a
shaded lamp. (Dickens, Bleak House)
---a different kind of historic present is found with verbs of
communication:
(41) a. Joan tells me you are getting a new car.
b. The ten o’clock news says that it’s going to be cold.
c. I hear poor old Mrs. Baxter has lost her cat.
130
The verbs tell, say and hear here refer to the initiation of a message
in the past; therefore, we expect the use of the past tense or of the present
perfect. However, it appears that the verbal meaning has been transferred
from the initiating end to the receiving end of the message.
---the following sentences illustrate a similar extension of the
present tense to cover information, which in strict historical terms belongs
to the past:
(42) a. In “The Brothers Karamozov”, Dostoevsky draws his
characters from the sources deep in the Russian soil.
b. Like Rubens, Watteau is able to convey an impression of
warm, living flesh by merest whiff of color.
When we discuss about an artist and his surviving work we are
justified in using the present tense because his work, and in a sense the
artist himself are still ‘alive’.
Thus, the only difference between Brahms IS the last great
representative of German classicism and Brahms WAS the last great
representative of German classicism is a difference of point of view:
whether one prefers to think of Brahms as a composer still living through
his compositions or as a man who died in the 19th century.
However, we do not have this choice when dealing with purely
biographical details of an artist’s life: the present tense cannot be
substituted for the past tense in Brahms WAS BORN in Hamburg /
Brahms COMPLETED his first symphony in 1876.
5.0. The Temporal Values / Uses of the Simple Past Tense

5.1. The Simple Past Tense with Deictic Value
The simple past tense has a basic time association with a past
moment of time, rendered by adverbs such as then, yesterday, two hours
ago. The past tense is deictically interpreted (i.e., with respect to the
moment of speaking now).
Other events that occur with the past tense can be simultaneous
with another past event as in (43), can be anterior to another past event as
in (44) or posterior to another past event as in (45) (c.f. Baciu, 2005):
131
(43) I bought this statue when I was in Venice.
(44) I misplaced my pencil a moment ago and I couldn’t find it.
(45) I signed and sealed the envelope.
The simple past tense describes a situation that occurred before the
present moment at a moment in the past understood as definite / specific;
this definite past moment is usually indicated by the adverb. The past
time adverbs most frequently associated with the past tense are
yesterday/last month/night/ year, the other day, once, two days ago, in the
year 2000, in June, when I was a child, etc.
The ATR of the past tense is [RT < ST, ET = RT].
The ART of the past tense explains why the eventuality is portrayed
in its entirety – as including its initial and final bounds (i.e., perfective
aspect). The described event is viewed in its entirety, because RT
coincides with the ET, from its initial to its final boundary.
There are, however, contexts in which the time adverbial
specification can be missing. This is possible only when the adverbial can
be inferred and retrieved from the larger context:
(46) Ann: This time last year I was in Vienna.
Bill: How curious! I was there too.
Bill’s answer is correct without a past tense adverbial because the
missing adverb can be equated with the adverb mentioned in the
preceding sentence (i.e., this time last year).
Another case in which a simple past tense sentence can occur
without a definite adverb involves sentences like the following:
(47) a. Joan has received a proposal of marriage. It took us
completely by surprise.
b. I have seen him already. He came to borrow a hammer.
In such contexts, the present perfect is used to introduce an event
that took place sometime before the moment of speech; once an anterior
frame of reference is established it is natural to resume reference to the
already introduced event by the simple past tense, which is thus uniquely
identified.
Thirdly, the simple past tense can be used without a definite
specification when a comparison is drawn between present and past
conditions (paraphrased by the phrase ‘used to’):
132
(48) England is not what it was (what it used to be) / Even dogs are
not what they were (what they used to be) / Life is not so
pleasant as it was (what it used to be)
5.2. The Non-Deictic Use of the Simple Past Tense

Beside its deictic usage, the simple past tense can also be used
non-deictically. This use is specific to the narrative mode and the simple
past tense occurs without a temporal adverb. The situations narrated
happened before the moment of speech but this moment is not given and
has to be identified as part of the information associated with the way
narrative functions. Here are two examples that constitute the opening
paragraphs of J. Joyce’s “Eveline” and W. Golding’s “Lord of the Flies”:
(49) She sat at the window watching the evening invade the avenue.
Her head was leaning against the window curtains, and in her
nostrils was the color of dusty cretonne. She was tired.
(50) The boy with fair hair lowered himself down the last few feet
of the rock and began to pick his way toward the lagoon.
Linguists have also identified other uses of the simple past tense:
a) the habitual use; b) the past perfect use; c) the present time use.
5.3. The Habitual Value of the Simple Past Tense

In a habitual sentence such as (51), the frequency adverb at noon
every day specifies the repeated ET of the predication. The adverbial
during his childhood specifies RT, the past interval during which the
recurring event took place:
(51) John got up at noon every day during his childhood.
In the case of habitual sentences, the determiner of the frequency
adverbial must be indefinite:
(52) They went to the movies three times a week.
*They went to the movies three times the week.
The frequency adverbial can be missing in case the object in the
plural, as indicated in (53), conveys the habitual reading:
133
(53) a. Fido chased cars. (habitual reading)
b. Fido chased a car / Fido chased the car. (non-habitual reading)
5.4. The Simple Past Tense with Past Perfect Value

Consider the sentences below:
(54) a. He enjoyed and admired the sonnets of Shakespeare.
b. He knocked and entered.
c. He shaved and listened to the radio.
In (54a) we have the description of state predicates: states denote
duration hence the sentence is understood to describe two simultaneous
states. On the other hand, the sentences in (54b,c) describe two events that
can be performed only sequentially (as a rule, one first knocks and then
enters). The event that is interpreted as taking place before another event
in the past has a past perfect value – we have to do with a shifted reading
of the simple past tense in the case of events.
Notice that temporal relations between two consecutive events can
sometimes be explicitly marked either by an adverbial or conjunction.
Consider the following examples:
(55) I thought of him very much AFTER I went to bed.
AS SOON AS he left us, Sir William very politely stepped up
to our new niece.
He dropped the letter BEFORE he went away.
In the above sentences, a formal morphological marker (after, as
soon as, before) indicates whether the event italicized occurred after or
prior to the event of the main clause.
5.5. The Simple Past Tense Referring to Present Time

In everyday conversation the simple past tense can be used with
present time reference with no adverbial specification. Consider:
(56) A: Did you want me?
B: Yes, I hoped you would give me a hand with the painting.
The use of the simple past tense with present time reference occurs
mainly in questions, which are thus perceived as more polite and less
pressing than the one in the simple present tense. Leech (1971:11) makes
134
the following comments on sentences in (56): “The subject of this
exchange would probably be the present wishes of speaker B, despite the
use of the past tense. The Present and the Past are, in fact, broadly
interchangeable in this context; but there is quite an important difference
in tone. The effect of the past tense is to make the request indirect, and
therefore more polite….The present tense (I hope…) in this situation
would seem rather brusque and demanding – it would make the request
difficult to refuse without impoliteness. The past tense, on the other hand,
avoids the confrontation of wills. Politeness also extends to the original
question Did you want me? The logically expected tense (Do you want
me?) might have peremptory overtones, and would seem to say ‘Oh, it’s
you, is it? You always want something! Other verbs similarly used are
wonder and think:
(57) a. I wondered if you’d look after my dog while I go shopping.
b. I thought I might come and see you later this evening.
As before, the speaker is purportedly testing the listener’s reaction
to a past attitude to which he confesses, whilst pretending that his present
attitude is undermined”.
6.0. The Temporal Values / Uses of the Simple Present Perfect Tense
6.1. General Properties of the English Present Perfect Tense
The present perfect, as distinct from the simple past, is often
described as referring to ‘past with present relevance’, or ‘past involving
the present’. Consider the contrast below:
(58) a. John read the book last year.
b. John has already read the book.
Both sentences involve reference to events that occurred prior to the
moment of speech. The sentences in (58) show that past events can be
predicated about either in the past tense or the present perfect but from
two different perspectives. In (58a) we understand that John’s reading the
book in its entirety (a perfective eventuality) is dated / is specified as
occurring during last year, which is prior and thus distinct from the
moment now. In contrast, from (58b) we understand that John’s reading
the book in its entirety occurred at some unspecified time in the past but
135
John’s reading the book is related, and thus relevant to the present
moment through its result: now, John knows what the book is about.
One of the widely accepted contrastive definitions of the past tense
vs. present perfect is that proposed by McCoard (1978): “the preterit tense
represents an action or state as having occurred or having existed at a past
moment or during a past period of time that is definitely separated from
the actual present moment of speaking; the present perfect is analyzed as
marker of prior events which are nevertheless included within the overall
period of the present”.
Due to the extra dimension of meaning (i.e., the relation of the
present perfect to the speech time) we say that the present perfect is not a
perfective tense but a perfect tense. However, both perfective and perfect
tenses describe completed eventualities; more technically, they are both
topologically closed eventualities (Giorgi and Pianesi, 1997). The
perfective past tense is topologically closed before the moment of speech
while the perfect present perfect is topologically closed at the moment of
speech.
The present perfect, as its name shows, is a paradoxical tense: it is
half past (i.e., the unspecified ET is situated before ST), half present (i.e.,
RT is situated at ST, as indicated by the present tense marker on have).
Thus its ATR is [ST = RT, ET < RT].
The perfect aspect has been treated in the linguistic literature as an
aspectual operator (it changes the aspectual properties of the eventuality
it operates on) similar to the progressive operator. Different from the
progressive operator, which selects only for non-stative eventualities, the
perfect operator can occur with and thus operate on all types of
eventualities (e.g., He has been here, He has sung, He has built a house).
The present perfect describes the result(ant) state of an eventuality. An
event in the present perfect as in (58b) occurred at an unspecified ET in
the past, and the perfect aspect specifies the state that results at ST after
the culmination of the event. Thus, sentences in the present perfect are
derived state sentences, different from past tense sentences, which inherit
the aspectual properties of the main predicate (state, process or event).
In spite of the fact that both the past tense and the present perfect
express relations of anteriority they evince highly restrictive
combinatorial abilities when they occur with temporal adverbs (McCoard,
1978).
136
In essence, the past tense occurs with specific / definite time
adverbs (e.g., two months ago, last weekend, yesterday, after the war, at
4, on Tuesday) while the present perfect occurs with non-specific /
non-definite time adverbs (e.g., since 3 o’clock, for 2 hours, so far, yet,
lately, before now).
The English present perfect contrasts with its equivalents in the
large majority of languages in that it resists occurrence with specific
temporal adverbs.
Compare: *John has come at 4 / yesterday / on Tuesday / before the
war with Ion a venit la 4 / ieri / marţi / înainte de război. This inability of
the present perfect to occur with definite adverbs is known as the ‘Present
Perfect Puzzle’ (Klein, 1992).
The intriguing properties of the present perfect have received many
kinds of explanations. Out of the most known theories on the semantics
and pragmatics of the present perfect we briefly present The Indefinite
Past Theory, The Current Relevance Theory and The Extended Now
Theory.
The Indefinite Past Theory focuses on the indefiniteness of the ET
of the present perfect eventuality, the Current Relevance Theory
underlines the current relevance of the resultant state at ST of the present
perfect eventuality while the Extended Now Theory focuses on the idea
of an ’extended’ present moment.
6.2. The Indefinite Past Theory on the Present Perfect

The crucial element of the Indefinite Past Theory (ID) of the
present perfect is the claim that this tense locates eventualities somewhere
before the moment of speaking, without identifying any particular point
or interval of time. The ET of the perfect is thus indefinite – any past
interval I can serve as ET. In contrast, remember that the ET of the past
tense is definite.
For instance, consider the difference between sentences (59) and
(60) below:
(59) “I have been to Carnegie Hall only once”.
“Did you hear the New York Philharmonic?”
(that is, on the same occasion)
137
(60) “Have you heard the New York Philharmonic?”
(that is, ever in your life, not necessarily on the same occasion)
Leech (1971) brings in one more factor in the selection of the past
tense, viz., the shared knowledge between speakers is rendered by using
the past tense not the present perfect:
(61) Did you put the cat out? (said between husband and wife who
have in mind a particular time when the cat is normally put
out)
McCoard (1978) and Klein (1992) argued that the present perfect is
an indefinite tense in the same sense as Leech did: the exact position of its
ET is not specified; it is left open. In Chris has been in Pontefract,
“Chris’s being in Pontefract could have been yesterday or twenty years
ago, it could have lasted for three years or for one day, he may have been
there once or fifty times” (Klein 1992: 539).
Indeed, a cursory inspection of the adverbs compatible with the
present perfect (since 3 o’clock, for 2 hours, so far, yet, lately, before
now, long since, in the past, once (one time), today, in my life, for three
years, recently, yet, often, always, ever, never, already) shows that they
are adverbs unable to establish ET.
However pertinent in its conclusions, ID stipulates only pragmatic
conditions under which the choice between the past tense and the present
perfect is made.
6.3. The Current Relevance Theory on the Present Perfect

The Current Relevance Theory (CR), which originated with
Jespersen (1931), claims that it is only the present perfect that purports
current relevance at the moment now, a feature that the past tense lacks.
Consider the contrast between (62) and (63):
(62) You woke him up when you went to the bathroom ten minutes
ago.
(63) You’ve waken him up when you went to the bathroom ten
minutes ago.
Sentence (62) would most probably imply that he is awake now,
but this inference is ensured by our knowledge of the world and is not
138
rendered by the verb form itself. When sentence (62) is compared with
sentence (63), in the latter it is the verb form itself that locates the effects
of the event at the moment of speech. If the implication that he is awake
now is contradicted or negated by other factors of context (such as
evidence that the state of being awake no longer holds at the moment of
speech) the use of the present perfect is ungrammatical (cf. Fenn, 1987).
In the same line, K. Inoue (1979) defines the notion of discourse
topic in order to account for the difference in acceptability between past
tense sentences and present perfect sentences. Discourse topic is
understood to be the subject matter under discussion in a certain context.
Consider the following contrast:
(64) a. *Einstein has visited Princeton.
b. Einstein visited Princeton.
c. Princeton has been visited by Einstein.
The well-formed sentence in (64b) can be uttered in a context about
Einstein and his activities in general. But since Einstein is no longer
living, he can neither visit Princeton nor engage in other activities. As
such, the speaker cannot use the present perfect and only the past tense is
appropriate. On the other hand, sentence (64c) is pragmatically felicitous
in case the discourse topic is Princeton University: its having been visited
by a renowned scientist such as Einstein represents a memorable occasion
in the history of Princeton University.
We can easily recognize CR theory and its ‘discourse topic’ variant
as defining pragmatic conditions on the use of the present perfect.
6.4. The Extended Now Theory on the Present Perfect

The Extended Now Theory (XN), set up by McCoard (1978), is
perhaps the most popular theory on the semantics of the present perfect.
McCoard appeals to the idea of an ’extended’ present moment:
speakers can psychologically ‘extend’ the present backwards.
The present perfect represents the means of the English tense
system to render this backward extension. “From the point of view of the
present the speaker looks back upon some continuous stretch of the past
and within this he places the action or state. This period of past time may
be momentary as in ‘The massager has just arrived’ or it may be of
considerable extent as in ‘The old house has been left untenanted for
139
many years’ or it may include all past time, as in Shakespeare’s ‘Men
have died from time to time and worms have eaten them, but not for love”
(c.f. Ştefănescu, 1988). Thus, the present perfect amounts to meaning that
a past event ‘constitutes’ a present state in some way.
The present perfect serves to locate an event within a period of time
that begins in the past and extends up to the present moment (and
includes it); in contrast, the past tense specifies that an event occurred at a
past time that is separated and distinct from the present.
McCoard’s XN theory is a semantic solution to the problem of the
present perfect in that it characterizes it in terms of RT (which is the same
as ST). Its ATR is [ST = RT, ET < RT].
6.5. The Values / Uses of the Simple Present Perfect

Linguists speak of four main uses of the present perfect: the
experiential present perfect, the continuative present perfect, the
resultative present perfect and the ‘hot news’ present perfect.
In current studies the experiential present perfect is renamed
Existential Present Perfect while the continuative present perfect, the
resultative present perfect and the ‘hot news’ present perfect all collapse
under the Resultative or Continuative Present Perfect (Demirdache and
Uribe-Etxtebarria 2002, 2004). We adopt this terminology.
The two main values of the present perfect crucially depend on the
aspectual properties of the eventuality (cf. Comrie 1976, Smith 1991,
Kamp and Reyle 1993, Julien 2001, Demirdache and Uribe-Etxebarria
2002, 2004 among many others).
6.5.1. The Existential Value of the Present Perfect

(65)a. Sam has broken my computer (twice) (accomplishment)
b. John has reached the top (twice) (achievement)
c. Susan has played the violin (twice) (activity)
d. Mary has lived in Cairo for three years (twice in her life) (state)
As can be easily noticed, in the present perfect sentences in (65) the
eventuality is presented as ‘bounded’ since it can be repeated. We say that
the present perfect has an existential value: it shows the existence of one
140
or several eventualities (states, processes or events) that are presented as
completed prior to the moment of speech (Demirdache and Uribe-
Etxebarria, 2002). In (65d), the adverb for three years specifies the
duration of two past eventualities (i.e., twice in her life Mary lived in
Cairo for a period of three years). All the predications in (65) have current
relevance at ST and they are interpreted as states.
6.5.2. The Resultative or Continuative Value of the Present Perfect

Consider the examples below:
(66) a. Oh! My God! Sam has broken my computer.
(accomplishment)
b. John has reached the top. (achievement)
c. Susan has played the violin for two hours. (now) (activity)
d. Mary has lived in Cairo for three years. (now) (state)
In the sentences in (66), the present perfect induces a resultative or
continuative reading: it presents a state as holding from a moment in the
past up to and including the moment of speech. This value of the present
perfect focuses on the resultant state of an eventuality.
Notice that the resultative reading obtains with predicates that
describe an accomplishment (as in 66a) or an achievement (as in 66b).
Thus, in (66a,b) the result state described by the sentence SAM BREAK
MY COMPUTER or the sentence JOHN REACH THE TOP are
presented as persistent at the moment of utterance. In other words, (66a)
implies (the result) that my computer is broken at the moment of speech
and (66b) implies that John has reached the top and is on the top of the
mountain at the moment of speech. The result state is presented as still
persisting at ST: it continues from a past interval up to the moment of
speech.
The present perfect in (66d) indicates that Mary still lives in Cairo
at the moment of speech. The adverb of duration shows that the state in
question began three years before the moment of speech and this state still
continues at the moment of utterance. The presence of the duration
adverb for three years is obligatory and it measures the whole eventuality.
Notice that the continuative reading obtains with a predicate that
describes a state or an activity.
141
6.5.3. The ‘Hot News’ Value of the Present Perfect
The third value of the present perfect is the ‘hot news’ present
perfect, dubbed as such by McCawley (1971). It is illustrated in (67). In
present day studies on the present perfect this value is not identified as a
separate value of the present perfect as it can be easily identified with the
resultative value of the present perfect with an event predicate. The
notion of recent past with a resultative state holding at the moment of
speech is conveyed by the use of just:
(67) Malcolm Jones has just been assassinated!
7.0. Other Temporal Uses of the Present Perfect

7.1. The Future Value of the Present Perfect
In adverbial clauses of time the present perfect is used with a future
value. Consider the sentence below:
(68) You can go when you have finished your work.
The conjunctions commonly used to introduce the adverbial clauses
of time are when, as soon as, before, after, until, once, by the time (that),
the moment (that).
In some contexts, the use of the present perfect is in free variation
with the present tense:
(69) I shall leave as soon as the meeting ends / has ended.
In other contexts, the choice between the two tenses is not free:
--– when the events in the main clause and the subordinate clause
temporally coincide the use of the present tense in the subordinate clause
is favored, as in (70a) below; when the event in the subordinate clause
occurs after the one in the main clause the use of the present perfect in the
subordinate clause gives well formed sentences, as in (70b) below:
(70) a. Come over and see us when our guests leave.
b. Come over and see us when our guests have left.
--– when a causal relation between the event in the main clause and
that of the subordinate clause is established, the use of the present perfect
is favored in the subordinate adverbial clause of time:
142
(71) a. We can go out as soon as we have had dinner.
b. *We can go out as soon as we have dinner.
8.0. Temporal Adverbs with the Present Perfect and the Past Tense
The temporal contrast between the present perfect and the past tense
would not be complete unless we examine in detail the strong restrictions of
the two tenses in terms of their occurrence with temporal adverbs.
A major contribution of McCoard’s study (1978) is the detailed
analysis of the way in which temporal adverbs relate to the present perfect
and/or past tense. Adverbs bring in their temporal meaning and they bear
on tense selection and even on tense interpretation.
McCoard identifies three classes of adverbs: those that occur with
the simple past tense but not with the perfect, those that occur with either
the simple past or with the perfect and those that occur with the perfect
but not with the simple past.
Occur with the Occur with either the Occur with perfect
simple past but not simple past or the but not with simple
with the perfect perfect past
long ago long since at present
five years ago in the past up till now
once (= formerly) once (= one time) so far
the other day today as yet
those days in my life during these five years
last night for three years herewith
in 1900 recently lately
at 3:00 just now since the war
after the war often before now
no longer yet
always
ever
never
already
before
143
The adverbs in the first column refer to points or stretches of time
that precede the moment of speech, either by their semantics or by
context (e.g., at 3:00). The adverbs in the third column coincide or are
oriented to the moment of speech. In context, these adverbs can be
thought of as beginning before the moment of speech and extending
beyond it. They only occur with the present perfect and exclude the past
tense. For the adverbs in column two, it is the context and in particular the
tense used which decide which time-sphere (past or present) is actually
being referred to. They are ‘neutral’ time-span adverbs (Fenn, 1987).
As far as the adverbs in column two are concerned, the following
comments are in order.
The duration adverb for-phrase (e.g., for three years) measures the
duration of a homogeneous eventuality (state or activity, never an event)
whether it is in the past tense or the present perfect (e.g., He sang for a
couple of minutes / I have known him for years / *John has built the
house for two years / *John built the house for two years).
In contrast with for-phrase, since-phrase “marks the beginning of
the period and the moment of utterance marks the end” (Heinamaki,
1978) and it only occurs with the present perfect with either states or
events. The inclusive since-phrase together with the eventuality in the
present perfect relates the predication to the moment of speech (e.g., John
has been here since 3 o’clock / John has worked in the garden since
morning / Since last summer John has moved to Paris).
Ever and never are used when the life experience of the subject is
predicated about. Both suggest the meaning ‘within a period of time’.
When they occur with the present perfect it is the present perfect that
relates their time-span to the moment of speech (e.g., A saner and more
practical man I’ve never met). On the other hand, their ‘within a period of
time’ meaning also makes them compatible with the past tense (e.g., I
never saw the St. Patrick’s Day Parade while I was in New York).
Adverbs such as often, always, sometimes, which refer to frequency
can, depending on the context, occur with either the present perfect or
with the past tense (e.g., I have always suspected your honesty / He
always made a lot of fuss about nothing when they were married).
Lately and recently are commonly regarded as synonyms but they
show different compatibility as to their occurrence with the past tense and
144
the present perfect. Lately accepts only the present perfect (e.g., I have
spent a great deal of money lately / *I spent a great deal of money lately)
while recently goes with both the past tense and the present perfect (e.g., I
have been ill recently / I was ill recently).
Adverbs such as today, this week, this year can occur with both the
present perfect and the past tense (e.g., I have seen John this morning / I
saw John this morning). Both sentences convey the meaning that “the act
occurred within the time span this morning. The difference lies in whether
the event is viewed simply as a factor of experience obtaining at the
moment of speech [with the present perfect] (i.e., the morning time-span
is not over) or whether it is viewed within the context of the time at which
it occurred [with the past tense] (i.e., the morning time-span is over)“
(Fenn, 1987).
The difference in uses between adverbs such as just and just now is
the following. Just can take either the present perfect or the past tense
(e.g., I have just seen your sister / I just saw your sister) while just now,
which is interpreted as a moment/second/minute ago, can only occur with
the past tense (e.g., I saw your sister just now).
Finally, there are adverbs that combine with either the present
perfect or the past tense but with a clear difference in meaning. Now is
mainly associated with present tenses: Now my ambition is fulfilled / has
been fulfilled. With past tense, it is a narrative substitute for then (= ‘at
this point in the story’): Now my ambition was fulfilled.
Once, with the meaning ‘on a certain occasion, at one time’ occurs
with the past tense, despite its indefinite meaning: He was once an honest
man. In a present perfect sentence it has a numerical value contrasting
with twice, three times, etc: I have visited the Highlands only once
(Leech, 1971).
Already, still, yet and before occur with the present perfect in the
sense ‘as early as now’, ‘as late as now’: I have seen him already / I (still)
haven’t seen him (yet). With the past tense they must have a meaning
involving a past point of orientation: I was already (= ‘as early as then’)
very hungry (Leech, 1971).
145
9.0. The Temporal Values / Uses of the Past Perfect
9.1. General Properties of the Past Perfect Tense in Simple Sentences
Consider the following sentence in the past perfect where the event
of Mary’s leaving the school is viewed as completed before a past
reference time:
(72) Mary had left school.
The ATR of the past perfect is [RT < ST, ET < RT]. Since ST
follows RT, which itself follows the ET, the event of Mary’s leaving the
school is viewed as completed before a past reference time.
Notice that because RT and ET denote two disjoint intervals, if we
add a temporal adverb such as at 5 to the sentence in (72), as in (73), the
sentence will have two distinct readings depending on whether the time
adverb modifies the ET or the RT:
(73) Mary had left school at 5.
If the time adverb at 5 modifies ET we understand that Mary’s
leaving the school occurred at 5 o’clock (this is the so-called event time
reading of the past perfect sentence).
If the time adverb at 5 modifies RT we understand that Mary’s
leaving the school occurred prior to the RT and the RT coincides with the
time denoted by 5 p.m. (this is the so-called reference time reading of the
sentence).
9.2. General Properties of the Past Perfect Tense in Complex

Sentences
In complex sentences, the main clause can establish the past RT of
the subordinate past perfect clause, as in the example below:
(74) a. They told us yesterday that Tom had arrived 3 days earlier.
b. *Tom had arrived 3 days earlier.
Sentence (74a) is well formed because the adverb yesterday in the
main clause also establishes the RT of the embedded clause: we
understand that Tom’s arrival occurred 3 days prior to yesterday. The
embedded clause specifies a time other than RT, namely ET. Thus, while
146
the adverb in the main clause specifies RT for both clauses, the adverb in
the embedded clause specifies only its ET (and its RT is shared with that of
the main clause). Notice that sentence (74b) is ungrammatical because it
contains an adverbial and a tense marker that together cannot establish RT.
The past perfect can be used in main clauses (as in 73 above) and in
subordinate clauses. It occurs mainly in subordinate adverbial clauses
introduced by the conjunctions when, after, before, till. Consider some
examples:
(75) a. He would not allow anyone to leave the inn till he had drunk
himself asleep.
b. As two and a half years had elapsed since he had made any
money Spencer returned to London.
c. When his mind had been weaker his heart led him to speak
out.
In subordinate adverbial clauses of time introduced by an explicit
conjunction, the past perfect can be substituted by the simple past tense:
(76) a. When he came back from India, he was made a member of
Parliament
b. As soon as he discovered them, he ran away
c. I ate my lunch after my wife came back
However, there are cases when the past perfect cannot be
substitutable by the past tense, because a marker of anteriority of the
event in the subordinate clause is necessary for the correct interpretation
of the whole sentence:
(77) a. When he had read the letter, he burned it.
b. *When he read the letter, he burned it.
10.0. Means of Expressing Future Time

10.1. General Characteristics
The auxiliary verbs shall and will used to convey future tense in
Modern English were, in older stages of English, full verbs. Shall meant
‘I must, I am under the obligation’ and will meant ‘I want’. ‘Obligation’
and ‘want’ are oriented to a future time sphere. In the course of time these
147
verbs developed into auxiliary verbs that signaled both modality (i.e.,
possibility / probability) and futurity. In Modern English they retain both
a temporal and a modal value but many other linguistic means are being
used to render future time more explicitly. There are also some
differences and nuances of usage that distinguish among these means.
The ATR of the future tense is [ST < RT, ET = RT] irrespective of
the linguistic means that render it.
Leech (1971) lists the following linguistic means that express
futurity in Modern English; these means belong either to the modal
system or to the aspectual paradigm.
a. The present tense + future time adverbs:
e.g., Exams start tomorrow.
b. The present progressive: e.g., The parcel is arriving tomorrow.
c. Will/shall + infinitive: e.g., The parcel will arrive tomorrow.
d. Will/shall + progressive infinitive:
e.g., The parcel will be arriving tomorrow.
e. Be going to + infinitive:
e.g., The parcel is going to arrive tomorrow.
10.2. The Simple Present Tense with Future Time Adverbs

(78) a. Tomorrow is Thursday.
b. I leave early tomorrow morning.
c. I am off tonight.
The future time meaning of the simple present tense refers to a
future occasion: at the moment of speech the speaker anticipates an event
or a state that is to take place at a RT = ET that is after/posterior to ST.
The present tense with future time adverbs is used in situations when the
event is scheduled by external factors, i.e., there is a decision taken or
plan fixed according to some external authority.
10.3. The Present Progressive with Future Time Adverbs

(79) I’m starting work tomorrow.
148
The present tense progressive plus future time adverbs is used when an
element of human volition is involved and it has the flavor of a planned,
arranged action that takes place in the future. The sentence in (80b) below is
ill formed because no conscious, human agency is involved:
(80) a. John is rising at 5 tomorrow.
b. *The sun is rising at 5 tomorrow.
Present progressive sentences with future time adverbs also convey
a sense of imminence that is absent from the use of the simple present
tense with future time adverbs:
(81) a. The Smiths are leaving tomorrow.
b. My aunt is coming to stay with us this Christmas.
Palmer (1978) contrasts the use of the simple present tense form
with the present progressive form with future time adverbs in the
following terms. Consider first the sentences:
(82) a. I’m starting work tomorrow.
b. I start work tomorrow.
“The first sentence suggests that the speaker now expects or intends
to start work – he may perhaps, have been ill. The second indicates that
tomorrow is the time fixed for him to start, e.g., by his firm or by the
doctor”.
10.4. Will and Shall plus the Infinitive

Traditional grammars have interpreted modal auxiliaries will and
shall as means of expressing future tense. In fact, the contribution of these
modal verbs in sentences as in (83) below (without temporal adverbs) is
modal, i.e., that of making predictions:
(83) a. Allan will be in Bucharest.
b. Tomorrow’s weather will be cold and windy.
c. You will feel better after you take this medicine.
It is evident that will/shall also refer to future predictions due to
their modal nuance. As said above, the mixture of modal and temporal
values of these modal verbs is due to the diachronic development of
149
English: at the beginning will/shall had only modal values and in time
they also developed a future reading when they occur with future time
adverbs. Leech (1971) makes the following comments with respect to
their usage: “frequently a sentence with will/shall is incomplete without
an adverbial of definite time: *It will rain / *The room will be cleaned.
These sentences are relatively unacceptable on their own, presumably
because of their factual emptiness: we all feel certain that ‘it will rain’ at
some time in the future, so there is no point in saying ‘it will rain’ unless
an actual time can be forecast”.
10.5. Will and Shall plus the Progressive Infinitive

(84) a. This time next week I shall be sailing across the Atlantic.
b. Don’t call me at 9 – I’ll be eating my supper.
In the first sentence in (84a) the verb is in the progressive form and
the modal shall contributes its (modal) predictive sense. Therefore, the
sentence predicts that this time next week the activity of sailing across the
Atlantic will be in progress. The prediction is made with such a great
degree of certainty that the event is presented as unfolding at a specified
future reference time. The same explanation can be considered for the
second sentence. Thus, the future interpretation of the sentences does not
result from will/shall plus progressive infinitive, but from the adverbial
specification in the sentences.
10.6. Be Going To
(85) I’m going to call him.
Be going to is a frozen form that cannot be analyzed into two
separate verb forms: it is listed as such in the lexicon. Jespersen (1931)
remarks that the structure is going to derives from the progressive form of
the verb to go: “going loses its meaning as a verb of movement and
becomes an empty grammatical word”. The same process occurred in
French with the form je vais faire. In contemporary English, be going to
is mainly used in colloquial speech.
150
The basic meaning of be going to is that of “future fulfillment of the
present” (Leech, 1971). Leech (1971) identifies two extensions of this
general meaning of to be going to:
---the first one is ‘the future fulfillment of the present intention’ that
is found with human subjects who consciously exercise their will:
(86) What are you going to do today? I am going to stay at home
and watch television.
On this reading the sentence I am going to watch television is felt as
stronger than I intend to watch television.
---the second extension of the general sense of be going to can be
stated as ‘future fulfillment of present cause’. This sense is common with
both agentive and non-agentive verbs:
(87) a. She is going to have another baby.
(i.e., she is already pregnant)
b. I think I’m going to faint. (i.e., I already feel ill)
c. There’s going to be a storm in a minute.
(i.e., I can see the black clouds gathering)
Notice that be going to can also be used when speaking about
periods remote from ST:
(88) a. I’m going to be a policeman when I grow up.
b. If Winterbottom’s calculations are correct, this planet
is going to burn itself out 200,000,000 years from now.
10.7. The Future of Past Situations

In case the sentence has a past time sphere, all the future time
expressions are modified to indicate a future plus past situation (i.e.,
future in the past):
(89) a. He was leaving town the day after we arrived.
b. He was going to be a policeman later in his life.
Palmer (1979:130) remarks that “for future in the past, be going to
is regularly used”, while in literary style would is likely to occur
(Leech, 1971):
151
(90) a. I was going to say that it looked a bit like a pheasant
in flight.
b. Twenty years later, Dick Whittington would be the richest
man in London.
To the above-mentioned expressions of futurity in English we can
also add the following: to be about to (used to express imminent future
situations; it is less colloquial than to be going to), to be ready to, to be
near to, to be on the point of/on the verge of/on the brink of:
(91) a. He was about to retrace his steps when he was suddenly
transfixed to the spot by a sudden appearance.
b. His finger was upon the trigger and he was on the point of fire.
c. He has been on the brink of marrying her.
d. He was just on the point of proposing to her.
e. I was very nearly offering a large reward.
11.0. The Temporal Values / Uses of the Future Perfect Tense

The future perfect tense, illustrated in the examples below, contains
both a tense morpheme (will / shall) and an aspectual one (have).
(92) a. By this time tomorrow, I shall have crossed the Channel.
b. If you come at 7 o’clock we shall not have finished dinner.
c. By the end of next week, he will have left the hospital.
d. If you keep forgetting to water the plants there won’t be any
by the end of the week. They will have died.
The meaning of a sentence in the future perfect such as By the time
you come I shall have finished writing the paper is: by the time you come
in the future (RT) I shall finish writing the paper (ET) before you come.
The ATR of the future perfect tense is [ST < RT, ET < RT].
If the situations are viewed from the vantage-point of the past, the
use of the future perfect in the past is dictated by the sequence of tenses:
(93) Four o’clock! In another hour the Whitsuntide party, for which
the house stood ready, would have arrived.
The sentence in (93) contains a future perfect in the past: its
morphological tense is ‘past’ (would) and its morphological aspect is
perfect (have arrived).
152
8. MODALITY AND MODAL VERBS

1.1. Modal expressions are linguistic expressions that allow people
to talk about alternative states of affairs. These states of affairs are
thought of as not present in the current situation and may never occur in
the actual world.
Typical examples of modal expressions are the modal verbs can,
may, must, will, shall, could, might, would, should, ought to, need and
dare and the grammatical category of mood (roughly the indicative mood
versus the subjunctive mood distinction).
1.2. Consider the contrast in meaning between an unmodalised
assertion as in (1a) and the modalised assertions in (1b,c) (cf. Cornilescu
2003):
(1) a. Tom is the murderer.
b. Tom must be the murderer.
c. Tom might be the murderer.
In uttering (1a) the speaker makes an unmodalised assertion. His
statement expresses a commitment that the proposition is true in the real
world (i.e., the proposition is taken as a fact in the real world). Sentences
(1b) and (1c) contain modal verbs and express modalised expressions.
These sentences are conclusions that the speaker draws on ‘what is
known’ in the context. They differ only regarding the strength of the
conclusion they embody. Sentence (1c), containing the modal verb might,
says that given ‘what is known’ the speaker concludes that it is possible
for Tom to be the murderer. In contrast, (1b) which contains the modal
verb must, says that given ‘what is known’ it is necessary that Tom
should be the murderer (yet it is not a fact).
153
Since modal verbs change the meaning of the sentence they occur
in they are interpreted as modal operators. A modal operator is an
operator which, when added to a proposition p, yields another proposition
q (A proposition describes the semantic content of a sentence, that is, its
meaning).
1.3. The notional definition of the category mood covers the
semantic (i.e., meaning) distinction between factual (i.e., related to facts
in reality – the indicative mood) and non-factual (the subjunctive mood).
Consider the contrast:
(2) a. James is the president of the company. (indicative mood)
b. I wish that James should become president of the company.
(subjunctive mood)
In uttering (2a) the speaker makes an unmodalised assertion and the
proposition is taken as a fact in the real world. In uttering (2b) the speaker
makes a modalised assertion and the proposition that Tom should become
president of the company is non-factual; it is evaluated for its truth-value
only with respect to my wish-world. It is clear that the verb wish in the
main clause also acts as a modal operator. Roughly speaking, the modal
operator expresses an attitude towards the operand proposition (towards
the proposition on which it operates).
1.4. Besides modal verbs and verbs that require the use of the
subjunctive in the subordinate clause (such as assume, believe, fancy,
fear, feel, guess, hope, imagine, presume, reckon, surmise, suspect, think,
trust, declare, etc.), various nouns, adjectives and adverbs can also
express modal concepts and attitudes:
Nouns: allegation, hypothesis, command, exhortation, request,

assumption, certainty, doubt, expectation, etc.
Adjectives: sure, certain, possible, necessary, probable, compulsory,
imperative, lawful, legal, permissible, etc.
Adverbs: allegedly, certainly, conceivably, evidently, likely,
necessarily, obviously, possibly, presumably, probably, seemingly,
surely, etc.
154
2.0. On Root and Epistemic Meanings of Modal Verbs
2.1. Modal auxiliary verbs such as can, may, must, could, might,
should, etc. are verbs with weak, incomplete semantic content. Their
interpretation is highly dependent on the (linguistic and non-
linguistic/pragmatic) context in which they occur.
In the linguistic literature it is acknowledged that modal verbs are
used to communicate at least two clusters of meaning: the root (or
deontic) meaning and the epistemic meaning.
The root meaning roughly deals with the necessity or possibility of
acts performed by morally responsible individuals (i.e., agents) and
covers notions such as obligation, volition or permission.
The epistemic modal meaning deals with the possibility or necessity
of an inference drawn by the subject from available evidence and covers
notions such as necessity, probability and possibility.
It is commonly the case in English that one single modal verb is
capable of conveying both root and epistemic modal meanings. Consider:
(3) a. Employees must feed the animals twice a day. (obligation)
b. Whoever has finished may go. (permission)
c. You should be grateful to your parents for their support. (moral
obligation)
d. I will do it no matter what. (volition/threat)
(4) a. You must be John’s wife. (necessity)
b. It may rain later in the afternoon. (possibility)
c. Bioethics lectures should prove interesting. (probability)
d. A lion will attack a man when hungry. (probability)
2.2. Many linguists have argued that modal verbs are not
ambiguous between a root/deontic and epistemic meaning. Instead, it is
claimed that each modal verb has a core meaning that constitutes the
basis used to derive the range of possible interpretations of each modal
verb function of its context of use (Perkins 1983, Haegeman 1983,
Ştefănescu 1988, Kratzer 1977, 1981, 1991, Papafragou 2000).
Let us consider more closely the notion of context of use. In
general, conversation unfolds against a common conversational
background. The sentences people utter when they converse are taken as
155
propositions (Remember that a proposition describes the semantic content
of a sentence). The common conversational background is the set of
propositions accepted as true by the discourse participants. The
propositions in the conversational background play an important role as
they are taken as implicit premises in the judgements speakers make (cf.
Cornilescu 2003).
The common conversational background must be modified when
dealing with modalised sentences. In this case, it is necessary to consider
not the whole common background but one of its subsets. A modalised
sentence such as (5):
(5) Mary must have climbed Mount Toby.
is understood against a subset of the larger conversational background
namely, the subset of ‘what is known’. This subset stands for evidence by
virtue of which the speaker draws the necessary conclusion that Mary has
climbed Mount Toby (cf. Giorgi and Pianesi 1997). The various subsets of
the conversational background of modalised sentences are called modal
bases.
Modal bases are again sets of propositions but they are contextually
determined. It is against modal bases that the truth of modalised sentences
is evaluated.
Consider the following set of examples that contain the modal verb
must, discussed in Kratzer 1977:
(6) a. All Maori children must learn the names of their ancestors.
b. The ancestors of the Maoris must have arrived from Tahiti.
The verb must in (6a) has a root/deontic meaning: it refers to a
duty/obligation. The verb must in (6b) is used epistemically: it refers to a
piece of knowledge. Kratzer tries to identify the feature that connects the
two distinct uses of the verb must in (6a) and (6b). She does this by means
of paraphrasing the sentences in (6) in the following way:
(6’) a’. In view of what their tribal duties are, all Maori children
must learn the names of their ancestors.
b’. In view of what is known, the ancestors of the Maoris must have
arrived from Tahiti.
156
These paraphrases show that the verb must has shifted its meaning
to different ‘in view of’ phrases. We notice that in (6’) the verb must is
used relationally. In these sentences we do not have to do with an
absolute must but with a relative ‘must in view of something’. The
phrases ‘in view of something’ constitute the two modal bases of the
modalised sentences in (6). Therefore, a sentence that contains a modal
verb (which is interpreted as a modal operator) has three components:
Operator Modal Base Proposition p

MUST in view of what their tribal all Maori children must
duties are learn the names of their
ancestors
in view of what is known the ancestors of Maoris
must have arrived from
Tahiti
The operator MUST operates on the proposition p and it relates p to
the modal base in which it is evaluated for truth-value. As indicated
above, it is the modal base that is responsible for the deontic or epistemic
meaning that a modal verb expresses in different contexts.
The modal base may be linguistically indicated as in (7) below or is
pragmatically inferred as in (8) (Papafragou 2000):
(7) a. In view of the political situation, you must leave the country.
b. If you want to avoid a scandal, you must leave the country.
c. In case you plan to save yourself, you must leave the country.
(8) You must leave the country.
Modal bases are organized in various subsets (domains) of the
conversational background: the factual domain (i.e., propositions that
describe the factual world), the regulatory domain (i.e., propositions that
include legal rulings, social regulations, religious rules, chess rules, etc.).
There is also the domain of moral beliefs (i.e., propositions that are
descriptions of states of affaires in ideal worlds/situations), the domain of
desirability (i.e., propositions that are descriptions of states of affaires in
worlds/situations desirable from someone’s point of view), etc. Thus,
individual modal verbs come out as permitting different kinds of domains
of propositions as modal bases.
157
There are instances when the speaker – hearer can misunderstand
each other because they may be mistaken in the recovery of the modal
base. Jokes are a good example to illustrate misunderstanding in the
identification of the right modal base. (cf. Papafragou, 2000). Suppose
that a zookeeper says sentence (9) to his new assistant:
(9) The monkey can climb to the top of the tree.
Later on in the day, the monkey is missing and the zookeeper is
angry with the assistant because he warned the assistant that the monkey
was able to climb to the top of the tree and so could escape. The assistant
replies that he interpreted the sentence to be about what the monkey was
allowed to do, and was therefore not worried when the monkey behaved
in just this way. The misunderstanding is due to a domain mismatch: the
zookeeper had in mind potentiality in terms of the monkey’s physical
abilities, while the assistant had in mind potentiality in terms of the zoo’s
regulations.
A further piece of evidence, which shows that interlocutors are
sensitive to subtle aspects of the modal bases, is that speakers and hearers
frequently shift and modify modal domains/bases during the same
conversational exchange. Imagine that Alice and her lawyer have been
discussing the prospect of Alice’s having a divorce. Alice utters sentence
(10) and her lawyer replies as in (11):
(10) I can’t leave my husband penniless.
(11) Of course you can – the law allows you to.
The modal domain in (10) includes assumptions about Alice’s
feelings and moral profile whereas in (11) it includes assumptions about
legal regulations. This example is a clear demonstration of how the modal
base affects the truth-conditional content of a modal sentence.
Hence, modal verbs are treated as expressions with vague/weak
semantic content whose deontic or epistemic interpretation is highly
dependent on the context of use.
There is further narrowing down of the subset of modal bases that
ensures the correct interpretation of utterances. Consider the following
example:
(12) James must be at work; he doesn’t answer the phone.
158
This sentence involves the epistemic modal base of necessity,
i.e., ‘in view of what is known, James must be at work’.
However, not all possible states of affairs contained in a modal base
are taken into account when interpreting a modalised utterance. For
instance, in (12) the speaker does not take into consideration some
implausible states of affairs. For example, the state of affaires in which
James does not want to answer the phone because he suffers from an anti-
social crisis or the state of affairs in which James has suddenly lost his
voice because he suffers from a sore throat. Why should that be? A first
approximation is that situations containing these states of affairs are
‘wilder’ and unusual possibilities, which need not be taken into account.
More precisely, these states of affairs in the modal base are removed from
what we take to be the normal or ideal course of events.
Therefore, we need an ordering source that will order situations in
terms of similarity to the norm or ideal and specify the minimal ‘distance’
from the ideal a situation should satisfy. The ordering excludes those
states of affairs that are farther from the normal course of events.
Ordering sources capture the observation that the understanding of a
modalised sentence often implies the use of stereotypical/idealized states
of affairs. They describe the way situations should be, not as they really
are – that is, according to the normal courses of events, according to the
law, according to what is desirable, etc. Ordering sources may be
explicitly introduced by phrases such as in view of what is normal, in view
of what is desirable, according to the law, etc. The joint effect of modal
bases and ordering sources (which are another subset of the
conversational background) is to force the evaluation of a modalised
proposition in those states of affairs of the modal base that better realize
the norm or ideal (Giorgi and Pianesi 1997).
3.0. More on the Deontic Meaning and Epistemic Meaning and

the Shift from One Meaning to the Other Meaning
3.1. There is robust linguistic evidence that epistemic modality has
historically developed from items that originally encoded deontic
meaning (mainly volition, obligation and permission) (Bybee and
Pagliuca 1985, Bybee 1988, Heine 1991, Bybee and Fleischman 1995, in
Papafragou 2000). In other words, there has occurred a shift from
159
root/deontic modality, which involves representations of external states of
affairs to epistemic modality, which involves the speaker’s internal
ability to perform deductive operations on ‘what is known’.
3.2. For example, it is argued that should shifted from ‘weak
(deontic) obligation’ to ‘probability’ (epistemic) modal meaning.
(13) a. You should clean the place once in a while.
b. That should be the plumber.
Should in (13a) expresses weak obligation relative to norms or
existing stereotypes and it involves a state of affairs that is external to the
speaker. Moreover, modal deontic meaning presupposes the existence of
a responsible individual (i.e., agent) who is supposed to carry out the
relevant activity. Sentence (13a) is interpreted as imposing obligation on
the hearer to clean the place from time to time; it is the hearer who is the
agent / doer of the activity commanded, not the speaker.
That is why a modal deontic interpretation is agent-oriented and is
said to involve relations among ‘circumstances’ (Kratzer 1991,
Papafragou 2000).
The verb should in sentence (13b) expresses probability in terms of
the speaker’s internal beliefs about the normal expected course of events.
According to the speaker’s beliefs, the plumber is expected to arrive some
time after he is called for. The speaker cannot be certain about the
plumber’s coming but given the normal course of events he concludes
that whoever is ringing the bell is indeed the plumber.
That is why many authors claim that since epistemic modality
involves the speaker’s beliefs on which he draws to derive a conclusion,
epistemic modality is subject-oriented and it typically marks a shift
towards ‘subjectivity’ in language. “Epistemic modality is the speaker’s
assessment of probability and predictability. It is external to the content,
being a part of the attitude taken up by the speaker: his attitude, in this
case, towards his own speech role as a declarer” (Halliday 1970:349).
As the comments on the sentences in (13a) and (13b) indicate, the
shift of the modal verb should from ‘weak obligation’ to ‘probability’ can
be explained as a shift in the normative modal base it chooses: either a
normative deontic modal base or a normative epistemic modal base
(Papafragou 2000). In other words, despite the fact that the root/deontic –
160
epistemic distinction is mainly pragmatic in nature (both the conversational
background and the ordering source involve non-linguistic/pragmatic
knowledge), this modal distinction is due to the differentiation in language
between two distinct major classes of modal bases.
3.3. Consider now the examples in (14): they illustrate the
development of epistemic must out of deontic must. In both cases, a sense
of ‘necessity’ is conveyed.
(14) a. You must be back by 10 o’clock.
b. ”Susana Foo” must be one of the most expensive restaurants
in Philadelphia.
Must in (14a) expresses ‘strong obligation’ in a context relative to
the system of social laws as modal base. In such a context, the speaker
has authority over the hearer and the speaker’s desires are expected to be
(necessarily) carried out by the hearer.
In (14b) must is interpreted epistemically (‘necessity’): the speaker
has evidence on the base of which he draws the conclusion that ‘Susana
Foo is the most expensive restaurants in Philadelphia’ must be the case.
The epistemic interpretation of must is highly subjective as it depends on
the speaker’s belief-set backed up with various degrees of support by
available evidence.
Papafragou (2000) argues that in the epistemic interpretation of all
modal verbs, the base proposition p is used as a representation of an
'abstract hypothesis’, which is compatible with or follows from the
speaker’s set of beliefs.
Let us illustrate this with the epistemic must. The sentence in (14b)
is analyzed into two components:
(15) a. Susana Foo is one of the most expensive restaurants in
Philadelphia.
b. As far as I know, it must be the case that (a).
The role of (15b) is that of ‘interpreting’ the base proposition in
(15a). That is, (15b) is a comment on (15a) and it conveys the degree of
the speaker’s commitment to the proposition p in (15a).
In (14b) above the speaker’s commitment is high and the level of
support in the extra-linguistic reality is strong since he uses must
161
(necessity). If the speaker’s evidence for the truth of the base proposition
p were weaker he would choose to use may or might.
The proposition p under a modal verb interpreted epistemically is
“not to be treated as directly picking out a state of affairs in the world but
as an epistemic object” (Sweetser 1990). In fact, the proposition p is a
representation formed in the mind of the speaker. The epistemic operator
expresses the speaker’s attitude towards this base proposition.
In their epistemic uses, all modal verbs communicate a type of
relation (necessity or possibility) between a certain proposition p (as in
(15a)) and the speaker’s belief-set (as in (15b)).
From the speaker’s point of view, the use of epistemic modality
rests crucially on his ability to reflect on the content of his own beliefs.
The speaker takes into account the reliability of those beliefs and
performs deductive operations on them.
From the hearer’s point of view, the comprehension of an
epistemically modalised utterance involves reference to the
communicator’s evaluation of a proposition in terms of mentally
represented evidence (cf. Papafragou 2000).
In conclusion, an utterance that contains a modal verb (i.e., a modal
operator) has three component parts, no matter whether it is interpreted
deontically or epistemically:
modal operator modal base proposition p
In case the utterance is interpreted deontically, the modal operator
operates on the proposition p and relates p to a modal base for its truth-
conditions. The deontic modal base is a subset of the conversational
background in which the proposition p is evaluated for its truth (see
section 2.2.).
In case the utterance is interpreted epistemically, the modal
operator operates on the proposition p (a representation formed in the
mind of the speaker), which is verified for its truth-value in a modal base
made up of the speaker’s set of beliefs.
In view of the above observations, it is no wander that epistemic
modality developed out of deontic modality in the process of
subjectification in grammar. The meaning of deontic modal verbs
broadened and bleached to the extent that modal verbs became
162
compatible with both deontic modal bases (i.e., subsets of conversational
backgrounds) and epistemic modal bases (i.e., the speaker’s belief-sets).
4.0. Morpho-syntactic Properties of Modal Verbs

4.1. Because of their morpho-syntactic properties, modal verbs
form a special class of auxiliary verbs. Their properties set them apart
from both lexical verbs and the other class of auxiliaries (i.e., the
aspectual auxiliaries have and be).
The most striking characteristics of the English modals are the
so-called NICE properties (Huddleston, 1976) (NICE is the acronym
from Negation, Inversion, Coda and Emphasis properties of modals).
The following properties distinguish the English modals from
lexical verbs:
(i) Negation can attach to the modal, without do-support:
(16) I cannot come / *I do not can come.
(ii) Inversion of the subject with the modal is obligatory in
interrogative sentences and in tags; do cannot be inserted:
(17) Must they leave? / *Do they must leave?
(18) You can speak English, can’t you? / You can speak English,
*don’t you?
(iii) Modals can appear in the “coda”:
(19) I can come and so can Bill / *I can come and so does Bill.
(iv) Emphatic affirmation is possible, again without do-support:
(20) You shall have the money by tomorrow! / *You do shall have
the money by tomorrow!
The following properties distinguish the modal auxiliaries from the
aspectual auxiliary verbs have and be:
(i) modals lack conjugation:
(21) *He cans swim / I am swimming / He has written.
(ii) modals cannot be followed by non-finite forms:
(22) *He can to swim / *He can swimming / He has left / He is writin.
163
(iii) modals always select a short infinitive as their complement:
(23) He can (*to) swim.
(iv) like aspectual auxiliaries, modal auxiliaries may have both past
and present tense forms (see 24); some of them have a past
tense form that can only be used in reported speech (see 25):
(24) They can play the piano / He could play the piano when he
was young.
(25) The boss said he might go right away.
(v) modals have no passive form
(vi) modals have no imperative form
(vii) modals cannot co-occur (with the exception of some dialects):
(26) *You might would say that.
4.2. Hoffman (1976) also noticed a systematic syntactic distinction
between the epistemic and the root interpretation of modal verbs.
When modal verbs are interpreted epistemically (when they span
notions such as possibility, probability, and logical inference) they evince
the following syntactic patterns:
(i) with this interpretation, modals can occur in the progressive
aspect:
(27) He can be singing now.
(ii) with this interpretation, modals can occur with the perfect
infinitive form:
(28) He must have already left
(iii) with this interpretation, modals have no selection restrictions
on the subject:
(29) The apple/the boy must have fallen from the tree.
When modal verbs are interpreted deontically (i.e., their meaning
spans notions such as permission, obligation, volition, ability, etc.) they
cannot occur in the progressive aspect or with the perfect infinitive form:
(30) *He can be singing now. (no ability interpretation)
(31) *He must have already left. (no permission interpretation)
164
Rather, deontic/root modals have a future time orientation:
(32) a. You may sing now. (permission)
b. You must leave now. (obligation)
This is explained in the following terms. A responsible agent
typically grants permission for an activity that will be carried out in the
future (not the past or the present); the same goes for imposing ‘obligation’.
More often than not root modality is associated with activities while
epistemic modality is associated with events (Papafragou 2000).
4.3. It has also been pointed out that root modals impose selection
restrictions on the subject: the subject should be animate.
(33) *My car must leave now. (no obligation interpretation)
This is so because deontic modals, which communicate permission,
obligation, etc., involve an agent responsible for carrying out the relevant
activity. However, this generalization is not absolute. Consider the
following examples:
(34) a. There must be law and order in the country.
b. The table should be ready for dinner before 7.
4.4. Moreover, it has also been noticed that since epistemic
modality involves inferential processes on the part of the speaker,
epistemic modals unlike root modals cannot appear sentence-initially in
yes-no interrogatives:
(35) May the race start?
‘Is there permission for the race to start?’
*’Is it possible that the race starts?’
(36) Should John leave?
‘Is it required that John leave?’
*’Is it predictable that John will leave?’
More specifically, it is hard to construct a context in which the
speaker asks whether a conclusion is possible or necessary with respect to
his own set of beliefs. However, once such a context becomes available,
interrogative-initial epistemics become acceptable. Consider:
(37) a. Might John be a liar?
b. Must John be a liar?
165
Such sentences involve deliberative questions in which the speaker
addresses a question to himself in an attempt to elaborate the evidence he
has for a certain conclusion.
4.5. It was also noticed that negation operates differently in root
and epistemic modality. As a rule, negation affects the modal predication
if the modal has root meaning (i.e., root modals scope inside negation) as
in (38) below while on the epistemic interpretation of a modal verb
negation affects the main predication (i.e., epistemic modals scope
outside negation) as in (39) below (Coates 1983):
(38) You may not enter. (root/deontic meaning)
You are not allowed to enter.
*You are allowed not to enter.
(39) You may not be given this opportunity again.
(epistemic meaning)
It is possible that you will not be given this opportunity again.
*It is not possible that you will be given this opportunity again.
The different locus of negation in the two interpretations of modal
verbs has been taken as a syntactic reflex in syntactically separating
deontic / root interpretation from epistemic interpretation.
However, it appears that the position of negation in the two modal
interpretations is not a reliable test for separating root meaning from
epistemic meaning. Rather the scope divide with respect to negation
seems to be drawn along the following lines: necessity operators scope
over negation (i.e., they occur before negation) while possibility
operators scope under negation (i.e., they occur after negation). The
following examples together with their analyses are taken from
Papafragou (2000:92).
4.5.1. Necessity Operators: they scope over negation
(40) a. These children must not be older than 3.
(epistemic meaning)
‘It is certain that these children are not older than 3’.
*’It is not certain that these children are older than 3’.
b. They must not leave. (root meaning)
‘It is required that they do not leave’.
*’It is not required that they do not leave’.
166
(41) a. He should not be late; he only went across the road.
(epistemic meaning)
‘It is predictable that he will not be late’.
*’It is not the case that it is predictable that he will be late’.
b. John should not drink alcohol. (root meaning)
‘It is advisable for John not to drink alcohol’.
*’It is not the case that it is advisable for John to drink alcohol’.
(42) a. She ought not to appear before noon. (epistemic meaning)
‘It is predictable that she will not appear before noon’.
*’It is not the case that it is predictable that she will appear
before noon’.
b. You ought not to come. (root meaning)
‘It is required that you do not come’.
*’It is not required that you come’.
4.5.2. Possibility Operators: they scope under negation

(43) George can not swim.
‘George does not have the ability to swim’.
(44) George could not swim.
‘George did not have the ability to swim’.
(45) a. You need not resign.
‘It is not the case that you need resign.
b. The proof need not exist.
‘It is not the case that the proof need exist.
(46) You dare not resign.
‘It is not the case that you dare resign.
The scope of negation and its effects will be signaled out when we
analyze the root and epistemic interpretations of modal verbs (especially
the cases of modal can and must).
In what follows we present a detailed description of the modal
verbs can, may, must, will, shall, could, might, ought to, would, should,
need and dare on both their deontic and epistemic interpretations when
their modal bases are not linguistically indicated but must be inferred
from the context.
167
5.0. CAN
The modal meaning of CAN generally covers the notion of
potentiality (Watson 1988, Bolinger 1989, Klinge 1993, Papafragou
2000). A state of affairs is characterized as potential when it is compatible
with the state of affairs in the real world. That is, the modal base of CAN
does not preclude an activity x from taking place now or in the future.
(47) Computer-aided instruction can co-occur with more traditional
methods of teaching.
The modal base of CAN in (47) is the set of assumptions/facts
about technological progress in current teaching practices. This modal
base can be further contextually narrowed down to picking out subsets of
these assumptions, which are again highly compatible with the states of
affairs in the actual world. The ‘in view of’ paraphrases are subsets of the
modal base and the interpretation of CAN in these contexts is root/deontic
‘potentiality’:
(48) a. In view of the way schools are run today….
b. In view of the technological equipment available for
education processes….
c. In view of the teachers’ encouraging stance….
computer-aided instruction can co-occur with more
traditional methods of teaching.
Consider now the following examples:
(49) a. Cigarettes can seriously damage your health.
b. Even expert drivers can make mistakes.
Traditional grammars have interpreted CAN in (49a,b) as having an
epistemic ‘possibility’ interpretation (i.e., available evidence allows the
speaker to infer that cigarettes can seriously damage your health). Notice
however that these sentences also allow a deontic ‘potentiality’
interpretation: given statistics on how dangerous smoking is, smoking
cigarettes is potentially dangerous.
168
That is why many linguists (Leech 1971, Ştefănescu 1988) argue
that the modal meaning of CAN is “best viewed as indeterminate outside
a larger context that helps us choose between these two alternative
interpretations” (Ştefănescu 1988:423).
Other linguists claim that in affirmative sentences CAN has no
epistemic interpretation at all since it always operates on factual modal
bases (Papafragou 2000). Instead, epistemic possibility in affirmative
sentences is conveyed by the use of the modal verb MAY. Consider the
following context of use that illustrates the claim:
(50) Do you think that James is hiding something from the authorities?
a. He may be, and again he may not.
b. *He can be, and again he can not.
In (50) only MAY gives a well-formed sentence and is interpreted
epistemically (i.e., possiblity) while CAN is precluded from occurring
with an epistemic modal base. In (50a) the speaker communicates that the
proposition ‘James is hiding something from the authorities’ is
compatible with his set of beliefs but goes on (after some deliberation)
asserting that the negation of this proposition (~p) is also compatible with
his belief-set. That is, the speaker’s current beliefs leave open the
possibility for either p or ~p to come out as true. In contrast, in (50b) since
the use of CAN is always compatible with a factual modal base, we arrive
at a contradiction in terms: on the one hand, the proposition p is
compatible with the set of factual propositions but on the other hand it is
not (~p).
In sum, since CAN operates on factual modal bases it acquires only
a deontic ‘potentiality’ interpretation (Papafragou 2000).
However, in negative and interrogative sentences CAN is
interpreted epistemically (i.e., possibility):
(51) a. He can’t be that impudent. (impossibility)
b. Can she be that old? (possibility)
In what follows we shall adopt the more conservative view, which
states that CAN has both a deontic and an epistemic interpretation. Yet,
we acknowledge the fact that in affirmative sentences the use of MAY on
its epistemic (possibility) interpretation is preferred (Leech 1971).
169
Traditional grammarians (Leech 1971, Palmer 1979, Jespersen
1931, Poutsma 1926) have identified three main modal meanings of
CAN: ability, permission (root/deontic meanings) and possibility
(epistemic meaning).
5.2. Ability CAN (root meaning)

On its root/deontic interpretation, CAN refers to both physical and
mental ability:
(52) a. He can lift a hundredweight.
b. John can swim like a fish.
c. Can you speak English?
The ability sense of CAN may be easily included in the broader
‘potentiality’ meaning of CAN with a present or future time orientation.
Consider the following contrast:
(53) John is swimming.
(54) John can swim.
In (53), the speaker asserts that John is now engaged in an ongoing
activity of swimming. In (54), the speaker does not assert that John is
actually swimming now. Sentence (54) conveys the meaning that the
speaker has in mind some sets of circumstances that include previous
occasions on which John demonstrated his ability to swim. These
previous occasions are such that they do not preclude a similar occurrence
in the future. On its ability/potentiality interpretation CAN covers present
and future time-spheres while the present progressive covers only a
present time interval. The ability/potentiality meaning of CAN is
paraphrased as ‘John can swim and will swim’ (Palmer 1979).
Due to its general interpretation of potentiality root CAN may also
indicate various kinds of acts such as suggestion, offer or request:
(55) We can meet one day after work. (suggestion)
(56) I can give you a lift. (offer)
(57) a. Can you just remind me? (request)
b. Can you give me a hand with the painting? (request)
These types of interpretation require the presence of a collection of
contextual assumptions that involve both the hearer and the speaker. In
170
(55) and (56), the embedded proposition p represents a state of affairs that
is desirable to the hearer from his own point of view as well as beneficial
to him. On the other hand, the speaker has the responsibility for bringing
about the state of affairs described in p but no obligation to do so. In
(57a,b) CAN is used in an interrogative sentence that has you as subject.
In this case, the interrogation does not function as a genuine question that
elicits information, but as a request addressed to the hearer.
Ability/potentiality CAN is associated with two additional
occurrences:
(i) CAN may occur with verbs of sensation (hear, see, smell, feel,
and taste) and verbs of cognition (understand, remember, think, stand,
bear):
(58) I can see the moon / A song can be heard in the distance / I can
feel the presence of a stranger in the room.
(59) How much competition can I stand from now on? / What you
can remember in two weeks is the thing that matters.
(ii) CAN may also have the so-called ‘characteristic’ reading, which
may also be subsumed to its ‘potentiality’ interpretation:
(60) He can tell awful lies / She can be catty (at times) / She can be
charming.
Notice that the ‘characteristic’ meaning of CAN has little to do with
the verb CAN itself; it is the whole sentence that purports this
interpretation rather than the sense of CAN. Moreover, if we leave the
modal verb aside, the sentences above acquire habitual/generic
interpretations (He tells awful lies / She is catty / She is charming).
It was long ago noticed that in many contexts deontic CAN may
occur when BE ABLE TO does since both are related to circumstances
that do not preclude an activity from occurring:
(61) a. He can / is able to lift a hundredweight.
b. Our team can / is able to beat your team.
However, the two verbs are not always interchangeable. While
CAN really shows general ‘potentiality’, BE ABLE TO rather specifies
the subject’s ability to accomplish a certain task under specific
circumstances. These shades of meaning, which distinguish CAN from
171
BE ABLE TO are made clear when these verbs occur with an embedded
conditional clause such as ‘if he likes’. Consider the contrast:
(62) a. John can swim, if he likes.
b. *John is able to swim, if he likes.
In (62a) the interpretation of CAN is ability/potentiality as John’s
swimming may be actualized any time (now or in the future). Sentence
(62b) is ill formed because the accomplishment of a task is dictated by
circumstances and does not depend on somebody’s liking or disliking to
carry out the activity of swimming.
Consider other examples with BE ABLE TO:
(63) a. In this way we are able to carry our research and not simply
to undertake consulting.
b. We are able, in mathematical terms, to find the optimum
solution.
However, BE ABLE TO is preferred to CAN under the following
conditions (Palmer 1979):
(i) since CAN has no non-finite forms, only BE ABLE TO is
available after other modal verbs: should be able to, might be
able to, must be able to.
(ii) BE ABLE TO is more formal than CAN and tends to be used
especially in written texts.
(iii) if the temporal reference of the sentence is past and if the
situation is a single accomplished occurrence, only BE ABLE
TO is allowed:
(64) I ran and I was able to catch the bus.
(65) *I ran and I could catch the bus.
5.3. Permission CAN (root meaning)

Beginning with the 18th century CAN has come to be used with the
sense of permission. With this meaning it has tended to replace MAY:
(66) a. You can go now.
b. You can smoke in this room.
172
In (66) the state of affairs described by the proposition p is desirable
from the hearer’s point of view. The speaker has authority over the hearer
but it is in the hearer’s power to bring about this state of affairs.
In present-day colloquial English CAN with permission
interpretation is more widely used than MAY although it is perceived as
less polite than MAY. CAN has the less specific meaning of ‘you have
permission’ rather than the more specific one of ‘I grant you permission’.
The sentence You can smoke in this room simply means that
rules/regulations allow smoking in this room:
(67) Mr. X: Can I smoke in here?
Mr. Y: So far as I know you can – there is no notice to the contrary.
However, permission CAN is not always used for granting
permission to the benefit of the hearer. Consider the sentences in (68):
(68) a. You can forget about your holidays.
b. If he doesn’t like it he can lump it.
c. You can jump in the lake.
All the sentences in (68) are more or less offensive and their
impolite tone contains an ironical touch: the speaker offers the hearer a
choice that is not convenient or desirable from the hearer’s point of view.
In the same line, a suggestion made by the speaker to the hearer is
so sarcastic that the suggestion is interpreted as a command of a brusque
and somewhat impolite kind:
(69) a. You can leave me out, thank you very much.
b. You can start looking for another job!
The explicit future time reference of ability CAN is rendered by the
future tense of BE ABLE TO:
(70) He’ll be able to run a mile in 4 minutes next year.
while the explicit future time reference of permission CAN is rendered by
the future tense of the verbs permit or allow:
(71) I shall allow/permit you to talk with her.
The past time reference of ability and permission CAN is conveyed
by COULD:
(72) a. Your mother was out and couldn’t have the key.(past ability)
173
a’. Why was the house so cheap? No one could get a mortgage
on it.
b. He said I could leave the next day. (past permission)
b’. She said that, if he wanted, he could come.
COULD may also acquire a habitual interpretation when a
recurrent event in the past is intended:
(73) My father could usually lay hands on what he wanted.
5.3.1. Hypothetical COULD

However, deontic COULD does not exclusively cover the past time
sphere of various actions. COULD may describe ideals of action or
behaviour in subsets of its modal base (i.e., ordering source) that are
closer and more similar to the way the world should be. In this case, the
time sphere of COULD is present or extended present. Consider the
following examples:
(74) a. Do you think you could command an army? (ability)
b. I could do it with the right moral support. (ability)
c. He could go if he wants to.
(more polite form of granting permission)
This use of COULD is also called ‘hypothetical’ COULD (Leech
1971).
Consider other examples of hypothetical COULD:
(75) a. Well, could we go on to modern novels, then?
(more polite form of asking permission)
b. Could I speak to the manager, please?
(more polite form of asking permission)
5.4. Possibility CAN (epistemic meaning)

As already said, possibility CAN is more frequently used in non-
assertions, that is, in negative and interrogative sentences:
(76) a. He can’t be that stupid!
b. Can she be that old?
In affirmative sentences, epistemic (possibility) MAY is preferred:
174
(77) a. Michael may well get his degree next year.
b. He may have been joking when he said that.
In colloquial English, epistemic CAN has come to be used as a
‘tactful/democratic’ imperative when addressing a person regarded as
one’s equal (Leech 1971):
(78) Mike and Wily, you can be standing over there; and Janet can
enter from behind the curtain.
Remember that epistemic CAN, unlike deontic CAN, may occur in
the progressive aspect:
(79) You can be standing on it, so be careful!
The past time reference of epistemic CAN is rendered by CAN +
the perfect infinitive form of the main verb:
(80) a. He can have been hiding at that time, you know.
b. Poor devil, he can’t have been too happy either.
c. Can you have made such a mistake?
5.4.1. Hypothetical COULD

In hypothetical situations, COULD may acquire an epistemic
interpretation covering a present or extended present time sphere:
(81) a. There could be trouble at the match tomorrow.
b. You could try to be a bit more polite.
In (81b) COULD is used to convey reproach and the sentence is
paraphrased as ‘it would be possible for you to be a little more civilized
but in fact you don’t behave yourself’.
Hypothetical COULD may also combine with the perfect infinitive
form of the main verb (it is the perfect infinitive that carries the past tense
marker of the sentence while COULD is interpreted hypothetically):
(82) You could have given me some notice (but you didn’t).
The combination in (82) leads to a ‘contrary-to-fact’ interpretation
and is contextually used as a complaint about a past omission.
175
6.0. MAY
The modal verb MAY is used with two different interpretations: a
permission (root) interpretation and a possibility (epistemic) interpretation.
6.1. Permission MAY (root meaning)

Permission MAY is perceived as more polite than permission CAN
because it includes the speaker’s preferences and conveys a greater
involvement on the part of the speaker. Consider the following example:
(83) During the seminar, you may interrupt as often as is needed.
Suppose that the person who teaches the seminar addresses his class
with the sentence above. The teacher makes it manifest that students are in
a position to interrupt as long as he does not object. (The modal base of the
sentence is thus, the teacher’s own preferences). Given that interrupting the
seminar as often as is needed is also desirable from the students’ point of
view and that the teacher has some authority as far as leading the seminar is
concerned, (83) conveys a permission interpretation.
Consider also the following sentence uttered by a bank employee
addressing a young customer:
(84) Our branch may convert your account into a student account;
you just need to supply us with proof of student status.
The young customer (the speaker, i.e., the student) has made
manifest that he is interested in changing the account he has and that he
considers the change desirable from his own point of view. The bank
employee (who has banking authority) informs the customer that the bank
regulations permit the branch to convert his account into a student
account. In this context, bank regulations constitute the modal base over
which the modal verb operates.
Consider also the examples below where permission MAY has
rules and regulations as a modal base:
(85) A local health authority may, with the approval of the
Minister, receive from persons to which advice is given under
this section….such charges, as the authority considers
responsible.
176
(86) ….and it is subject to the final prerogative of mercy of the
Home Secretary who may recommence a reprieve.
However, it is not the case that MAY is always felt to be polite.
Imagine that a general utters (87) thus interrupting a soldier who is
talking:
(87) You may go home, Jones.
In this context, Jones does not consider going home as desirable
from his own point of view; actually, the fact that Jones was in the middle
of talking rather makes manifest the opposite assumption (i.e., that he
does not wish to leave immediately). The utterance is perceived as rude
because the general ignores the hearer’s preferences, although they are
contextually salient.
When permission is granted in the future the future tense of be
allowed to or be permitted to is used:
(88) He’ll be allowed to/permitted to leave the hospital in two
weeks time.
The past time reference of permission MAY is rendered by
MIGHT, used in reported speech:
(89) He said she might leave.
6.1.1. Hypothetical MIGHT

MIGHT is also used to convey hypothetical, formal / very polite
requests in the present:
(90) Might I have a light, please?
6.2. Possibility MAY (epistemic meaning)

MAY has a possibility interpretation in contexts such as:
(91) a. A friend may betray you.

b. If he should marry, it may well be that his wife would like a
house of her own.
c. Perhaps she may be his daughter, though he is not married.
d. Careful, that gun may be loaded.
177
MAY on its possibility interpretation indicates that evidence
available to the speaker is such that the sentence is not currently assumed
to be true but nor can it currently be false. MAY focuses primarily on the
current verifiability of the truth of the sentence.
The past time reference of epistemic MAY is conveyed by MIGHT
+ perfect infinitive forms:
(92) He may have left earlier.
6.2.1. Hypothetical MIGHT

Hypothetical possibility MIGHT is used to render polite
suggestions in the present:
(93) We might meet again after Christmas if you like.
In familiar speech MIGHT is also used in reproaches (as COULD
may):
(94) a. You might stop grumbling at me, for a change!
b. You might help me with the luggage.
The assumption is that although it is possible for you to stop
grumbling at me, you don’t.
In the past, hypothetical epistemic MIGHT + perfect infinitive form
acquires a contrary-to-fact interpretation (in fact, the event did not take
place):
(95) a. Arthur might have taken it away again.
b. You might have told me about it in due time.
The sentence in (95b) means that although you had the
possibility/opportunity to tell me about it in due time, you didn’t. The
sentence qualifies as a complaint or reproach about a contrary-to-fact
situation.
6.3. A Parallel between CAN and MAY

From an epistemic point of view, both CAN and MAY render a
possibility and in many contexts the distinction between them is hardly
perceivable:
178
(96) After his divorce, Peter can/may visit his daughter whenever
he likes.
However, CAN and MAY are not always interchangeable.
Consider the following pair of examples:
(97) a. The dollar can devalue.
b. The dollar may devalue.
Out of the two statements, it is (97b) that is far more worrying for
those who are interested in the health of the dollar currency. Sentence
(97a) may be uttered at any time while sentence (97b) only at a time of a
financial crisis. The contrast between the following sentences is
interpreted in the same way:
(98) a. A friend can betray you.
b. A friend may betray you.
Sentence (98a) is a general observation made about friends, that is
the circumstances include at least an instance of a friend committing a
betrayal; hence, the circumstances are such as not to prevent another
instantiation of betrayal (potentiality). Sentence (98b) is more likely to be
a warning about one particular friend.
7.0. MUST
The modal verb MUST has both a root/deontic meaning
(obligation) and an epistemic meaning (necessity, logical deduction).
7.1. Obligation MUST (root meaning)

When MUST is used in a context relative to the system of social
laws and a person is in a position of authority MUST is interpreted as
indicating obligation or compulsion:
(99) a. You must be back by 10 o’clock.
b. Tell him he must stop this dishonest behavior.
c. She must be made to do it.
d. You must get in permanent jobs.
The root/deontic interpretation of (99) arise under the following
conditions: (i) the modal base involves the speaker’s desires and factual
179
assumptions; (ii) the speaker has authority over the hearer; (iii) the hearer
is in a position to bring about the state of affairs described in the
embedded proposition.
A variant of the ‘obligation’ interpretation of MUST arises in cases
of quasi-imperative suggestions or offers:
(100) a. We must go for a drink some day.
b. You must come and visit us sometime.
The first sentence in (100) conveys the meaning that the speaker’s
desire is that he and his addressee go for a drink one day. Since the
addressee most probably shows sensitivity to his interlocutor’s desire
(due to the social rules governing their relations), the sentence is
interpreted as an urgent form of suggestion / offer.
When self-compulsion is inflicted MUST may also occur with
subjects in the first person I or we:
(101) a. I must do what I can to protect my wife.
b. Yes, I must ask for that Monday off.
c. We must use it once or twice.
When MUST occurs in interrogative sentences it is the hearer’s
authority that is questioned and not that of the speaker’s. Thus a question
such as Must I answer all these letters myself? means Are these your
orders?.
Moreover, there is also a sarcastic or ironical usage of MUST in
Must you make that dreadful noise? (i.e., For heaven’s sake, stop it!) or in
If you must behave like a savage, at least make sure the neighbors aren’t
home.
Contextually, MUST can be used as a firm piece of advice almost
equivalent to an imperative, as in You must keep everything to yourself, be
discrete!
MUST is often used with a limited set of verbs related to the act of
conversation: I must say / admit / be honest / ask you / reiterate / confess /
mention and you must remember / admit / realize / understand, etc.:
(102) a. I must say, I’ve never known that.

b. I think I must make a confession here.
180
The form I admit is a stronger assertion than I must admit; hence, the
effect of the modal verb is to weaken the assertion made by the plain verb.
HAVE (GOT) TO is a synonym of deontic MUST and indicates
external compulsion while MUST itself implies internal compulsion.
(103) a. I have to be at the airport at four.
b. There is a whole lot of literature you’ve got to read.
c. He’s got to go into hospital, you know.
The examples in (103) contain both instances of HAVE TO and
HAVE GOT TO. According to Palmer (1979) there are two points of
difference between these forms:
(i) HAVE TO is more formal; HAVE GOT TO is generally used
in spoken language.
(ii) HAVE GOT TO has no non-finite forms (the forms *to have
got to, *having got to do not exist); instead, the forms of
HAVE TO must be used: to have to, having to.
The negative forms of HAVE TO and HAVE GOT TO are distinct
from each other: the negative form of HAVE TO is does / do / did not
have to while that of HAVE GOT TO is has / have / had not to:
(104) a. They haven’t got to juggle about. They’ve got the total
page copy.
(They are not obliged [to juggle about])
b. You don’t have to pay that fine.
(You are not obliged [to pay that fine])
7.1.1. Negation of Obligation MUST

When deontic (obligation) MUST is negated the paraphrase of
MUST assumes wide scope with respect to negation:
(105) They must not (mustn’t) leave. (interdiction)
‘It is required that they do not leave’.
When modality itself is negated there is no appropriate form with
MUST and NEED NOT (NEEDN’T) is used:
(106) You needn’t pay that fine. (lack of interdiction or obligation)
‘You are not obliged to pay that fine’.
181
7.1.2. If future time reference is explicitly made the forms will /
shall have to are used:
(107) a. Well, I’ll have to think about it.
b. I shall have to go into total silence for half an hour.
The time sphere of MUST itself is present or extended present and
it conveys a sense of ‘internal’ obligation:
(108) a. We must do something about it.
b. You must tell me the truth at once.
The past time reference of deontic MUST is taken over by HAD
(GOT) TO and no distinction between internal vs. external obligation is
retained:
(109) a. She had to sleep in the kitchen last night.
b. In his youth he had to work hard for a living.
c. He gave the children their presents in early December but
they didn’t have / hadn’t got to open them until Christmas
day.
In reported speech, however, the difference in interpretation
between MUST and HAVE (GOT) TO is retained and MUST is retained
as such:
(110) a. She said he must go.
b. She said he had to go.
c. He had trifled with life and must now pay the penalty.
7.2. Necessity MUST (epistemic meaning)

(111) a. In opening a game of chess, the players must move a pawn.
b. The president must formally approve the new Government
before it can undertake its duties.
c. The accused must remain silent throughout the trial.
Sentence (111a) expresses a necessity with respect to the rules of
chess, (111b) a necessity with respect to the Constitution and (111c) a
necessity with respect to judicial rules (Papafragou, 2000). As can be seen
these utterances require regulatory modal bases of different types.
182
In other contexts, the epistemic use of MUST occurs relative to the
system of rational laws as a necessary conclusion: “knowledge arrived at
by inference or reasoning rather than by direct experience” (Leech,
1971:148). Consider the following relevant examples:
(112) a. He must be working late in his office.
b. There must be some mistake.
c. You must have left your handbag at the theatre.
In each case, a chain of logical deduction can be postulated: given
the evidence, there can be no other conclusion. Thus, MUST represents
the strongest epistemic judgement one can make; but making the
strongest of all judgements is not the same as making a factual assertion.
TO BE ABOUT TO serves as a non-modal expression that in most
occurrences has an epistemic interpretation that runs parallel to epistemic
MUST. Consider:
(113) a. It’s bound to come out.
b. Self exploration and exploration in a small group at that
level of complexity and so on is bound, it seems to me, to
generate special languages.
However, there is a difference in the meaning of BE BOUND TO
and MUST. Consider (114) and (115) below:
(114) John’s bound to be in his office.
(115) John must be in his office.
Sentence (114) is the more certain of the two. It has little or no

sense of ‘logical conclusion’. In (115), the presence of MUST indicates
that the speaker draws the most obvious conclusion (because, for
instance, the lights were on in John’s office).
Besides BE BOUND TO, HAVE (GOT) TO can occur with
epistemic interpretation especially in American English as illustrated in
(116) while in British English MUST is used, as in (117):
(116) You’ve got to be joking.
(117) You must be joking.
183
8.0. WILL
There are four modal bases in which WILL can be evaluated. We
distinguish between volition WILL, power WILL, habitual WILL (root
meanings) and probability WILL (epistemic meaning).
Remember that the auxiliary verbs WILL and SHALL were full
lexical verbs in older stages of English (see the Category of Tense
sections 10.1. and 10.4.). SHALL meant ‘I must, I am under the
obligation’ and WILL meant ‘I want’. In the course of time these verbs
developed into auxiliary verbs that signaled both modality (i.e.,
possibility / probability) and futurity. In Modern English they retain both
a modal and a temporal value.
8.1. Volition WILL (root meaning)

The volitional use of WILL expresses want or wish.
Volition WILL has the following characteristics:
(i) it is used to express actual volition (i.e., with respect to the
present).
(ii) volition WILL is subject-oriented, i.e., the want rests with the
subject of the sentence.
(iii) in case a future time is intended, it is separately indicated in the
sentence (e.g., I will write tomorrow / We’ll celebrate this
tonight).
It has often been noticed that WILL expresses not only pure ‘want’
but also ‘determination, ‘resolve’ to carry something out. Thus, a
sentence such as:
(118) I will do that.
means “I am willing or determined to do that, and I shall do it”.
Volition WILL with first person subject may convey, function of
the context, a promise or a threat:
(119) a. You shall sign a statement…..I will bring it with me for
you signature, when I come again.
b. Marry me and I will save you life.
c. I will begin again.
184
The affirmative form you will is used as a request or order, most
often to a subordinate person.
(120) You will put the box into the van.
A question in the second person such as Will you? is the equivalent
of a request, that is, asking if someone is willing is taken as a request for
action:
(121) a. Will you remind me of this?
b. Will you dine with us on Saturday?
Contextually, will you….? is a mild form of an imperative.
Although it is a question about the hearer’s willingness, this question is in
fact used as a polite substitute for an imperative. The question can be
made even politer when accompanied by further markers of politeness:
Will you be good enough to…?, Will you kindly….?, Will you like to…?
In questions in the third person volition WILL is used emphatically:
(122) Poor Joe, why will he be so shy?
Volition WILL in the second and third person is most commonly
used in conditional clauses after if and unless. In these contexts, WILL
expresses willingness rather than determination:
(123) a. I shall be glad if he will come.
b. I shan’t be happy unless she will come.
The past time reference of WILL in reported speech is WOULD or
the past tense of be willing to:
(124) I asked him whether he would it / was willing to do it.
8.1.1. Hypothetical WOULD

WOULD is also used as a hypothetical marker of volition WILL
when it expresses formal or polite requests in the present or extended
present time:
(125) a. Would Mr. Smith come to the information desk?
b. Would you pass the dictionary, please?
185
These are not genuine questions asking for information about the
hearer’s will; rather, they are formal, polite requests for action undertaken
by the hearer.
Hypothetical volition WOULD is also used in the following
characteristic contexts:
(i) in constructions with the main verb pre-posed:
(126) a. Doors were shut upon him, go where he would.
b.….said John, losing his self-command, try as he would to
keep calm.
(ii) in negative contexts, WOULD NOT (WOULDN’T) is used as a
tentative refusal:
(127) a. He was angry and he would not go.
b. The moneylender said he would not renew the bill.
c. I told you, Pa, but you wouldn’t believe it.
(iii) in subordinate adverbial clauses of condition introduced by the
conjunction if and unless:
(128) a. If you would sit by me every night I shall work better.
b. I dare say she would prefer to go.
c. If she would prefer to go she would g.
(iv) in complement clauses after a verb expressing wish:
(129) a. I wish he would stop that noise.
b. I wish he would come soon.
c. I wish the snow would melt.
8.2. Power WILL (root meaning)

Power WILL is used with inanimate objects and ‘indicates how
such objects will characteristically behave’ (Palmer, 1979):
(130) a. The boat will hold only half of these.
b. Will the ice bear?
c. The hall will seat five hundred people.
Both volition WILL and power WILL are subject-oriented, i.e., the
source of ‘will’ or ‘power’ is seen as intrinsic to the subject of WILL.
186
For both uses of WILL discussed so far, the modal bases are
interpreted as someone/something being ‘positively disposed towards
something’ (Perkins, 1983). Volition WILL refers to the system of social
laws within which someone is disposed towards something while power
WILL refers to the system of natural laws in which something
characteristically behaves in a certain way.
The past tense reference of power WILL is conveyed by WOULD:
(131) I tired hard but the door wouldn’t open.

Hypothetical power WOULD is used with inanimate subjects in the
present or extended present:
(132) a. A house in London would cost a lot of money.
b. The hall would seat 100 persons.
8.3. Habitual WILL (root meaning)

Habitual WILL is used to state that, given certain empirical
circumstances, a situation regularly or occasionally takes place as a
consequence of a natural tendency of a person or object:
(133) a. A falling drop will hollow the stone.
b. Boys will be boys.
c. A cat will often play with the mouse before she kills it.
As in the case of power WILL the modal base of habitual WILL is
again the system of natural laws.
The past tense of habitual WILL is rendered by WOULD or USED TO
with the following difference in meaning: the form USED TO only correlates
with past time reference while WOULD is not thus restricted. Consider:
(134) a. I used to mow the lawn.
b. I would mow the lawn.
Sentence (134a) communicates a habitual activity in the past.
Sentence (134b) communicates a habitual activity in the past only if a
past time index is added (such as when I was young); otherwise iterativity
may also hold in the present.
187
8.4. Probability WILL (epistemic meaning)
In a sentence such as:
(135) You will be John’s daughter.
the circumstances are interpreted as evidence that the addressee is most
probably John’s daughter. Consider other examples:
(136) a. This will be the tower of London, I suppose.
b. Mother will be expecting me and will be getting uneasy.
c. That will be the postman.
The epistemic interpretation of WILL is naturally suited to
scientific and quasi-scientific statements:
(137) a. A lion will attack a man when hungry.
b. Truth will out.
c. Oil will float on water.
In (136) and (137) the inferences concern the present or extended
present time.
If the inference concerns a past time sphere, WILL plus the perfect
infinitive form is used:
(138) a. You will have received the box by that time.
b. He will have brought his mother back by that time.
c. It will have been perceived that old Lady Lufton had heard
nothing of Major Grantly’s offence.

The hypothetical epistemic form WOULD is more tentative than
probability WILL. Consider:
(139) a. How long would that take?
b. That would be in the year 1879.
c. That’s what a sensible man would do.
Its past time reference is rendered by WOULD plus the perfect
infinitive form:
(140) That’s what a sensible man would have done.

188
9.0. SHALL
The original meaning of SHALL, the Old English SCEAL, is
‘owe’. In Old English and Middle English, sceal took objects such as
money and tribute. During the period from 1370-1570 SHALL meant ‘I
must, am under obligation’, while WILL meant ‘I want, will’. In time this
distinction has very much eroded. In Modern English, SHALL is rarer
than WILL and it is largely restricted to first person interrogatives, at least
in informal English.
The modal verb SHALL has both a root meaning (obligation) and
an epistemic meaning (probability, prediction).
9.2. Obligation SHALL (root meaning)

The root/deontic sense of SHALL is obligation. It is used to
indicate that it is the will of the speaker/agent that imposes obligation:
(141) a. You shall never see me again.
b. Good dog, you shall have a bone when we get home.
c. You shall not go tonight. You shall not be excused if you go.
Depending on the larger context in which it occurs SHALL may
carry strong, undemocratic overtones of imperiousness:
(142) a. You shall obey my orders.
b. No one shall stop me.
c. She shall be mine.
With this meaning SHALL is felt as obsolete in Modern English. It
has survived in fairy-tales (where it is frequently on the lips of ogres,
wicked witches, jealous stepmothers), in the biblical Ten Commandments
(e.g., Thou shalt not kill), in legal and quasi-legal documents, in rules for
card games and academic dress:
(143) a. The first condition of legal justice is that it shall hold the
balance impartially.
b. A player who bids incorrectly shall forfeit fifty points.
c. The hood shall be of scarlet cloth, with a silk lining of the
color of the faculty.
189
In first person interrogative sentences, SHALL is used to inquire
about the wish of the addressee/hearer:
(144) a. Shall I go?
b. Sir, there is a gentleman below who wishes to see you.
Shall I show him into the parlor?
In second person interrogatives, shall you is used in situations that
are independent from the will of the person addressed (distinct from will
you, which inquires about the other person’s will / willingness):
(145) a. Shall you be in if I call in the afternoon?
b. When shall you do it?
c. Shall you stay in Paris?
In the second and third person statements, SHALL is interpreted as
a promise or threat on the part of the agent:
(146) a. If you hurt a hair of her head, you shall answer for it.
b. You shan’t escape.
9.2.1. Hypothetical SHOULD

SHOULD is not interpreted as the past tense of obligation SHALL
(Warner 1993, Papafragou 2000). SHOULD communicates a hypothetical
meaning, expressing obligation relative to existing norms or stereotypes.
Hypothetical SHOULD conveys obligation of a weaker type than
MUST:
(147) You should clean the place once in a while.
In such a context SHOULD expresses obligation with respect to
norms, normal course of events and its time sphere is present or extended
present.
Most probably SHOULD entered the modal system of English as a
separate individual item. However, it retains overtones of its original
meaning: I must, I am under obligation.
In Modern English, SHOULD is used as a hypothetical marker
with inverted word order or in the quasi-subjunctive construction and its
time sphere is present or extended present:
190
(148) a. Should you require any further assistance, please feel free
to contact us.
b. I do not desire that I should be left alone to the task.
If past time sphere is intended SHOULD plus the perfect infinitive
form is used:
(149) He should have been more kind to her at the time of their
reconciliation.
The sentence in (149) renders the ‘contrary-to-fact’ meaning of the
deontic, normative hypothetical SHOULD.
9.3. Probability, Prediction SHALL (epistemic meaning)

SHALL is interpreted epistemically when its modal base is the
system of rational laws and where the empirical evidence implies the
truth of the sentence:
(150) a. A flower shall produce thousands of seeds, of which
perhaps not one shall fall upon fertile ground and grow into a
fair plant.
b. Who touches pitch shall be defiled.
The general meaning of epistemic SHALL is that ‘someone /
something is disposed towards something’ (Perkins, 1983).
9.3.1. Hypothetical SHOULD

Epistemic SHOULD covers the notions of prediction, expectation
or probability and its time sphere is present or extended present. Consider
the following examples:
(151) a. That should be the plumber.
b. As I remember that should be the house.
c. By heaven, this should be my book.
For instance, the first example in (151) shows that the speaker
cannot be certain that the circumstances in the actual world guarantee that
the proposition p is true. The only evidence that he has for the embedded
proposition comes from his beliefs about the normal, expected course of
191
events. According to these beliefs, the plumber is expected to arrive some
time after the speaker called for him. So, if the circumstances in the real
world are as the speaker expects them to be, it follows that whoever is
ringing the bell is indeed the plumber.
If the time sphere is past, SHOULD and the perfect infinitive form is
used:
(152) a. The night should have turned wet since I came in, for he
had a large hat on.
b. In the prevailing temper of the public, the evidence should
have been very clear indeed to induce an ordinary English
jury to convict him.
In sum, although SHOULD was originally the past tense form of
SHALL, expressing past time reference, the situation has changed
considerably in Modern English. SHOULD occurs nowadays with
present time reference in both a deontic and epistemic interpretation. Its
sense is that of conveying something stereotypical or normative. “Its root
interpretations occur when normative assumptions are regarded as repre-
sentations of external states of affairs, whereas epistemic interpretations
arise when the expectation-conforming evidence is focused upon (as set
of internal propositional representations)” (Papafragou, 2000:232).
10.0. OUGHT TO
OUGHT TO has both a deontic meaning (moral obligation) and an
epistemic meaning (weak necessity).
On its deontic/root interpretation the modal base on which OUGHT
TO operates is that of ideal/morally commanded state of affairs.
(153) One ought to respect one’s country.
On this interpretation it is close in meaning to the use of SHOULD.
When used epistemically OUGHT TO conveys a sense of weak
necessity and it is weaker and more formal than MUST.
10.1. Moral Obligation OUGHT TO (root meaning)
OUGHT TO (whose semantics refers to ideals or moral obligation /
imperatives) resembles obligation MUST but with a slight difference in
meaning.
192
The obligation meaning of OUGHT TO, different from obligation
MUST, rather conveys lack of confidence in the fulfillment of the
situation described. “For example, if one says You must buy some new
shoes one assumes that the purchase will be carried out. But You ought to
buy some new shoes is a different matter – the speaker could well add
under his breath But I don’t know whether you will or not” (Leech
1971:157).
When deontic (obligation) OUGHT TO is used it is perceived as
more formal and polite than obligation conveyed by MUST and its time
sphere is present or extended present:
(154) a. You oughtn’t to pick your nose in public. (formal /
normative obligation)
b. You mustn’t pick your nose in public. (strong obligation /
interdiction)
10.2. Weak Necessity OUGHT TO (epistemic meaning)

On its epistemic interpretation OUGHT TO communicates a much
weaker sense of necessity than MUST. Consider the following contrast
(Leech 1971):
(155) a. This is where the treasure must be.
b. This is where the treasure ought to be.
Suppose the above sentences are uttered by a treasure seeker while
looking at a map. When saying (155a) he is optimistic and he is almost
sure he has found the treasure. When uttering (155b) the treasure seeker is
more cautious, acknowledging that there might be something wrong in
his calculations.
The time sphere of epistemic OUGHT TO is present or extended
present:
(156) Our guests ought to be home by now.
Past time reference is rendered by OUGHT TO and the perfect
infinitive form:
(157) We have left undone those things, which we ought to have
done, and we have done those things, we ought not to have
done.
193
11.0. NEED
The modal verb NEED has to be kept apart from the full lexical verb
NEED. The lexical verb NEED has a third person singular indicative
present tense form –s, it occurs with to and has non-finite forms. Consider:
(158) a. He needs to talk with you.
b. I needed another dress so I bought one.
In contrast, the modal verb NEED has the NICE properties (except
for coda and emphatic affirmation). As modal verb NEED is a suppletive
form of MUST.
NEED has both a deontic meaning (obligation) and an epistemic
meaning (necessity).
11.1. Obligation NEED (root meaning)

Modal NEED does not occur in affirmative contexts. It is either
used in questions or in negative sentences and supplies these forms for the
deontic/obligation MUST:
(159) a. I have no less than five morning dresses so that I need
never be seen twice in the same dress.
b. Watson need not spend a single evening alone if he didn’t
like to.
c. Need I tell my reader that…
There is a difference in interpretation between the modal form
NEEDN’T and the lexical verb form DON’T / DOESN’T NEED.
Consider the contrast below (Leech 1971):
(160) a. Lady P to her gardener: “The hedges needn’t be trimmed
this week, Smithern”.
b. (same situation): “The hedges don’t need to be trimmed”.
Sentence (160a) means ‘I don’t oblige you to trim the hedges this
week’ (i.e., negation of obligation MUST) while sentence (160b) means
‘the hedges have not grown enough to need trimming (i.e., negation of
the lexical verb NEED).
In interrogative sentences both NEED and MUST can be used but
with the difference in meaning as indicated below:
194
(161) a. Need I come tomorrow?
‘Is it necessary for me to come tomorrow?’
b. Must I come tomorrow?
‘Do you ask me to come tomorrow?’
As seen above the time sphere of NEED is present or extended present.
In indirect discourse NEED, like MUST, is maintained and
indicates a back-shifted event:
(162) a. She believed she need not fear any persecution.
b. He told her that she need not worry.
If past time reference is intended, NEED plus the perfect infinitive
form is used:
(163) a. She need not have been uneasy.
b. He made the cottage smaller than it need have been.
There is also a difference in interpretation between the negative
past tense forms of the two verbs NEED; their distinct interpretations are
indicated below:
(164) I didn’t need to go there. (so I stayed home)
(165) I needn’t have gone there. (but I went there nevertheless)
12.0. DARE
As in the case of NEED, there is both a modal verb and a lexical
verb DARE. They can be distinguished formally in that the lexical verb
DARE has non-finite forms, has an –s inflectional form for third person
present indicative and generally occurs with to.
Modal DARE has the NICE properties (except for coda and
emphatic affirmation). Compare:
(166) a. John dare not come. (modal DARE)
b. John doesn’t dare to come. (lexical DARE)
(167) a. Dare John come? (modal DARE)
b. Does John dare to come? (lexical DARE)
The lexical verb TO DARE means ‘to venture, to challenge’ while
the modal verb DARE conveys the meaning of ‘to have the courage or
impudence to do something’.
195
12.1. Hypothetical DARE
Modal DARE is always hypothetical and subject-oriented. Consider:
(168) a. How dare he judge?
b. I was so upset that I dare not even go upstairs and call Sissie.
In reported speech the form DARE is maintained and signals a
back-shifted event (see (168b)).
Consider also the following examples:
(169) a. She dare not tell him this is only an insurance policy.
b. You know you dare not have given the order if you
hadn’t seen us.
c. He dare not have done it if I had been with him.
As can be noticed the time sphere of DARE is present or extended
present as in (169a) while the past time reference is formed by DARE
plus the perfect infinitive form as in (169b) and (169c).
196
REFERENCES
Blackwell, Baltin, M. and Collins, C. eds., A Handbook of Contemporary

Syntactic Theory 2001.
Bach, E.1981, On Time, Tense and Aspect: An Essay in English Metaphysics, in
Cole, P. ed.
Baciu, I. 2004, Functional Categories in English, Bucharest University Press.
Barwise, J. and Perry, J. 1983, Situations and Attitudes, MIT Press.
Bennett, M. and Hall Partee, B. 1978, Towards the Logic of Tense and Aspect in
English, Bloomington, ms.
Bolinger, D, L 1989, Extrinsic Possibility and Intrinsic Potentiality, Journal of
Pragmatics, 13.
Borer, H. 2004, Structuring Sense. An Exo-Skeletal Trilogy, Book 1,
forthcoming.
Bybee, J. and Pagliuca, W. 1985, The Diachronic Dimension in Explanation, in
Explaining Language Universals, Hawkins (ed).
Bybee, J. and Fleischman, S. 1995, Modality in Grammar and Discourse: An Int
roductory Essay, in „Modality in Grammar and Discourse”, Amsterdam.
Carson, G.1977, A Unified Analysis of the English Bare Plural, in „Linguistics
and Philosophy” 1.
Chesterman, A. 1992, On Definiteness, Cambridge University Press.
Chiţoran, D. and Cornilescu, A. 1986, Elements of English Sentence Semantics,
Bucharest University Press.
Chomsky, N. 1981, Lectures on Government and Binding, Dordrecht, Foris.
Chomsky, N. 1995, The Minimalist Program, MIT Press.
Christophersen, P., The Articles. A Study of their Theory and Use in English,
Copenhagen.
Coates, J. 1983, The Semantics of Modal Auxiliaries, Croom Helm, London and
Canberra.
Comrie, B.1976, Aspect, Cambridge University Press, London.
Cornilescu, A. 2003, Complementation in English, Bucharest University Press.
Cornilescu, A.1995, Concepts of Modern Grammar, Bucharest University Press.
197
Crăiniceanu, I. 1996, The Category of Aspect in English and Romanian with
Special Reference to the Progressive Aspect, University of Craiova Press.
De Swart, H., Aspect Shift and Coercion, Natural Language and Linguistic
Theory 16.
Demirdache, H. and Etxtebarria, M. 2002, Temps et Aspect, Presses
Universitaires de Vincennes, Laca, B. ed.
Demirdache, H. and Etxtebarria, M. 2004, The Syntax of Time, MIT Press,
Guéron, J. and Locarme, J.eds.
Donnellan, K. 1966, Reference and Definite Descriptions, in Steinberg and
Jakobovits eds.
Dowty, D.1979, Word Meaning and Montague Grammar, Dordrecht.
Farkas, D.1985, Intensional Descriptions and the Romance Subjunctive Mood,
New York.
Fenn, P.1987, A Semantic and Pragmatic Explanation of the English Perfect,
Ph.D. Dissertation, Tübingen.
Filip, H.1999, The Category of Aspect, Cambridge University Press.
Fillmore, J. C.1971, Some Problems of Case Grammar, mimeo.
Giorgi, A. and Pianesi, F.1997, Tense and Aspect: From Semantics to
Morphosyntax, Oxford University Press.
Halliday, M.1970, Functional Diversity in Language as Seen from a
Consideration of Modality and Mood in English, in Foundations of
Language 6.
Hausser, R. 1974, Syntax and Semantics of Plural, in CLS 10.
Hawkins, J. A. 1987, Definiteness and Indefiniteness, Croom Helm, London.
Heim, I.1982, The Semantics of Definite and Indefinite Noun Phrases, Ph.D.
Dissertation, University of Mass., Amherst.
Hoffman, T. J.1976 , Past Tense Replacement and the Modal System, McCawley
(ed).
Hornstein, N.1982, The Study of Meaning in Natural Languages, in Hornstein
and Lightfoot, eds.
Huddlestone, R.1976, Past Tense Transposition in English, Journal of Linguistics
13.
Inoue, K.1979, An Analysis of the English Perfect, Linguistics, vol. 17-18.
Jackendoff, R. 1977, X-bar Syntax: A Study of Phrase Structure, MIT Press.
Jespersen, O. 1933, Essentials of English Grammar, in Allen and Unwin,
London.
Kamp, H and Reyle, K. 1993, From Discourse to Logic, Dordrecht: Kluwer
Academic Press.
Kaplan, D. 1973, On the Logic of Demonstratives, mimeo.
198
Kearns, K. S. 1999, The Semantics of the English Progressive, Ph.D.
Dissertation, MIT.
Kenny, A.1963, Action, Emotion and Will, Humanities Press Inc. New York.
Kim, J. B. 2004, Hybrid Agreement in English, Linguistics 42-6.
Klein, W. 1992, Time in Language, New York: Routledge.
Klinge, A. 1993, The English Modal Auxiliaries: From Lexical Semantics to
Utterance Interpretation, Blackwell, Oxford.
Kratzer, A. 1977, What ‘Must’ and ‘Can’ Must and Can Mean, Linguistics and
Philosophy 1.
Kratzer, A. 1981, The Notional Category of Modality, in Worlds, Words, and
Contexts, de Gruyter, Berlin.
Kratzer, A. 1991, Modality, in Semantics: An International Handbook of
Contemporary Research, de Gruyter, Berlin.
Kripke, S.1973, Meaning and Necessity, in Harman and Davidson eds.
Leech, G. N. 1971, Meaning and the English Verb, Longman.
Levinson, S. 1983, Pragmatics, Cambridge University Press, London.
McCawley, J. 1975, Syntax and Semantics, vol. 7, Academic Press, New York.
McCawley, J.1981, Everything that Linguists Always Wanted to Know about
Logic, The University Chicago Press.
McCoard, D. 1978, The English Perfect – Choice and Pragmatic Differences, in
Dik, Kooji, Thompson, eds., Amsterdam.
Mittwoch, A. 1988, Aspects of the English Aspect, Linguistics and Philosophy 11.
Mourelatos, A. 1978, Events, Processes and States, Linguistics and Philosophy 3.
Nunberg, G. 1976, Validating Pragmatic Explanations, in Cole, P. ed.
Palmer, F. R.1979, Modality and the English Modals, Longman, London.
Papafragou, A. 2000, Modality: Issues in the Semantics-Pragmatics Interface,
University of Pennsylvania, USA.
Perkins, M. 1983, Modal Expressions in English, Norwood, New Jersey.
Portner, P. 2005, What is Meaning? Fundamentals of Formal Semantics,
Blackwell
Poutsma, J. 1926, A Grammar of Late Modern English, vol. 1-4, Groningen,
Noordhoff.
Reichenbach, H. 1947, Elements of Symbolic Logic, the Macmillan Company.
Rothstein, S. 2004, Secondary Predication, Syncom, internet source.
Russell, B.1905, On Denoting, in Mind 14.
Ryle, G.1949, The Concept of Mind, New York.
Searle, J.1969, Speech Acts, Cambridge University Press.
Smith, C.1977, The Vagueness of Sentences in Isolation, in CLS 13.
Smith, C.1991, The Parameter of Aspect, Kluwer Academic Publishers,
Dordrecht.
199
Ştefănescu, I. 1988, English Morphology, vol. 2, TUB.
Strawson, P. F. 1950, On Referring, in „Mind” 59.
Sweet, H.1891, A New English Grammar, Oxford University Press.
Sweetser, E.1990, From Etymology to Pragmatics, Cambridge University Press.
Vendler, Z.1967, Verbs and Times, Linguistics and Philosophy 7.
Wechsler, S. and Zlatić, L. 2000, A Theory of Agreement and its Application to
Serbo-Croatian, „Language” 76.
200

An1 Elements of English Morphology Crainiceanu

Încărcat de

Informații document

Descriere originală:

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

An1 Elements of English Morphology Crainiceanu

Încărcat de

Drepturi de autor:

Formate disponibile

ILINCA CRĂINICEANU

ELEMENTS OF ENGLISH MORPHOLOGY

© Editura Fundaţiei România de Mâine, 2007

Redactor: Andreea DINU

Bun de tipar: 5.02.2007; Coli tipar: 12,5

Editura Fundaţiei România de Mâine

EDITURA FUNDAŢIEI ROMÂNIA DE MÂINE

1. ON LEXICAL AND GRAMMATICAL CATEGORIES

2. THE CATEGORY OF NUMBER

3. THE CATEGORY OF DETERMINATION

4. THE CATEGORY OF GENDER

5. THE CATEGORY OF CASE

6. THE CATEGORY OF ASPECT

7. THE CATEGORY OF TENSE

8. MODALITY AND MODAL VERBS

1.0. Introductory Remarks

1.1. What does Morphology study?

1.1.1. Derivational Morphology

1.1.2. Inflectional Morphology

Nearly all English nouns exhibit a number paradigmatic set of two

2.1. Lexical Categories vs. Functional Categories

3.1. The Lexicon

3.2. Syntactic Phrases

3.3. The Morpho-syntactic Nature of Functional Categories

2.0. Singular-Plural Opposition. Countable-Uncountable Distinction

1.2. Sortals vs. Non-Sortals

(1) a. Please, tell me isn’t there a ghost?

The most important semantic property of countable nouns is that

(2) a. What does he do for a living?

Languages like English (and Romanian) use classifiers for dividing

The conclusion is that the distinction [+/-count] is independent

a tin of sardines an alloy of copper and tin

(5) a lump/piece of sugar; a loaf/piece of bread; a sheet/piece of

It is clear that the classifier (i.e., the partitioning expression) can

3.1. Collective Nouns. Distributive and Collective Plurals

(a) Politics: assembly, air-force, cabinet, House of Commons,

3.1.1. Properties of Collective Nouns

3.1.2. Agreement Patterns

(10) Morph-syntactic agreement

Det noun verb…

3.1.4. Collective Nouns in the Plural

The examples in (11a, 11b) receive a distributive reading. This

4.0. Morphologically Defective Nouns

5.1. Number Recategorization of Mass Nouns into Countable Nouns

6.0. General Terms with No Corresponding Mass Lexicalization

7.0. Plurale Tantum Nouns

7.1.2. Names of certain physical or mental illnesses such as: creeps,

7.1.3. Nouns that name sciences such as: aesthetics, acoustics,

1.0. Classes of Determiners

(1) a. The boys will all/each/both/*some/*every get a prize.

2.0. Philosophical Background

3.0. Characteristics of Indexicals

3.1. Indexicals or Deictic Elements

4.0. The Values/Uses of Demonstrative Descriptions

4.1. The Deictic/Gestural Value

4.2. The Symbolic Value

4.3. Non-deictic Functions of Demonstratives

4.3.1. The Anaphoric Value

b. What looked like a white lace poncho covered him from

4.3.3. The Emotional Value

5.0. Definite Descriptions

5.2. The Values/Uses of Definite Descriptions

5.2.2. The Symbolic/Deictic Use of Definite Descriptions

(1) a. The boys will all/each/both/some/every get a prize.