○○○○○○○○○○●●●●○○○○○○○○○○○○○○○○○○○

○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○
○○ ●●●●○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○○○○
○○○ ●●●●○○○○○○○○○○○○○○ ●●●●○○○○○○○
○○○○○○○○○○○○ ●●●●○○○○○○○○ ●●●●○○○○○
○○○○○○ ●●●●○○○○○○○○○ ●●●●○○○○○○○○○
○○○○○○○○○○○○○○ ●●●●○○○○○○○○ ●●●●○○○
○○○ ●●●●○○○○○○○○○○○ ●●●●○○○○○○○○○○
○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○
○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○
○○○○○○○○○○○○○○○ ●●●●○○○○○○○○ ●●●●○○
○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○
○○○○○○○○○○○○○○○ ●●●●○○ ●●●●○○○○○○○
○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○
Articles in English
○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○ ●●●●
○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○
○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○
as a Second
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○
○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○
Language
○○○○○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○
○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○
●●●●○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○
A Phraseological
○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○
○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
Perspective
○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○
○ ●●●●○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○ ●●●●○○ ●●●●○○
○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○ ●●●●○
○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○
○ ●●●●○○○○○○○○○○○○○○○○○○○○○ ●●●●○○
○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
○○○
●●●
○○○
○○○
○○○
○○○
○ ●●
○○○
○○○
○○○
○○○
○ ●●
○○○
○○○
○○○
 ●●●● ●●●●
●●●●○○○○○○○○○ ●●●●○○○○○○○○○
○○○○○○○○○○○○○○ ●●●●○○○○○○○○ ●●●●○○○
○○○ ●●●●○○○○○○○○○○○ ●●●●○○○○○○○○○○
○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○
○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○
○○○○○○○○○○○○○○○ ●●●●○○○○○○○○ ●●●●○○
○○○○○○○○○○○○○○○○○○○
Justyna Leśniewska ●●●●○○○○○○○○○
○○○○○○○○○○○○○○○ ●●●●○○ ●●●●○○○○○○○
○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○
Articles in English ●●●●○○○○○○
○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○ ●●●●
○○○○○○○○○○○○○○○○○○○○○
as a Second ●●●●○○○○○○○
○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
Language
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○
A Phraseological ●●●●○○○
○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○
Perspective ●●●●○○
○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○
●●●●○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○
○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○
○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○○
○ ●●●●○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○ ●●●●○○ ●●●●○○
○○○○○○○○○○○○○○○○ ●●●●○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○ ●●●●○
○○○○○○○○○○○○○○○○○○○○○○
Jagiellonian
●●●●○○○○○○○
○ ●●●●○○○○○○○○○○○○○○○○○○○○○ ●●●●○○
University
Press
○○○○○○○○○○○○○○○○○○○○○○ ●●●●○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○○
Contents 5

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Chapter 1
Articles: Descriptive approaches . . . . . . . . . . . . . . . 15
1.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.2. The classification of articles as parts of speech and sentences . . . . 16
1.3. The indefinite article . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.4. The definite article . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.5. The zero article / bare noun phrase . . . . . . . . . . . . . . . . . . 26
1.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Chapter 2
Key concepts in the study of articles . . . . . . . . . . . 31
2.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2. Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3. Information flow . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.4. Countability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.4.1. The count / non-count distinction . . . . . . . . . . . . . . . 35
2.4.2. Implications for L2 acquisition . . . . . . . . . . . . . . . . 40
2.4.3. Some uses of articles are very difficult to explain . . . . . . . 41
2.5. Definiteness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6 Contents

Chapter 3
Other approaches to articles . . . . . . . . . . . . . . . . . 49
3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2. Socio-pragmatic approaches . . . . . . . . . . . . . . . . . . . . . 49
3.3. Cognitive accounts of definite article use . . . . . . . . . . . . . . 51
3.4. The generative tradition . . . . . . . . . . . . . . . . . . . . . . . 56
3.5. Corpus-based perspectives . . . . . . . . . . . . . . . . . . . . . . 58
3.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Chapter 4
Articles as a source of difficulty in SLA . . . . . . . . . 63
4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.2. Theoretical approaches to difficulty . . . . . . . . . . . . . . . . . 64
4.2.1. Difficulty as cognitive complexity . . . . . . . . . . . . . . . 66
4.2.2. Feature-related difficulty . . . . . . . . . . . . . . . . . . . . 68
4.2.3. Context-related difficulty . . . . . . . . . . . . . . . . . . . 68
4.2.4. Learner-related difficulty . . . . . . . . . . . . . . . . . . . 69
4.2.5. Some final words on objectivity . . . . . . . . . . . . . . . . 70
4.2.6. Articles and feature-related difficulty . . . . . . . . . . . . . 71
4.2.7. Salience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.2.8. Context- and learner-related difficulty . . . . . . . . . . . . 76
4.2.9. Language patterns versus metalinguistic
propositions and rule difficulty . . . . . . . . . . . . . . . . 78
4.3. Implicit versus explicit learning and knowledge . . . . . . . . . . . 79
4.4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.4.1. High difficulty level . . . . . . . . . . . . . . . . . . . . . . 87
4.4.2. Implicit learning . . . . . . . . . . . . . . . . . . . . . . . . 87
4.4.3. Limited usefulness of metalinguistic propositions . . . . . . 88

Chapter 5
Articles in SLA research . . . . . . . . . . . . . . . . . . . . . 91
5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.2. Crosslinguistic aspects . . . . . . . . . . . . . . . . . . . . . . . . 92
5.3. Semantic universals and the acquisition of articles . . . . . . . . . 94
5.4. The sequence of acquisition of articles in L1 and L2 English . . . . . 99
Contents 7

5.5. Studies on countability . . . . . . . . . . . . . . . . . . . . . . . . 103

5.6. Abstractness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.7. Articles and learners’ interim rules . . . . . . . . . . . . . . . . . 106
5.8. Type of task . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.9. Lexical chunks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.10. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Chapter 6
Articles and ESL teaching . . . . . . . . . . . . . . . . . . . . 111
6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
6.2. Articles and the efficacy of corrective feedback . . . . . . . . . . . 112
6.3. Other treatments and recommendations concerning the teaching
of articles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
6.4. The treatment of articles in teaching materials and pedagogical
grammars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

Chapter 7
Formulaicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.2 Speaker-external formulaicity . . . . . . . . . . . . . . . . . . . . 132
7.3 Speaker-internal formulaicity . . . . . . . . . . . . . . . . . . . . 136
7.3.1 The idiom principle . . . . . . . . . . . . . . . . . . . . . . 136
7.3.2 Theoretical support for the formulaic nature of language
processing . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
7.3.3 Empirical evidence supporting the formulaic nature
of language processing in native speakers . . . . . . . . . . . 142
7.4 Formulaicity, frequency and recent trends in linguistics . . . . . . 147
7.5 Formulaic language in L2 speakers . . . . . . . . . . . . . . . . . . 150
7.5.1 Speaker-external aspects . . . . . . . . . . . . . . . . . . . . 150
7.5.2 Speaker-internal aspects . . . . . . . . . . . . . . . . . . . . 151
7.6 Formulaicity and second language teaching: A brief history . . . . 159
7.7 Articles and formulaicity . . . . . . . . . . . . . . . . . . . . . . . 160
7.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
8 Contents

Chapter 8
Investigating article use by advanced Polish learners
of EFL: The role of formulaicity . . . . . . . . . . . . . . . 165
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
8.1.1 Purpose and rationale . . . . . . . . . . . . . . . . . . . . . 166
8.1.2 Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
8.2 Study 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
8.2.1 Participants . . . . . . . . . . . . . . . . . . . . . . . . . . 168
8.2.2 Instrument and procedure . . . . . . . . . . . . . . . . . . . 169
8.2.3 Analysis, results and discussion . . . . . . . . . . . . . . . . 174
8.3 Study 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
8.3.1 Participants . . . . . . . . . . . . . . . . . . . . . . . . . . 180
8.3.2 Instrument and procedure . . . . . . . . . . . . . . . . . . . 180
8.3.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . 190

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Appendix 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Introduction 9

Introduction

… article usage in English is complex and, in many

instances, arbitrary. There are several different uses
for each article, articles are often omitted, and there
are dialectal differences in the use of articles.
(Brinton & Brinton, 2010, p. 123)

… the English article system… appears to be virtually

impermeable to instruction.
(Gass & Selinker, 2008, p. 383).

The aim of this book is twofold: firstly, to provide an overview of

research findings on the acquisition and use of articles in English as
a second language (L2); secondly, to investigate articles in L2 English
from a phraseological perspective.
Learners of English as a second language are well known to expe-
rience numerous problems with mastering the article system. The
extent of the problem depends heavily on the mother tongue of the
learners and on other languages they may speak. Articles are par-
ticularly challenging for speakers of article-less languages such as
Polish, Russian, Mandarin or Korean. Research on second language
10 Introduction

acquisition (SLA), apart from confirming that problems with arti-

cles are strongly dependent on the first language (L1), also yields sev-
eral other insights: that learners’ problems are most pronounced at
the beginning stages of acquisition, that the indefinite article tends
to be acquired late, and that the definite article is often overused.
Also, there is evidence that, for some learners, articles remain a prob-
lematic area even at advanced – possibly even native-like – levels of
attainment.
No straightforward linguistic description can capture the func-
tioning of the article system in English. A number of syntactic, seman-
tic and pragmatic factors play a role in their use. Traditionally, arti-
cles are presented as elements of the morphosyntactic system of the
English language, which grammaticalizes the features of countability
and number. Equally important are semantic and pragmatic accounts
of the use of articles, since articles encode the notions of definiteness
and specificity, and play a major role in structuring information and
maintaining topic continuity in discourse. The central concept behind
the use of the definite article – definiteness – has a long history as
a topic of investigation not only in linguistics, but also in philosophy.
As a syntactic phenomenon, articles have been studied within the gen-
erative tradition as components of noun phrases (NPs) and even as
heads of determiner phrases (DPs). They have also been investigated
from a historical perspective, as well as in computational linguistics
studies, and in connection with natural language processing. There
is a vast collection of writings on articles that employ contrastive
and functional approaches, while the most recent work on articles
includes sociolinguistic approaches.
Given such a vast body of academic insights into the deceptively
modest-looking a and the, it remains a daunting, if not impossible,
task to try to offer a complete overview of the issues involved in the
study and description of the English articles. However, the first two
chapters of this book are intended to provide at least the essential,
Introduction 11

basic information on existing linguistic accounts of the English arti-

cles, in order to provide a descriptive framework for the rest of the
book and to set the scene for further discussion. Chapter 1 thus pro-
vides an overview of the traditional descriptive treatments of the arti-
cle, as they are found in grammars of the English language. The main
lines of linguistic inquiry into the English article system, which were
briefly mentioned above, are discussed in Chapter 2. This chapter pre-
sents an overview of the major aspects of article use that have been
investigated in long-standing approaches to the study of articles in
English, and the key notions in semantics and the philosophy of lan-
guage which are relevant to the study of articles. To complement the
first two chapters, Chapter 3 summarizes the way various approaches
to language use have conceptualized the nature and role of the English
article, underscoring the fact that none of them manages to account
for all possible uses made by speakers of that language.
The complexity of the English article system highlighted in the
first three chapters makes it obviously a difficult aspect of the lan-
guage for its learners. However, despite the uncontroversial character
of this observation, actually explaining the reasons for this state of
affairs is not a trite task. Since it is so commonplace to say that articles
are difficult for Polish learners of English, it seems worthwhile to take
a closer look at the concept of difficulty in second language research.
Chapter 4 applies this concept, as construed and operationalized in
the literature, to the English articles. The available theoretical frame-
works allow for a detailed analysis of reasons why articles are such
a stumbling block for many learners. The topic of pedagogical rules
and their difficulty is connected to the important topic of explicit
knowledge and explicit learning, as contrasted with implicit knowl-
edge and learning. Those concepts are also discussed in this chapter.
A more detailed overview of what is known about the acquisition
and use of articles in L2 English is provided in Chapter 5 of this book,
which discusses available research on this topic, bringing together
12 Introduction

findings from various publications, despite the very different method-

ologies and various limitations of the discussed studies which reduce
their generalizability. As will be seen, the body of research on arti-
cles in SLA revolves mostly around the nature and extent of L1 influ-
ence, the role of Universal Grammar, semantic universals, the time
course of development, and ultimate attainment. Since L2 learning
often takes place in instructional contexts, Chapter 6 complements
Chapter 5 by discussing the way articles are treated in the teaching
of English to speakers of other languages.
While articles have been examined from a number of angles,
there is a perspective which has not received much attention in SLA
research, namely the role of phraseology in the acquisition and use
of articles. While the correct use of articles by learners of L2 English
is traditionally seen as the result of the eventual mastery of grammar
rules, it stands to reason that it may also be to some extent colloca-
tional in nature, i.e., it may be aided by the storage and retrieval of
prefabricated and/or conventionalized language chunks (of various
kinds). However, phraseological phenomena are very complex. While
interest in all kinds of formulaicity-related issues has been growing
rapidly in recent years, there is still no clarity as to how chunking and
formulaicity might interact with grammatical development. This is
why Chapter 7 presents the trends and main findings on the role of
phraseological phenomena in language use and examines their rele-
vance to the case of the English articles.
Descriptive grammars of English, pedagogical materials, and
studies on learner language do include references to “idiomatic” or
“fixed” uses of articles. However, this category is invoked to account
only for those uses of the definite, indefinite or zero article which can-
not be explained by the “rules.” Consequently, formulaicity is sim-
ply a convenient label for those cases of article use which fall out-
side the rule-governed spectrum. My argument is that the interplay
between formulaicity in language and the use of articles is much more
Introduction 13

complex than just the existence of some odd uses of articles frozen in
idiomatic expressions. It is possible that the formulaic nature of lan-
guage to some extent contributes to correct article use. This possibil-
ity is explored by means of two studies presented in Chapter 8. The
findings of those studies indeed offer some support for the view that
frequency-driven conventionality in language may play a role in the
use of articles in L2 English.
1.1. Introduction 15

Chapter 1

Articles: Descriptive approaches

1.1. Introduction

Traditional descriptive treatments of articles are by far the most

widely used in language teaching. They inform a vast majority of
textbooks, pedagogical grammars, and teaching materials. They are
also important for studies on article acquisition and use in L2, though
in research there is a greater variety of theoretical approaches, with
a number of studies adopting the functionalist, generative, or cogni-
tive approaches, as will be seen in Chapter 3 of this book.
In order to establish a descriptive framework for the discussion
of articles, I will draw on a number of descriptive grammars of Eng-
lish. The two most often used sources will be referred to in an abbre-
viated form: the widely-used and influential comprehensive gram-
mar entitled The Longman Grammar of Spoken and Written English
(Biber, Johansson, Leech, Conrad, & Finegan, 1999) will henceforth
be referred to as Longman Grammar, while the Cambridge Grammar
of the English Language (Huddleston & Pullum, 2002) will be referred
to as Cambridge Grammar. A result of a large cooperative project, the
16 Chapter 1. Articles: Descriptive approaches

Longman Grammar was very innovative in its day, as it was based

on corpus-derived data. It remains a popular source of information
and terminology for researchers in the field of SLA and for devel-
opers of language teaching materials. Regarding terminology, the
Longman Grammar is mostly consistent with the classic Comprehen-
sive Grammar of the English Language (Quirk, Greenbaum, Leech,
& Svartvik, 1985). The Cambridge Grammar is another extensive and
detailed description of the English language system, which in many
ways departs from the Longman Grammar in terms of theory and
terminological aspects.

1.2. The classification of articles as parts

of speech and sentences

In descriptive grammars of English, articles are usually classified as

belonging to the larger category of determiners, which are function
words used to specify the reference of a noun (Longman Grammar,
p. 259). Determiners, in turn, are studied as part of the noun phrase.
Various descriptive grammars use different terminologies to pres-
ent the structure of the noun phrase (NP), or, as it is sometimes called,
noun group (Downing & Locke, 1992). Barring differences in termi-
nology, the structure of the NP in English is presented as consisting
of pre-head elements, the head, and post-head elements. Determiners
(and thus articles) figure among pre-heads and always precede other
pre-head elements.1

1 This only applies to languages like English, i.e., those that are categorized as
“head-first” languages in generative grammar. For languages that are “head-last,”
the sequencing is different.
1.2. The classification of articles as parts of speech and sentences 17

The structure of the noun phrase is typically presented in the fash-

ion illustrated below (Carter & McCarthy, 2006, p. 324):

pre-head head post-head

determiners premodifiers complements postmodifiers
a mother of three children
his recent claim that he was which was
innocent broadcast on state
radio and TV

While complements (as their name suggests) complete the meaning of the
noun phrase (as in the mother of three little children, a rise in interest rates),
postmodifiers are either relative clauses or are modifiers that can be par-
aphrased as relative clauses (the house nearby, the house which is nearby).
The various classifications show divergences when it comes to
post-head elements; for example, the noun phrase may be represented
as (Greenbaum, 1991):

[determiners] [pre-modifiers] noun [post-modifiers]

or as (Kreidler, 1998):

[determiners] [modifier] head [complement].

Some sources use other terms, e.g. modifier, for all pre-head elements,
and qualifier for post-head elements (Downing & Locke, 1992, p. 410),
others classify all post-heads as postmodifiers except for appositive
clauses (Greenbaum, 1991). Luckily, the first element of the noun
phrase, the determiner, is the one for which there is no terminolog-
ical confusion.
Even though specific typologies of determiners vary, generally
speaking, determiners typically specify whether the reference is defi-
nite or indefinite (the, a), possessive (my, her, etc.), demonstrative (this,
18 Chapter 1. Articles: Descriptive approaches

those, etc.), or they specify number or quantity (three, first) (Carter

& McCarthy, 2006, p. 335).
For example, the Longman Grammar and Greenbaum (1991) divide
determiners into the following categories, on the basis of their posi-
tion inside the NP:
• pre-determiners (e.g. all, both, half; multipliers, e.g.: double,
once, twice);
• central determiners: articles (the, a[n]), demonstrative deter-
miners (e.g. this, these), possessive determiners (e.g. his, her);
• post-determiners: ordinal numerals, semi-determiners (same,
other, former, latter, last, next), cardinal numerals, and quan-
tifying determiners.

As Kreidler (1998) points out, the typical classification of determiners

is essentially semantic in nature, as illustrated, for example, by the
categories of demonstrative determiners (e.g. this, these), possessive
determiners (e.g. his, her), quantifiers: specific (e.g. five), general (e.g.
much, several), collective (e.g. all), distributive (e.g. every).
The Cambridge Grammar, in turn, makes a useful distinction
between determiners and determinatives. A determiner is a func-
tion in the NP structure, whereas a determinative is a lexical cate-
gory; a distinct, primary category. While it is true that the function
of the determiner is usually realized by a determinative (e.g. the, a,
those) or a determinative phrase (e.g. too many, almost all), it can be
realized by a genitive noun phrase (as in the minister’s speech). Deter-
minatives, in turn, do not always function as determiners, they can
also function in other roles, for example as modifiers. The following
types of determiners are distinguished (Cambridge Grammar, p. 355):

1. basic determiners:
• determinatives: the tie, a book, those shoes
• determinative phrases: almost every tie, at least two shoes
1.3. The indefinite article 19

2. subject-determiners:
• genitive NPs: my tie, the boy’s shoes

3. minor determiners:
• plain NPs: what colour tie, this size shoes, Sunday morning,
tomorrow evening
• preposition phrases: over thirty ties, up to thirty shoes, from
ten to fifteen judges.

As the name suggests, the determinatives are the basic type of deter-
miners. The Cambridge Grammar distinguishes 14 main categories
of determinatives (for example, personal, universal, distributive, dis-
junctive, etc.), of which articles are the first one (see the Cambridge
Grammar, p. 356, for the complete list).2

1.3. The indefinite article

The English indefinite article a (or an in the liaison form) is found

only in a determiner function. The basic grammatical category rele-
vant to the use of the indefinite article is number, since number cor-
responds to the semantic concept of countability, and the indefinite
article can occur only with singular countable nouns. The two major
grammatical types of nouns, count (or countable) and non-count
(or non-countable) differ with respect to whether they can express
contrast of number (Carter & McCarthy, 2006, p. 335). The most

2 As can be seen, traditional or descriptive approaches consider the article to

be a component of the NP. However, as will be seen in Chapter 3, generative
approaches in syntax posit the existence of DPs (determiner phrases) in which
the determiner is the head, thus highlighting the importance of determiners.
20 Chapter 1. Articles: Descriptive approaches

common approach in descriptive grammars is to treat the count/

non-count distinction as based on the syntactic behaviour of nouns.
As a consequence, the test for whether a noun is countable or not is
its occurrence with cardinal numbers (Cambridge Grammar, p. 334).
The distinction is, of course, to some extent motivated semantically,
since the general principle is that count nouns denote entities which
are treated as units, while non-count nouns denote entities which are
treated as non-divisible, or that do not have clear boundaries (Carter
& McCarthy, 2006, p. 335). However, this distinction cannot be made
in a predictable manner on the basis of the physical properties of the
entities denoted by nouns. This issue will be discussed in more depth
in Chapter 2, section 4.
The indefinite article is the most basic indicator of indefinite-
ness, which can be understood as lack of identifiability: “with indef-
inite NPs, the addressee is not being expected to identify anything”
(Cambridge Grammar, p. 371). The Cambridge Grammar (p. 372) dis-
tinguishes between two kinds of indefiniteness: quantitative and
non-quantitative. Quantitative uses, for example in I bought a new
car, a cat just came into my house, express existential quantification
by indicating reference to one item. This use results from its histori-
cal origin as an unstressed variant of the numeral one (Hewson, 1964;
Mitchell & Robinson, 1992). Non-quantitative uses, as in Jill is a doc-
tor, where Jill is characterized as belonging to the class of doctors,
indicate simple set membership. Generic uses of the determiner a are
also always non-quantitative (the lion is a ferocious beast), and so are
uses of a with count nouns (all examples in this passage are from the
Cambridge Grammar). In many sources, the non-quantitative uses
are labelled simply as “classification of an entity” and “generic use”
(e.g. in the Longman Grammar, p. 260).
It is important to note that the English indefinite article encodes
indefiniteness, not lack of specificity, so it can be used to introduce
new entities (not previously mentioned, not identifiable to the listener)
1.3. The indefinite article 21

that are either specific or non-specific, as the following examples from

Downing & Locke (1992, p. 429) illustrate:
• a new non-specific entity: I need a new car;
• a new specific entity: I’ve bought a new car.

References which follow the occurrence of a noun referring to a spe-

cific entity with an indefinite article usually take the form of definite
noun phrases or personal pronouns:

A cat was the victim of a cruel attack when she was shot in the neck
by a pellet. (…) The pellet went right through the cat’s neck and came
out on the other side.
(Longman Grammar, p. 260)

Most of the complications in the use of the indefinite article arise

from the complexities of the countability-related characteristics of
nouns in English; for example, Downing and Locke (1992, pp. 422–
423) enumerate six main different types of countability. This issue will
be discussed further in Chapter 2, section 4. However, there are also
some intricacies related to the notion of indefiniteness; for example,
it is possible to use the indefinite article with proper nouns, which
are definite by default, as in (examples from Downing & Locke, 1992,
p. 430):

Is there a John Smith is this class?

I can’t imagine a Greece without an Athens.
I’ve got a Goya at home, but it’s only a reproduction.

In such cases, the nouns refer to entities which are treated as members
of a class, or “something abstract or imaginary is predicated” about
them, or they are used as metonyms.
22 Chapter 1. Articles: Descriptive approaches

1.4. The definite article

As its name suggests, the definite article’s main role is to indicate defi-
nite entities, i.e., ones that are known, familiar, or identifiable to the
speaker and the hearer. This is a very general statement; the notion of
definiteness is complicated and a topic of much discussion in various
schools of linguistics, as will be explained in the next chapter. This
section will present only the most basic facts about the definite article.
The definite article is compatible with all types of common nouns:
count singular, count plural, and non-count, as illustrated by the fol-
lowing utterances:

I’m getting the screwdriver.

I’m getting the shelves.
I’m getting the paint.

Specific types of uses of the are variously classified by different

sources; however, by putting together a number of accounts offered by
different descriptive grammars, we arrive at a set of roles that are most
often played by the definite article. Before they are listed, it should be
noted that a useful distinction is made by Downing and Locke (1992,
p. 431) between definiteness which is identified endophorically (i.e.,
within the text) and exophorically (i.e., outside the text). Three of the
common functions of the definite article, listed below, are exophoric,
i.e., their definiteness results from what is known about the outside
world, and the speakers’ knowledge:

1. The referent being unique in the wider discourse (the sun, the sky).

2. The uniqueness of a specific entity in a given situation, for a spe-

cific speaker and a specific hearer, who share the non-linguistic
1.4. The definite article 23

context. In the sentence The President has been assassinated, the

determined noun phrase “the President” is most likely identifia-
ble because of the specific context in which this message is being
communicated. With an utterance like Be quiet, or you’ll wake up
the baby!, the interlocutors know which baby may be woken up.

3. The deictic role of the article (dependent on the situational con-

text). In sentences like Put the vase on the table, Could you do some-
thing about the noise?, the referent is assumed to be identifiable on
the basis of the physical or social context.

Other commonly recognized roles of the definite article, listed below,

appear to be endophoric:

1. Indicating the identifiability of referents denoted by nouns which

are modified, either by means of pre-modification (the red dress),
or by means of post-modification or complementation (the blue
t-shirt at the bottom of the pile, the man who was the last one to exit
the station). The latter type of identification is sometimes called
“cataphoric” (i.e., forward-looking), since the modifying element
comes after the noun (e.g. Cambridge Grammar).

2. The anaphoric (i.e., “backward-pointing”) role of the definite

article: the entity is identifiable because it has been mentioned
previously in the same discourse (e.g. Brinton & Brinton, 2010,
p. 123). In other words, it could be called the rule of prior men-
tion: the referent of the noun phrase is assumed to be known to the
speaker because of the preceding reference. This kind of reference
is sometimes labelled as direct anaphoric reference, as it involves
the repetition of the same noun head, but with the definite article.
(We were approached by a young girl and an old woman; The girl
looked distressed.) However, as will be seen later in this book, this
24 Chapter 1. Articles: Descriptive approaches

description of article use is not entirely sufficient, as immediate

anaphoric reference is in many cases more likely to be expressed
by a pronoun (for a discussion of the as an indicator of a low level
of accessibility, see section 3.3).

3. Indirect anaphoric references. Other types of anaphoric references

are variously classified. I will use the label “indirect anaphoric ref-
erences” (e.g. Kreidler, 1998, p. 148) as it seems to make the most
sense when applied to uses like the one below, which involves
synonymy:

I saw a strange picture on the wall. The painting was of the most un­usual
kind.

Likewise, the indirect anaphoric link between co-referential nom-

inals may be based on other semantic relationships, for example mer-
onymy or hyperonymy (Halliday & Hasan, 1976, p. 278). There is also
the special case of references which some sources label “bridging ref-
erences” (Clark 1977, cited in Saeed, 2003, p. 203). They can be illus-
trated with the following examples:

The car wouldn’t start. The battery was flat.

I had two hours to kill before the conference began, so I went out for
a walk. The park was beautiful.

The above examples illustrate how the listener is expected to create

links with the preceding sentences by referring to general knowledge.
The assumption is that the hearer can make the necessary implica-
ture to relate a new reference to a previous one and establish coher-
ence. In the first example, both the car and the battery are men-
tioned in the text, but exophoric knowledge about the battery as part
of a car is required to understand the connection between the two.
1.4. The definite article 25

This phenomenon of bridging will be discussed in more detail in sec-

tion 3.3, which is devoted to cognitive treatments of articles.
The concept behind the above uses of the definite article is labelled
as identifiability by the authors of the Cambridge Grammar. Identi-
fiability occurs when only one relevant entity is assumed to satisfy
the description expressed in the noun phrase. This identifiability of
the referent may be established through a number of different means;
for example, identifiability by virtue of uniqueness characterizes the
following utterances:

Where did you put the keys?

Where did you put the milk?

They are identifiable because the speaker has in mind a certain spe-
cific set of keys, a specific carton of milk.
Contrary to the indefinite article, which occurs only as a deter-
miner, the definite article has yet another syntactic function (beside
that of a determiner): a modifier in constructions with superlatives
and comparatives (Cambridge Grammar, p. 371).
The Longman Grammar also introduces the category of “idio-
matic” uses of the definite article, but it is the briefest of mentions.
Finally, the Cambridge Grammar also mentions that the to some
extent also expresses totality, which means that it indicates the entire
set or quantity of the items or substance which are referred to (even
though the concept of totality is somewhat weaker than that expressed
by universal quantification).
26 Chapter 1. Articles: Descriptive approaches

1.5. The zero article / bare noun phrase

Most traditional descriptive grammars of English use the term zero

article to refer to the absence of an article before a noun. For example,
the Longman Grammar defines the zero article as one that occurs with
uncountable nouns in contexts corresponding to the use of the indef-
inite article with countable nouns, as well as with plural countable
nouns. However, the Cambridge Grammar and some other grammars,
for example, The Collins Cobuild English Grammar (Sinclair, 1990) do
not include the zero article in their descriptions of the article system.3
Rather, they speak of conditions where the nominal (a unit interme-
diate between a noun phrase and a noun) can head the noun phrase
without a determiner. In the Cambridge Grammar, the noun phrases
containing a determiner are called determined noun phrases, while
bare noun phrases are formed, under certain circumstances, by nomi-
nals in the absence of a determiner. Nominals headed by plural count
nouns, or by non-count nouns can freely be admitted as indefinite by
default. The examples from the Cambridge Grammar are as follows:

New cars were stolen. – bare noun phrase (indefinite by default)

Two new cars were stolen. – determined indefinite noun phrase
The new cars / Ally’s new cars were stolen. – determined definite noun
phrase.

At first it may seem that the use or non-use of the term “zero arti-
cle” is just a matter of terminology. However, as Berezowski (2009)
explains, the concept of zero article is rooted in historical and struc-
tural linguistics. The term came into being as an explanation for why

3 In fact, the concept of the zero article is in itself problematic and not universally
recognized by linguists (see Berezowski, 2009).
1.5. The zero article / bare noun phrase 27

there is no article with some nouns, even though the use of articles
in English is generally obligatory. The missing article was conceived
of as existing but realized without an overt form. Berezowski argues
that the situation is different: the English article system has not fully
evolved, and articles are not used in all grammatical environments
(2009). I will use the term zero article for the sake of convenience,
simply to denote a missing article, as this kind of use seems to be the
most common in the literature, also in language teaching.
In the case of singular count nouns, the inclusion of a determiner
is generally obligatory. However, there are special cases of nouns
which otherwise behave like ordinary countable nouns, but which
in some phrases appear within a bare noun phrase; that is, they take
the zero article. This happens, for example, to names of meals when
they are used in a general sense, and names of places/institutions
such as school, hospital, jail, etc. In the sets of sentences below, this
use is contrasted with the ordinary, countable use of the same nouns.

We are going out for dinner. The dinner was delicious.

He is in hospital, awaiting surgery. The hospital is located on the

outskirts of the city.
The surgery lasted two hours.

The ceremony took place in church. The church is known for its
beautiful altar.

I worked in the evenings while The university is building a new

I was at university. campus.

She knew she would go to jail if The car was parked outside the jail.
they caught her.

The zero article is also used with means of transport and with rela-
tion to communication used in a general sense. This use, restricted
to prepositional phrases opening with by, affects nouns that are both
countable and uncountable: go by bus/car/coach/plane/sea/taxi/train,
28 Chapter 1. Articles: Descriptive approaches

travel by air/car/horse/train, contact by radio/telephone, send by mail/

post. (But: she took the train to Warsaw, it was in the mail, the tele-
phone was black). Other typical uses of the zero article include times
of day (at dawn, at night (but: she waited for the dawn, it was in the
middle of the night), days of the week (unless there is postmodifica-
tion), months, seasons, and vocatives.
Another context in which the zero article is used is with some
predicatives that have a unique reference. While English normally
requires an article with singular countable nouns in a predicative
position, the use of predicative noun phrases which name a unique
role or position is quite common with the zero article, especially with
the nouns president and chairman. Biber et al. found that uses such
as he was re-elected OPEC president in November, with the nouns
chairman and president, outnumbered the chairman / the director by
about 5 to 1 (1999, p. 262).
Another use of the zero article which is extremely interesting is
its occurrence in parallel structures (between lawyer and client, father
to son, door to door). The Longman Grammar makes the following
comment about this phenomenon: “as these nouns are countable, we
would normally expect an article. Examples of this kind are often fro-
zen idiomatic expressions, as in from start to finish, from time to time,
eye to eye, face to face” (p. 263). Consequently, we are dealing here with
a set of exceptions to a rule. Those exceptions are cases where the use
occurs in fixed idiomatic elements of language. In other words, the
production of an expression in its entirety, as a fixed phrase, over-
rides the need to insert an article where it should normally appear.
We could treat all the manifestations of this phenomenon as nothing
more than cases of idiomatic expressions containing usage which is
not usual outside those idiomatic expressions. But there is more to
this phenomenon: there is a pattern here which may be productive,
in that when completely new structures are generated, one may say,
for example, I made the repairs, armchair by armchair.
1.6. Conclusions 29

There is also a special use of the zero article which is completely

genre-dependent, namely, its use in newspaper headlines, notices,
and user guides, which is what the Longman Grammar calls “block
language”: Cat finds treasure, insert batteries here.

1.6. Conclusions

Some conclusions can be drawn from the above review. Firstly, the
descriptions of the article system in English are extremely extensive
and complicated. Secondly, there are very substantial differences
between the available descriptive grammars in terms of taxonomy
and terminology. Thirdly, all the rules abound in exceptions.
From a language learning perspective, the fact that there are so
many different possible explanations of why a or the is used in a spe-
cific context is already very challenging. To make matters worse,
none of the different explanations may make sense in a specific con-
text. For example, let us consider the last definite article of the fol-
lowing sentence: The insurgent was spotted looking out the window.
The referent “window” lacks definiteness, and the use of a definite
article cannot be accounted exophorically or endophorically: the
referent is not unique, even in that specific context (the building
where the insurgent was may have had several windows), the serves
no deictic purpose, the window is not readily identifiable, there is
no direct or indirect anaphoric reference involved, no bridging ref-
erence is made possible, and reference cannot be made by means of
meronymy or hyperonymy.
Another issue is the fact that the numerous exceptions to the rules
are usually assumed to be idiomatic, in the sense of being fixed over
time through continuous use. However, such “conventional” uses may
nevertheless be productive to some extent, as for example, the use of
30 Chapter 1. Articles: Descriptive approaches

the zero article in parallel structures, which enables one to say, for
example, the technicians removed the virus laptop by laptop.
Above all, several facts suggest that the way article usage has tra-
ditionally been described in grammar books may be inefficient or
insufficient to account for the full range of article uses. The follow-
ing chapters will move beyond descriptive approaches to see if other
traditions have succeeded in capturing the elusive nature of the rules
that govern the use of articles.
2.1. Introduction 31

Chapter 2

Key concepts
in the study of articles

2.1. Introduction

The previous chapter has shown that even basic descriptive accounts
of article use are very complex, with seemingly more exceptions than
regularities. This chapter attempts to answer the question of why this
is the case, by looking first at some crucial concepts which under-
lie article use, namely reference, information flow, countability and
definiteness.

2.2. Reference

The use of articles is inextricably related to the notion of reference.

A basic concept in semantics is the difference between a referring
expression and its referent. The latter is a certain entity outside of
the realm of language, and the former is the expression used to refer
32 Chapter 2. Key concepts in the study of articles

to that entity (J. Lyons, 1977, p. 177).4 The difference between the
two is the source of humour in the following joke and the cause of
the oddity of the following sentence (examples from Kreidler, 1998,
p. 131): Question: Where can you always find sympathy? Answer: In
a dictionary, Washington has three syllables and 600,000 inhabitants,
or in the saying The dictionary is the only place where success comes
before work.
Referents can be classified on the basis of a number of character-
istics. One basic distinction is between fixed (also called constant)
referents and variable referents.5 Fixed reference occurs when the ref-
erent is a universally unique entity (independently of context), as in
Lake Erie. Variable reference characterizes those referring expressions
whose referents may be different every time they are used, as in we
swam in a lake. The referent of a noun phrase which has variable ref-
erence is established on the basis of a variety of sources of information,
including context (both physical and linguistic), general knowledge,
etc. However, it is possible for nouns with fixed reference to be used
with variable reference, as in these examples: every city has a Green-
wich village, this fellow is an Einstein, no Shakespeare wrote this play
(Kreidler, 1998, p. 135).
In fact, the fixedness of reference appears not to be absolute: even
with expressions that are normally classified as having a fixed refer-
ence, for example, the Eiffel Tower, it is possible for that expression in
a sentence, for example, look at the Eiffel Tower to refer to a model of

4 Saeed (2003, p. 26) introduces a different definition of “referring expression.”

To him, a referring expression is a nominal (name or noun phrase) which has
a non-generic interpretation. Generic uses are thus, confusingly, labelled as
non-referring.
5 Considerable confusion results from the fact that the terms “unique” / “non-
unique” are sometimes used to refer to this distinction. I find this extremely
problematic, because uniqueness is also used interchangeably with specificity
by some writers.
2.2. Reference 33

the famous tower, or a piece of jewellery that is shaped to resemble it,

or to the half-size copy that stands in Las Vegas.
Another important distinction exists between specific and
non-specific reference. In a sentence such as I have a cat, reference
is made to a specific cat. On the other hand, in a sentence like I’d love
to have a cat, the cat in question is most likely non-specific. Under
some circumstances, a referring expression can even have two ref-
erents, one that is specific and one that is non-specific. An example
of that possibility was imagined by Donnellan (1968), who hypoth-
esized someone saying: Smith’s murderer is insane. One interpre-
tation is that of a non-specific referent: the speaker does not know
who killed Smith, but in his opinion the person was insane. Another
possible interpretation involves a specific referent: the murderer can
be a specific person whose identity is known and who is known to
be insane.
Finally, it is possible to contrast generic reference with non-ge-
neric reference, as illustrated by the following pairs of sentences (some
examples taken from Kreidler 1998, p. 141):

A dog is a fine pet. / A dog is lying in the middle of the street.

Dogs are fine pets. / I can hear dogs barking in the distance.
The dog was man’s first domestic animal. / The dog really frightened me.

Generic reference is usually expressed by means of the zero article,

but it can be expressed also through the use of a singular countable
noun with either the definite or the indefinite article, as shown in the
examples above. It is interesting to note that corpus analyses reveal
generic uses to be relatively uncommon (less than 5% of all article
uses, according to the Longman Grammar).
34 Chapter 2. Key concepts in the study of articles

2.3. Information flow

The way in which articles are used in general terms depends on the
information flow in discourse. Low (2005, cited in Głaz, 2012) makes
a useful distinction between anaphoric-oriented and accessibility-ori-
ented approaches to definiteness. Anaphoric-oriented approaches are
centred on discourse: the motivation of the definite article is to be
found in the preceding discourse. These approaches are very impor-
tant in the structuring of the information flow in discourse. Acces-
sibility-oriented approaches, for their part, are based on the notions
of familiarity (or identifiability), with the crucial feature being the
extent to which the referent is assumed to be known to the hearer.
In any clause, there are elements which present new information
(“the new”), and elements which present information available from
the preceding discourse (“the given”). The distribution of articles in
corpora reflects what the Longman Grammar calls the “information
principle”: there is a preferred distribution of information in a clause,
“corresponding to a gradual rise in information load” (p. 896). A con-
sequence of that principle is that it is more usual for clauses to open
with given information and continue to reveal new information.
Adherence to the information principle contributes to the cohesion
of a text. Reliance on this principle speeds up language production
for the speaker and facilitates the processing of speech for the listener.
The information principle is also reflected in the distribution of
articles in texts. According to the Longman Grammar (p. 269), across
registers, definite articles are much more common in noun phrases
which are in the subject position or which are the complement or
object of a preposition than in phrases which are in the object posi-
tion. For indefinite articles, the situation is reversed, which is in keep-
ing with the fact that the object comes in the clause that is typically
associated with new information. This kind of evidence provided by
2.4. Countability 35

analyses of text corpora confirms that the distinction between the

given and the new is indeed a guiding principle behind the structur-
ing of discourse. However, there are always numerous exceptions to
this rule, as will be seen in the section devoted to the treatment of
articles in cognitive linguistics.
Of importance to the issues discussed in this book is the fact that
while definiteness as a semantic and pragmatic category is closely
related to the definite article in languages with article systems, it has
been observed that article-less languages have other means of express-
ing definiteness, ranging from information flow and structure to word
order and the use of some nominal and verbal categories, such as
aspect (for an in-depth treatment of this topic, see C. Lyons, 1999).
Firbas (1992) develops the theory of functional sentence perspective
(FSP), which originated in the Prague school of linguistics, in order
to explain how the semantics, grammar, the context of a sentence, as
well as prosodic aspects, all contribute to a certain “orientation” or
“perspective” of the sentence, in which some elements are more prom-
inent than others. For Polish, this issue was investigated by Szwedek
(1973), who showed how Polish displays a tendency to mark the defi-
niteness of nouns by means of their sentence-initial placement. The
positioning of the noun in mid-sentence or at the end is often inter-
preted as conveying indefiniteness.

2.4. Countability

2.4.1. The count / non-count distinction

Countability is a concept of high importance for the use of the indef-

inite article. As was said above, the use of articles depends on the dis-
tinction between countable (also called “count”) and uncountable
36 Chapter 2. Key concepts in the study of articles

(also called “non-count” or “mass”) nouns. It is possible to describe

all nouns in the English language in terms of their countability. This
distinction appears quite unproblematic at first; Palmer called it “a
fairly clear one” (1981, p. 127).
However, the count/non-count distinction, which is reflected in
the syntax and semantics of nouns, verbs, adjectives, and quantifi-
ers, is characterized by notable crosslinguistic variation and has been
the subject of much discussion in linguistics and in the philosophy
of language. It is interesting to look at how this seemingly unprob-
lematic distinction is at the heart of a major debate about some of the
core beliefs concerning the nature of the relationship between lan-
guage and reality.
Some linguists do not differentiate between count and non-count
nouns because “nouns display a range of ‘countability’” (Kreidler,
1998, p. 137). Kreidler, for example, speaks of the “non-countable end
of the continuum” (1998, p. 138). This view is problematic in that in
any particular instance of use, a noun is either treated as countable or
not, so from a grammatical point of view the distinction is a dichot-
omy. A grammar-based distinction between the two classes is pre-
sented in Palmer (1981, p. 126):

Formally the two classes are easily distinguished. Count nouns alone
may occur in the singular with the indefinite article a, while only mass
nouns may occur with no article or with the indefinite quantifier some.

Palmer thus treats the indefinite article as actually defining what

is countable or what is not countable, because the distinction between
count/mass in English can be made on the basis of what article a word
can be used with: those lexemes which take the indefinite article are
countable. This reasoning is somewhat circular, as the use of arti-
cles is usually considered to be dependent on the count/mass status
of nouns.
2.4. Countability 37

He argues strongly that the membership of nouns in the count or

non-count categories is arbitrary:

The semantic difference between these two classes is clear enough.

The count nouns “individuate” – they indicate individual specimens,
while the mass nouns refer to a quantity that is not individuated
in this way. But the distinction does not correspond closely to any
semantic distinction in the world of experience, and this should be
no cause for surprise. (…) But there is no explanation in semantic
terms why we can refer to a single mass of jelly as a jelly but not to
a mass of butter as a butter.
(Palmer, 1981, p. 126)

To Palmer, whether a noun belongs to both the count and the

mass category is arbitrary and unpredictable. He gives the example
of cake, which can be count as well as non-count, contrasted with
bread, which is only non-count, even though you can get separate
loaves of it. “A foreigner,” writes Palmer, in one of very rare mentions
of the foreign-language perspective, “could not guess, then, whether
such words as soap, trifle, cheese, would be count nouns in English.
He has, moreover, to learn the “individuating” nouns: loaf of bread,
cake of soap, pat of butter” (Palmer, 1981, p. 127).
Thus, in Palmer’s view, countability and uncountability are gram-
matical categories. While they tend to correspond to real-word prop-
erties of entities, they do not always do, as can be shown by both the
differences in the distribution of the count/mass distinction across
languages, but also by the inconsistencies within the same language
(wheat vs. oats, for example).
While it seems that some nouns may belong to both categories,
Palmer treats such nouns as belonging either to the “mass” cate-
gory, while the uses of these nouns with an indefinite article are sim-
ply a kind of individuation that can be applied to them for specific
38 Chapter 2. Key concepts in the study of articles

purposes, for example, to denote a specific kind, or, with liquids,

a familiar amount (as in a coffee meaning a cup of coffee), or to the
“count” category, for example in the case of animals, can sometimes
be used in the non-count manner, indicating meat (therefore the
meaning of could I have some cat is clear, even though it is not cus-
tomary to eat cat meat).
In the Semantics of Grammar, Wierzbicka (1988) includes an inter-
esting discussion of countability, offering an argument against the
axiom that the relationship between the count status of nouns and
its meaning is arbitrary. She argues that form-classes are semanti-
cally motivated, and that “differences in grammatical behaviour reflect
iconically differences in meaning” (1988, p. 501). Often referring to
various languages, she argues that there always is an underlying jus-
tification behind the count status of nouns in a language. Brussels
sprouts are countable because their size is just right, watermelon is
too big, rice – the particles are too small. Onions are (or at some point
were) more likely to be eaten as a whole, thus they are countable, gar-
lic – less likely, hence uncountable. Prototypical fruits, such as apples,
are likely to be countable in most languages. She tries to explain the
differences between languages with references to eating habits: “people
(in Russia) normally eat cucumbers raw, holding them individually in
one hand, as one eats an apple or a banana, they don’t chop and cook
them…” (Wierzbicka, 1988, p. 504). It is true that different languages
reflect different ways of conceptualizing the world: “The different con-
ceptualizations have their own logic, which grammar reflects and illu-
minates” (Wierzbicka, 1988, p. 509). Indeed, a contrastive analysis of
the semantics of foodstuffs in different languages tells us that there are
a number of different possibilities for conceptualization.
The example Wierzbicka gives when discussing solids with a dou-
ble (both count and non-count) status, chocolate vs. soap, is actually
an excellent counter-argument to what she proposes. Soap cannot
be conceptualized in terms of individual objects. A cake of soap is
2.4. Countability 39

regarded as a piece of uncountable soap. Chocolates, on the contrary,

can be visualized as individual entities. This has to do, according
to Wierzbicka, with their fundamental qualities: a chocolate has an
attractive appearance as a whole, and dimensions suitable to be eaten
as one bite. Soap, on the other hand, can be viewed as “infinitely divis-
ible”: a piece of soap divided into parts will not lose its function as
soap (1988, p. 510). It seems that Wierzbicka’s argument has explana-
tory power in a diachronic sense: we can easily imagine that the first
bars of soap were considered in English to be just lumps or chunks
of some bigger mass, and this perception stuck and is reflected in
the language. Nevertheless, this does not change the fact that from
a contemporary perspective, there is hardly any difference between
a pretty chocolate bar and nicely packaged piece of soap. In fact, bars
of soap are similar in size and shape more than chocolates, so a soap
may produce a clearer mental image than a chocolate.
Her argument is that what seems arbitrary is actually not; it is true
that grammar does not reflect objective properties of entities in the
real world, because it reflects human conceptualization. That is a very
good point. But even when we accept, as Wierzbicka says, that “gram-
mar reflects the human perception of the world and human anthro-
pocentric interests” (1988, p. 520), there is still no predictive power to
such an explanation. Even if a learner of English takes a very care-
ful look at a bar of soap and a chocolate, there is no way to figure out
which is countable and which one is not. We can come up with an
explanation of why watermelon in uncountable (too big to be eaten as
a whole), but that does not mean that all things too big to be eaten
as a whole will be uncountable. Such post-hoc explanations do not
translate into any kind of predictive power which would be of use in
linguistic descriptions of how articles are used in a language.
Another example for the search for some kind of reality-based jus-
tification of the idiosyncratic countability behaviour of some nouns
can be found in the Cambridge Grammar, which gives the following
40 Chapter 2. Key concepts in the study of articles

explanation of why some nouns, such as dregs, grits, oats only have
the plural form. They denote substances consisting of particles, and
they have no singular form because the individual particles are of
no significance. They are different from non-count singulars (like
sand, gravel, rice, sugar) because “the particles tend to be relatively
larger in the case of the plural nouns (i.e. relative to the total amount
of substance under consideration)” (p. 342). This is somewhat far-
fetched – rice has bigger particles than dregs, and there are no reasons
to believe it is necessarily considered in larger quantities.
The Cambridge Grammar distinguishes between count and non-
count nouns, with the former denoting entities that can be counted
(as tested by occurrence with cardinal numbers). In the Cambridge
Grammar, chocolate as a count and non-count noun is treated as
a case of polysemy (chocolate has more than one sense). It is common
for nouns to have a count and non-count interpretation; the authors
of the Cambridge Grammar call it “a widespread phenomenon.” The
fact that some nouns (for example, piece) have no established non-
count uses, and some, like crockery, have no established count uses,
does not matter: “these are simply the limiting cases” (p. 335). Count
nouns denote what they call “atomic” entities, and non-count nouns
are either substances which are not inherently bounded (water), or
they are a heterogeneous aggregate of parts (crockery, bedlinen, furni-
ture, luggage). This approach does not seem to account for the exist-
ence of non-count nouns such as upholstery.

2.4.2. Implications for L2 acquisition

The differences in the conceptualization and lexicalization of counta-

bility across languages suggest considerable difficulties in second lan-
guage learning. The count/non-count distinction seems to be present
in most languages, and there are some common trends (for example,
2.4. Countability 41

a substance is likely to be non-count), but there are many differences.

In English, furniture is a non-count singular, contents is a non-count
plural; in French it is the other way round (meubles, contenu). In Eng-
lish, advice and information are conceptualized as substances, in Pol-
ish and French they are countable. It is not surprising that a learner
of English as a foreign language transfers the conceptualization of
the referent from one language to another, and accordingly says “an
information.”
A study by Yoon (1993) showed that native speakers of English,
when asked to classify nouns (which were presented out of context)
with respect to their countability, were not unanimous in their judge-
ments. This might not seem very surprising, as most nouns can func-
tion as either countable or not. In the same study, in a subsequent
task, the same native speakers were nevertheless entirely successful in
choosing the correct article to use with the nouns. Despite the meth-
odological limitations of this study, it points to a certain problem: the
level of metalinguistic knowledge, as well as the level of agreement,
related to article use among native users of the language may be lower
than for other aspects of the language system.

2.4.3. S ome uses of articles are very difficult

to explain

In many cases the use of the indefinite article seems to be in contra-

diction to the basic rules concerning countability. However, various
instances of such “violations” differ in the extent to which they form
certain patterns or regularities. For example, the use of primarily
non-count nouns in a count sense can sometimes be quite regular, as
in the case of drinks, which easily and regularly allow a count inter-
pretation as in a serving (a bottle, a cup) of the drink (I don’t like beer /
She offered me another beer).
42 Chapter 2. Key concepts in the study of articles

Similarly, nonce substance interpretations of primarily count

nouns are also quite understandable, as in this example:

The termite was living on a diet of book.

(Cambridge Grammar, p. 337)

However, some interesting unpredictable cases are discussed under

the heading of “abstract and event instantiations” in the Cambridge
Grammar (p. 337): while many abstract nouns which have a primary
non-count meaning can be used in secondary count senses denoting
an event which constitutes an instance of the abstract concept, some
do not, and this is not predictable, as shown by the following pairs of
sentences (examples from the Cambridge Grammar):

Considerable injustice was revealed during the enquiry.

*Two fundamental injustices were revealed during the enquiry.
Serious harm was done to the project’s prospects.
*Two serious harms were done to the project’s prospects.

The same goes for abstract nouns derived from verbs, as illustrated
by the following examples:

discussion – discussions
permission – *permissions
Full discussion of the land question is vital.
Two discussions of the land question took place.
Permission was required.
*Two separate permissions are required.

Finally, it should be mentioned that while countability is the main

criterion which enables the correct use of the indefinite article, there
are some cases when the indefinite article is used in structures with
countable plural nouns, as in:
2.4. Countability 43

A very pleasant three days was spent with Kim’s aunt in Brighton.
(Cambridge Grammar)

The explanation offered by the Cambridge Grammar is that the NP

containing the plural form is construed as singular, by virtue of
denoting a continuous stretch of time.
While with definite count singular NPs, identifiability is nor-
mally ensured by the fact that there is only one entity that satisfies
the description in the head of the phrase, as noted in the Cambridge
Grammar (p. 370), under certain circumstances, the definite article
can be appropriate with a count singular even when the context does
not strictly limit the number of entities satisfying the description
given in the nominal to just one:

Put down the cup on the arm of your chair.

He married the daughter of his bank manager.

This is very counter-intuitive, as the latter sentence does not mean

that the bank manager had only one daughter.
Another example of one more strange use of articles which lin-
guists try to explain is illustrated by the following examples:

He took the child by the hand.

He grabbed me by the arm.
She hit him in the face.

This usage is possible only with the referent being prototypically a body
part (cf. *(?) She grabbed me by the tie, Cambridge Grammar, p. 370).
Secondly, “the body-part NP and the NP with which it is associated are
respectively oblique and direct complements of a single verb,” which
explains why this kind of interpretation is not possible in the case of
I used the arm to help me get over the fence (Cambridge Grammar, p. 370).
44 Chapter 2. Key concepts in the study of articles

Let us consider this example from the Cambridge Grammar (p. 339):

Jill has a good knowledge of Greek.

This sentence is correct, even though knowledge is a clearly non-count

noun: it has no established plural and combines with determinatives
such as much, little, enough (as in they have little knowledge of the
matter). The explanation we find in the Cambridge Grammar is that
the function of the indefinite article is to “individuate a sub amount
of knowledge, her knowledge of Greek.” However, as the authors note,
“this individuation does not yield an entity conceptualized as belong-
ing to a class of entities of the same kind” (p. 339), so the following
sentences are not acceptable (p. 339):

*Jill has an excellent knowledge of Greek and Liz has another.

*They both have excellent knowledges of Greek.

From a language-learning point of view, the crucial aspect con-

cerning article use and countability is whether a certain observable
pattern is generalizable, that is, whether it constitutes a productive
“rule.” Certain patterns in the use of articles, even if very subtle, are
at least regular. Others seem to be completely arbitrary.

2.5. Definiteness

Definiteness is a concept that has been much discussed in semantics

and in the philosophy of language, which deals with the logic of the
notion of definiteness, and the truth values of definite descriptions. In
the early 20th century, Russell (1905) classified definite noun phrases
as “definite descriptions,” and this term continues to be used, even
2.5. Definiteness 45

without commitment to Russel’s views on definiteness (Neale, 1990).

Neale (1990) offers an in-depth analysis of Russell’s framework.
An early study on articles from a diachronic perspective was pub-
lished by Christophersen (1939), who introduced the notion of famil-
iarity in the study of definiteness: “the article the brings it about that
to the potential meaning (the idea) of the word is attached certain
association with previously acquired knowledge” (Christophersen,
1939, p. 72).
As Epstein (2002) notes, despite the divergences, most theories of
the definite article have a lot in common: they focus primarily on one
specific function of the definite article, namely the way it is used to
pick out a discourse referent. They try to answer the question of what
it is that allows one speaker to assume that a discourse referent will be
identifiable (uniquely) or familiar to another speaker. The sources of
definiteness that have been identified in the literature vary, and many
terms have been used by different authors to capture the notion of
definiteness: uniqueness, identifiability, unique identifiability, famil-
iarity, inclusiveness, determinedness, individualization, concretiza-
tion, actualization, specialization, particularization, salience (based
on Epstein, 2002; Głaz, 2012; von Heusinger, n.d.).
Articles are often studied as one of the many semantic and dis-
course-pragmatic devices used in English for structuring informa-
tion and maintaining topic continuity in discourse. The most widely
cited and most influential typology of the sources of definiteness is
to be found in the works of John Hawkins (1978, 1984, 1991). Firstly,
Hawkins defines the “previous discourse set” – the entities that the
interlocutors have talked about. For example, a

mention of a professor permits subsequent reference to the pro-

fessor. The speaker may have discussed numerous professors with
other interlocutors and have entered them into the previous dis-
course sets that he shares with these individuals, he may be able to
46 Chapter 2. Key concepts in the study of articles

access professors within the other kinds of pragmatic sets to be illus-

trated below, but as long as the previous discourse set shared with
the current interlocutor contains only one individual satisfying the
description professor, there will be no referential ambiguity through
non-uniqueness.
(J. A. Hawkins, 1991, p. 408)

Secondly, the

immediate situation of utterance in which the speaker and the hearer

find themselves can constitute a pragmatic set of entities for the pur-
pose of uniqueness: Pass me the bucket will be unambiguous for the
hearer if there is just one bucket in his field of vision, irrespective of
the existence of countless buckets elsewhere.
(J. A. Hawkins, 1991, p. 408)

The third possibility is provided by what Hawkins calls a “larger sit-

uation set,” that is, knowledge shared by people in the same physical
location, which explains why

inhabitants of the same town who have never met before can imme-
diately talk about the mayor, meaning the unique mayor of their
town, and inhabitants of the same country can do likewise with the
president or the queen, despite the existence of numerous mayors,
presidents and queens elsewhere.
(J. A. Hawkins, 1991, p. 408)

Finally, there comes the most interesting source of definiteness in dis-

course, namely the general knowledge shared by speakers concern-
ing “predictable co-occurrences of entities: (…) mention of a wed-
ding permits immediate reference to the bride; a house sanctions the
roof; a car, the steering wheel; and so on” (J. A. Hawkins, 1991, p. 409).
2.6. Conclusions 47

Most semantic and pragmatic descriptions now recognize the above

uses of the definite article. Labels often used for the above four uses
are anaphoric, deictic, unique, and bridging (i.e., indirect) uses (von
Heusinger, n.d.).
Definite and indefinite reference have long been a subject of study
in the philosophy of language and in linguistic semantics (for an
overview, see J. Lyons, 1977, pp. 179–187, or 1995, pp. 294–298). John
Lyons points out the fact that the meaning of all referring expressions
is context-dependent. A common noun preceded by a definite arti-
cle may suffice without further description, even though the referent
has not been previously mentioned, because the speaker can fairly
assume in the given situation, that the hearer will know which of the
potential referents satisfies the description he is referring to (J. Lyons,
1995, p. 181). Performing a definite description is an act of reference,
in which the speakers “tacitly assure the addressee that the descrip-
tive part of the expression will contain all the information that is
required, in context, to identify the referent” (J. Lyons, 1995, p. 300).

2.6. Conclusions

At first sight, the place articles firmly hold in grammatical accounts

of language use appears to be very much justified: they seem to form
a system of oppositions and lend themselves to systemic descrip-
tions, which aid learners of English to a certain extent. However,
those descriptive accounts fail to reflect the full complexity of the
English article system and to cover the entirety of existing and poten-
tial uses of the articles. Upon closer inspection, rules governing the
use of articles turn out to be so complex and elusive that it remains
something of a miracle that (most) language users end up using arti-
cles in a consistent fashion.
48 Chapter 2. Key concepts in the study of articles

We have seen that articles are very common in English, that their
distribution is connected not only with their syntactic role in a sen-
tence, but that it also follows the information flow principle, which
means that there is a strong pattern of the new information being
indefinite and the given information – definite. Also, it was noted
that some uses of articles, e.g. generic ones, are in fact less common
in language than it may seem from descriptive approaches and ped-
agogical materials. (Pedagogical materials traditionally include an
extensive discussion of generic uses.)
On the topic of definiteness, the following comment by Kambarov,
even though it refers to natural language processing, seems to hold
true for definiteness in language in general:

a typical grammatical reference work addressing the concept of defi-

niteness usually contains a large number of rules with numerous
exceptions and special cases, which hinders effective use of infor-
mation provided by determiners in nominal reference resolution.
(2008, p. 1)

Definiteness is a much-studied and intriguing topic, but none of the

theoretical treatments it has received so far have translated into clear
rules which could be of pedagogical use.
Yet another source of considerable difficulty for learners of Eng-
lish is the notion of countability, due to the conceptual differences
across languages as far as the count/non-count distinction is con-
cerned. Finally, there are numerous instances where the use of the
definite, indefinite or zero article is in contradiction to other rules
which seemingly should apply, as, for example, in many expressions
where the zero article is used with singular countable nouns.
3.1. Introduction 49

Chapter 3

Other approaches to articles

3.1. Introduction

The first chapter of this book presented descriptive approaches to

articles, while the second one looked at the key concepts behind the
use of articles, most notably reference, definiteness and countability.
It was shown that the philosophy of language and semantics have
contributed to the understanding of these notions. This chapter pro-
vides a brief overview of some other important linguistic approaches
to articles.

3.2. Socio-pragmatic approaches

One important development in the philosophy of language which

has direct relevance to the study of articles is the work of Paul Grice.
A closer look at instances involving the use of the definite article
makes it clear that the instances of its use are not only dependent
50 Chapter 3. Other approaches to articles

on the preceding text as such, but on general pragmatic knowledge,

and on the situational context. Therefore, the motivation for the use
of the definite article lies outside language as such. As is acknowl-
edged in the Longman Grammar, “the interpretation of definite noun
phrases often requires extensive pragmatic inferencing on the part
of the addressee” (p. 264). In fact, making inferences is what the lis-
teners do in order to preserve a sense of coherence in what they are
told. This is the reason why traditional sentence-based semantics can-
not fully account for many linguistic phenomena, including the use
of articles. It is clear that an analysis of the use of articles cannot be
complete without recourse to pragmatics.
The first theory that managed to account for inferencing was for-
mulated by Grice as the principle of conversational implicature (Grice,
1975), which was later extended and developed into relevance theory
by Sperber and Wilson (Sperber & Wilson, 1995). The Gricean max-
ims (Grice, 1975) of conversational cooperation (of quality, quantity,
relevance, manner) explain why, in the following exchange:

A: Did you give Mary the money?

B: I’m waiting for her right now.

The inferred answer to A’s question is negative. In the theory of rele-

vance, the hearer’s inferential strategy is motivated by a single prin-
ciple of relevance, which holds that “Every act of ostensive communi-
cation communicates the presumption of its own optimal relevance”
(Sperber & Wilson, 1995, p. 158).
A recent very interesting line of research on the English definite
article is of socio-pragmatic character. Acton (2014), in an analysis
rooted in Gricean pragmatics, looks at the English definite article and
analyses how the definite article contributes to the social meaning of
an utterance. For example, his corpus study shows that when referring
to a group of people, using the definite article (e.g. “the Americans”)
3.3. Cognitive accounts of definite article use 51

instead of a bare plural (“Americans”) presents the group as a homo-

geneous entity of which the speaker is not a part (Acton, 2014).

3.3. Cognitive accounts of definite article use

As can be seen from the above overview, reference is a concept which

is crucial to the understanding of how definite and indefinite arti-
cles function in English. Cognitive linguistics has contributed sig-
nificantly to providing a theory which elucidates many aspects of
reference. It is now common (see Kambarov, 2008) to see referents of
nouns in a text as elements of one of several alternating mental spaces,
which provide “a level of conceptual organization between the text
and reality” (Kambarov, 2008, p. 4). The specific theory within cog-
nitive linguistics from which this concept arises is the Mental Spaces
Theory (Fauconnier, 1994). It is a theory of dynamic meaning con-
struction, the cornerstone of which is the proposition that discourse is
divided into distinct mental spaces, which feature counterparts linked
together by connectors. Mental spaces are set up to recruit temporary
structure from the local discourse context and from other sources,
including long-term memory. This process is known as “schema
induction.” As language users make sense of a text, they assign truth
values to entities and relations between them in a given mental space,
updating them on the basis of whatever information becomes availa-
ble in the discourse, as well as on the basis of the existing knowledge
they have, which is organized in hierarchical structures known as
“frames.” Existing mental spaces are altered to form new ones, which
results in the creation of a mental spaces lattice.
As cognitive approaches to language typically seek to ground
language description in general cognitive mechanisms, also in this
case, the process of forming mental spaces is seen as governed by
52 Chapter 3. Other approaches to articles

non-linguistic cognitive principles (Evans, 2007, pp. 135–136), and lan-

guage is only a source of prompts which initiate this process. Since, as
Fauconnier observed, language does not carry meaning, but guides it
(1994, p. xxii), articles can be seen as signals which help the receiver
to manipulate mental constructs containing referents of nouns.
A detailed cognitive account of the working of the English article
system is provided by Epstein (e.g. Epstein, 2002), who drew on the
work of Fauconnier (1994) and Ariel (1990). Epstein invokes a state-
ment by Birner and Ward that “none of the previous analyses can
account for all uses of the definite article in English” (Birner & Ward,
1994, p. 101) and argues that none of the approaches to defining defi-
niteness are fully satisfactory. He points out that, because of its “enor-
mous complexity” (Epstein, 2002, p. 338), indirect anaphora is very
difficult to define in precise formal terms. Like Ariel and Fauconnier,
Epstein considers grammatical elements in general to be discourse
processing instructions. In this view, the basic meaning of the is to
signal to the addressee the availability of an “access path.” In other
words, the definite article indicates that the knowledge required for
interpreting a noun phrase is accessible (either already active or able
to be activated) somewhere in the dynamic configuration of spaces.
The use of the definite article signals to the recipient that there is
a connection between the discourse entity and some other entity. The
exact connection has to be worked out by the recipient, which explains
why some utterances may be ambiguous (Epstein, 2002, pp. 334–346).
Epstein’s account of article use is particularly convincing with
respect to certain “strange” uses of the definite article which can be
found especially in literary fiction and journalism, such as, for exam-
ple, the use of the with first-mention nouns which are neither accom-
panied by any modification, nor can be assumed to be identifiable for
the hearer. Epstein explains how the definite article in such contexts
derives from prominence in subsequent discourse. For instance, it
may signal the point of view of the narrator, or the protagonist. This
3.3. Cognitive accounts of definite article use 53

can be illustrated with the following example from Epstein’s paper,

which is the opening sentence of Hemingway’s A Farewell to Arms:

In the late summer of that year we lived in a house in a village that

looked across the river and the plain to the mountains.

According to Epstein, the passage encourages the reader to adopt

the point of view of the narrator, by using the definite article with
nouns that have not been mentioned before, and which cannot be
assumed to be familiar or identifiable to the reader (the river, the
plains, the mountains).
Another similar example comes from a newspaper:

Sierra Madre resident Andy Dotson might not have needed to breach
security barricades to return to his threatened home. He had forgotten
his tattered, 19-year-old blanket with the distinctive penguin design. “The
kids and the animals are my security blanket, they come first,” he said.
“But my family didn’t get [the blanket], so I went back there. It means
something to me. I was gonna bust through the barricades if I had to.”
(Los Angeles Times, 30 October 1993, p. A10, cited by Epstein, 2002,
p. 365)

The definite article in the distinctive penguin design signals that the
blanket is being described from the point of view of the man who
went back into a building on fire in order to retrieve it.
Epstein also makes the very interesting point that the definite arti-
cle signals a low level of accessibility to the referent. One of Epstein’s
examples is a comparison of the following two utterances (Epstein,
2002, p. 340):

a. There’s a cat in the yard. It’s eating a mouse.

b. There’s a cat in the yard. The cat is eating a mouse.
54 Chapter 3. Other approaches to articles

There is no doubt that (a) is the more usual, natural of the two. The
referents of both it and the cat are uniquely identifiable, as they have
just been introduced. However, they remain highly accessible, because
a cat is introduced in the first sentence, and there is no intervening
discourse between the first and the second sentence. The is reserved
for contexts with a lower degree of accessibility, this is why it seems
unnatural in (b), where accessibility is high.
Another example from Epstein illustrates how speakers use the
definite article with a referent that is new, so it is not yet identifiable
to the interlocutor, but the speaker signals that he is intending for
this to become the new topic of conversation, i.e., he assumes that
the introduction will need elaboration on, and that the interlocutor
knows it too (Epstein, 2002, p. 335):

M: Did you hear about the fight?

A: What fight?
M: Between Bob and Grandpa.

In this account, the definite article emerges as essentially a discourse

processing instruction,

… signalling that the means for interpreting the NP in which it occurs is

available somewhere in the configuration of mental spaces, as long as the
appropriate spaces, elements and connections – i.e., access path – can
be constructed by the addressee. Some of the functions fulfilled by the
are: unique identifiability, prominence, role/value status, point-of-view
shifts. Each function represents a conventional interpretation poten-
tially associated with a definite description. However, none of these
functions is specifically conveyed by the article itself. Instead, the inter-
pretation of a given definite description arises in a particular context
through a combination of lexical, grammatical, and pragmatic factors.
(Epstein, 2002, p. 371)
3.3. Cognitive accounts of definite article use 55

Epstein’s approach is useful in that it demonstrates that neither

familiarity nor identifiability are sufficient conditions for determining
definite article use. The mental spaces theory is very good at explain-
ing some uses of the definite article which the previous theories could
not account for, such as indicating the discourse prominence of an
entity, the entity’s status as a role function, or the fact that an entity
is presented from a noncanonical point of view.
Another cognitive account of English articles has been put for-
ward by Głaz (2012), whose work was inspired by Vantage Theory,
a model of categorization developed by linguist and anthropologist
Robert E. MacLaury, whose work concerned mostly colour catego-
rization and the way it reflects the way humans orient themselves in
space and time (MacLaury, 1995). Like Epstein, Głaz arrives at a plau-
sible explanation of a number of uses of English articles which could
not be accounted for in terms of reference, specificity or uniqueness.
An introductory illustration of his approach is provided in the fol-
lowing excerpt, analysed in Głaz (2012), of the beginning of Doris
Lessing’s novel The Good Terrorist:

The house was set back from the noisy main road in what seemed
to be a rubbish tip. A large house. Solid. Black tiles stood at angles
along the gutter…

In this excerpt, the house is first introduced with a definite article,

then referred to again, this time with an indefinite article, which is the
opposite of should be happening according to the “rule” of prior men-
tion. The first sentence assumes familiarity with the house, while the
second treats the house as merely a member of the category of houses.
Drawing on vantage theory, Głaz models the conceptualization of the
house as the succession of three vantages: recessive, dominant, and
recessive again. In Głaz’s opinion, applying Vantage Theory to the use
of articles makes it possible to explain some seemingly arbitrary uses,
56 Chapter 3. Other approaches to articles

usually considered to be motivated by convention only, by revealing

the “cognitive motivation” behind them (Głaz, 2012, p. 18).
Recently, reference and definiteness have been studied in the
context of natural language processing. In this field, reference is an
important issue, as one of the main challenges in computerised pro-
cessing of human language is co-reference resolution, that is, deter-
mining which words refer to the same entities (for example, anaphor-
ically). The challenges in nominal reference resolution centre around
identifying cases of co-reference, and instances of new entities being
introduced into discourse. Kambarov (2008) has established a set of
rules that “capture the factors essential to the concept of definiteness”
in the form of an algorithm, the purpose of which is to aid the process
of resolving reference in natural language processing. He proposes an
algorithm for nominal reference resolution that uses the definiteness
value of each determiner phrase to guide the search for its referent.
This work is inspired by cognitive linguistics, as the author draws on
the theory of mental spaces. The algorithm organizes referents of all
encountered nouns in structures that correspond to mental spaces
that are being described.

3.4. The generative tradition

What generative approaches have in common is that they do not

treat articles as a word category in itself. Rather, like most descrip-
tive grammars, they include them among the four categories of
determiners, along with demonstratives (this), quantifiers (many,
three), and possessives (my). A second assumption that most gen-
erative approaches share is that L2 acquisition does not start from
scratch like in L1 development, but relies on the entire set of para-
metric values fixed in the L1; this assumption is known as the “full
3.4. The generative tradition 57

transfer hypothesis” (Schwartz & Sprouse, 1996). Consequently, the

learning and processing of L2 articles would begin with the para-
metric values that the learners established for articles in their L1.
In this view, L2 acquisition is a question of parameter resetting (see
Kaltenbacher, 2001).
Beyond those two assumptions, the multiplicity of approaches
in generative grammar and the various roles they assign to articles
make it difficult to identify commonalities among them regarding the
acquisition of L2 articles. Represented for a long time as mere specifi-
ers of nouns inside Noun Phrases (NPs), determiners have been con-
sidered to play a more important role in some recent syntactical mod-
els, where they are perceived as heads of the phrases in which they
appear. Those phrases are consequently labelled Determiner Phrases
(DPs) rather than Noun Phrases (NPs) (see Hudson, 2004; Van Lan-
gendonck, 1994). Within the Determiner Phrase (DP), the determiner
is one of a series of functional categories that head a maximal projec-
tion with abstract features such as [±definite], [±specific], [±count].
However, disagreement persists among theoreticians, and Granfeldt’s
statement that “there seems to be no consensus on the internal struc-
ture of the DP” (2000, p. 264) apparently still holds true.
Moreover, the fact that articles are morphemes raises further
divergences. Some researchers agree with White’s claim that “UG
does not have anything to say about morpheme acquisition as such:
morphemes are lexical items; they are language specific and have to
be learned” (L. White, 1989, pp. 30–31). Other researchers in Gener-
ative Grammar (e.g. Vainikka & Young-Scholten, 1996) suggest that
the L2 initially develops exclusively from the parametric values and
syntactic projections of lexical items project, and that functional items
such as articles only come into play at a later stage in L2 development.
In some languages, articles carry a limited information load,
such as English, where they only indicate definiteness. In other lan-
guages, such as French or Spanish, they convey information about
58 Chapter 3. Other approaches to articles

definiteness, number and gender.6 Consequently, mastering an arti-

cle system in a second language is challenging not only for speakers
of article-less languages;7 they also pose greater difficulty for learners
whose L1 contains articles, in cases where those articles function dif-
ferently from the L2, or – in generative terms – when the “parameter
settings” differ across languages. Researchers have formulated var-
ious parameter-based hypotheses to explain the emergence or mas-
tery of articles. A combination of two parameters is found in L2 stud-
ies on the acquisition of English articles: the [±specific referent] and
the [±hearer knowledge] parameters. The former concerns whether
the article and noun refer to a specific entity, and the latter concerns
whether their referent is known to the hearer. This issue will be dis-
cussed in more detail in section 5.3.

3.5. Corpus-based perspectives

Corpora are covered here in this last section because, thanks to

advances in technology, only recently has corpus-based research gained
considerable ground and made notable advances. Apart from inform-
ing descriptive accounts of the article system (most notably the Long-
man Grammar), corpora also provide information on the distribution
of articles across registers and on the frequency of their occurrence.
According to the Longman Grammar, the indefinite article
appears with a frequency of approximately 20,000 per million words

6 The picture is further complicated by the fact that in some languages, such as classi-
cal Arabic, definiteness is conveyed by the article, while indefiniteness is conveyed by
the noun (as is the case of English but only in the plural). In Swedish, it is the oppo-
site situation: definiteness is conveyed by the noun, indefiniteness by the article.
7 One of the rare exceptions to that finding is Deprez, Sleeman and Guella (2011).
3.5. Corpus-based perspectives 59

in the written registers (which amounts to 2% of the running text)

and about 13,000 words per million in conversation (i.e., 1.3%). While
the indefinite article is distributed in a more or less similar manner
across registers, there is much greater variance in the distribution of
the definite article. The definite article is very common in academic
prose (with a frequency of ca. 55,000 per million words, i.e., 5.5%),
fiction (ca. 42,000, which means 4.2% of all the words are definite
articles), and is the least frequent in conversation, with only around
20,000 instances of use per million words (2%).
The lower frequencies of articles in conversation than in texts
are caused, among other things, by the lower frequencies of nouns,
and higher frequencies of proper names and personal pronouns. In
conversation, the indefinite article with its function of introducing
new entities is sometimes replaced by the demonstrative determin-
ers this/these, but corpus analyses show that this use is predomi-
nantly restricted to informal spoken language.
According to the Longman Grammar, the definite article is the
most frequent determiner of all. It occurs more often than its most
popular “replacements,” that is, other definite determiners, posses-
sive and demonstrative ones. The proportion of the occurrences of
the definite article with respect to other words that can appear in
the same position varies across registers; the relative frequency of
the definite article is the highest in academic prose.
The authors of the Longman Grammar observe that anaphoric
reference is intuitively perceived as the most common use of the
definite article. However, they note that when the definite article
use is investigated in a corpus, other uses emerge as equally or even
more common. In fact, anaphoric reference accounts for less than
one third of the uses of definite noun phrases, while situational
reference is surprisingly common. This observation resulted most
likely from the inclusion of spoken language in the corpus under
investigation.
60 Chapter 3. Other approaches to articles

There are also quite notable differences between registers. For

example, situational uses account for 55% of cases in conversation,
while they make up only 5% in the news or in academic sub-corpora.
This is fully justified, because, as the authors say,

conversation is embedded in a situation which is shared by the

speaker and hearer. Moreover, conversational partners are usually
closely related as family members or friends, and can thus rely on
a great deal of shared knowledge.
(Longman Grammar, p. 267)

Thinking of future research in this area, it is worth noting that, in

their attempt to classify the uses of the definite article across a variety of
registers, Biber and his colleagues (1999) were unable to classify (label:
“uncertain”) between 5 and 25% of the uses, depending on the register.

3.6. Conclusions

This chapter has looked at some of the theoretical approaches to arti-

cles and how their proponents attempt to make sense of the seemingly
confusing character of the English article system. The above review
provides only a very brief overview of the most important approaches,
but it is enough to show that articles have received a considerable
amount of attention from scholars representing various schools of
linguistic thought. The fact that there exist such extensive treatments
of various concepts relevant to the use of articles (for example, defi-
niteness or specificity), brings home the fact that the issues involved
are indeed very complex.
One approach which offers a very inspiring look at the English arti-
cles is cognitive linguistics. However, even though cognitive accounts
3.6. Conclusions 61

of article use are fascinating, and they apparently cover a much wider
range of article uses than competing accounts, providing insights into
the conceptualisations behind article use, the explanatory power of
this approach does not easily translate into practical applications.
The explanations offered by cognitive linguists have an essentially
post-hoc character. Also, even if some of its aspects could be used
for teaching purposes, the intricacies of such concepts might be too
complex to grasp for many practitioners.
4.1. Introduction 63

Chapter 4

Articles as a source of difficulty in SLA

4.1. Introduction

When English is taught as a second language in an English-speaking

country, that means, mostly to learners who are immigrants, teach-
ers tend to be less aware of specific L1-induced areas of difficulty for
their learners, due to the heterogeneity of the learners’ L1 background.
However, English is also taught in what is sometimes referred to as
a “foreign language” context, in non-English speaking countries,
which are sometimes very linguistically homogeneous, as is the case
with Poland. In such contexts, teachers tend to be well aware of the
typical problem areas to expect from their students, and it is com-
mon knowledge which target language features tend to cause diffi-
culty. The English articles are a classic example of a well-known prob-
lem area, and it is also possible that some of the teachers themselves,
being speakers of L1 Polish, might be less confident than usual when
it comes to correcting their students’ article errors.
Since the statement that “articles are difficult” is such a com-
mon remark in the English teaching context in Poland, this chapter
64 Chapter 4. Articles as a source of difficulty in SLA

provides an overview of theoretical perspectives on the concept of

difficulty, to see if those theoretical perspectives allow for a better elu-
cidation of the reasons which underlie the challenges posed by this
specific feature of the English language.

4.2. Theoretical approaches to difficulty

A number of criteria have been put forward in the literature, in

a rather chaotic and scattered matter, to account for what it means
for a structure to be difficult, including formal or structural complex-
ity, form-meaning relationships, developmental stages, learnability,
typological markedness, and teachers’ perceptions of learner diffi-
culty (see Ekiert & Han, 2017). Difficulty has been described as a very
challenging concept to define (see DeKeyser, 2005 for an in-depth dis-
cussion). According to Housen and Simoens, the existing literature
on the topic is scarce,8 difficulty is usually investigated in connec-
tion with a different concept which is the main focus of research, and

the study of difficulty in SLA has been plagued with terminologi-

cal difficulties, conceptual confusion, and misunderstandings of its
relationship with other related yet conceptually distinct constructs
such as linguistic complexity, learnability, and developmental stages
and orders of acquisition.
(2016, p. 164)9

8 However, one could point out that with respect to specific aspects of SLA, for
example vocabulary, the concept of difficulty has been defined (see e.g. Nation’s
concept of learning burden (2001)) and investigations have been conducted to
determine what constitutes difficulty (e.g. Laufer, 1997).
9 For a more comprehensive review of the situation, see Housen and Simoens (2016),
also Ekiert and Han (2017).
4.2. Theoretical approaches to difficulty 65

Importantly for this discussion, many accounts of difficulty are entirely

L2-centric, which means that they do not mention L1/L2 differences as
a potential source of difficulty, which renders them inadequate for ana-
lysing the case of articles, since, as has already been mentioned and as
will be made clear by the review of research in the next chapter, cross-
linguistic differences play an important role in determining the level of
difficulty for learners. In the days of contrastive analysis, articles were
an ideal argument supporting the hypothesis that differences between
the L1 and the L2 lead to difficulty with specific aspects of the L2. How-
ever, despite the predictive power of contrastive analysis when it comes to
articles, the approach generally lost its popularity in the 1960s and 1970s,
when generative grammar and cognitivism replaced behaviourism as
theories informing approaches to the study of L2 acquisition (see Pichette
& Leśniewska, 2018). Actually, Housen and Simoens (2016) suggest that
difficulty was for a long time an unpopular topic of inquiry precisely
because it was associated with contrastive analysis, which was considered
outdated and insufficient (see Lightbown & Spada, 2006).
The concept of difficulty re-emerged in language studies near the turn
of the century, partly due to the advent of available large language cor-
pora, and in connection with studies of structural and formal complexity,
which tend to be L2-centric. Formal complexity may be defined in many
different ways, depending on which linguistic theory is applied, for exam-
ple, it may be operationalized as the number of transformations needed
to arrive at a certain construction. However, regardless of which linguis-
tic theory is adopted by the researcher, the general rule is that the more
“complicated” a construction is, the more difficult it is to learn and use.
In this view, articles are not difficult at all, since they are formally very
simple. Interestingly, Spada and Tomita (2010) use the example of arti-
cles to argue that structural complexity does not function as a predictor
of difficulty, since structurally they are an example of the simplest struc-
tures possible, derived by one transformational rule, in contrast to more
complex structures, which require a greater number of transformations.
66 Chapter 4. Articles as a source of difficulty in SLA

4.2.1. Difficulty as cognitive complexity

Fortunately, a very recent publication by Housen and Simoens (2016)

offers a useful taxonomy in which there is a place for L1-L2 differences,
and which combines all the different approaches to difficulty into one
framework. I will attempt to apply this framework for analysing the
specific difficulty caused by articles. To avoid terminological confu-
sion, Housen and Simoens propose to distinguish between structural
complexity (or linguistic/absolute complexity), which means

the inherent linguistic properties of a language feature or (sub)sys-

tem (…) typically operationalized in terms of the number and vari-
ety of the discrete components of which a language feature consists,
(2016, p. 166),

and cognitive complexity, which

has to do with how costly, demanding, or difficult a given language

feature is for a given language learner in a given learning context,
particularly in terms of the mental resources allocated and cognitive
mechanisms deployed in processing and internalizing the feature.
(2016, p. 166)

It thus follows that structural complexity may be a contributing fac-

tor to cognitive complexity, but the two do not coincide. For the rest of
this discussion, I shall adopt the terminology proposed by Housen and
Simoens, with difficulty used to refer to cognitive complexity. Thus, “a
language feature is more difficult than another if its processing and
learning requires more time and/or more mental activity from a par-
ticular learner in a particular learning context” (2016, p. 166).
Within this framework, three main types of difficulty are recog-
nized: feature-related difficulty (which is described as “objective”),
 L2 Difficulty
(cognitive complexity)

Feature-related Context-related Learner-related

difficulty difficulty difficulty

Intrisic feature Input properties of form- Learning Individual learner

properites meaning connections conditions characteristics

Function Form Naturalistic Instructed Socio-

Regularity Redundancy Frequency ... Cognitive ...
complexity complexity (implicit) (explicit) affective

Transparency Saliency

Figure 1. A taxonomy of learning difficulty offered by Housen and Simoens (2016).

68 Chapter 4. Articles as a source of difficulty in SLA

context-related difficulty, and learner-related difficulty (described as

“subjective”). An important characteristic of this framework is the
recognition that all the main types and sub-types interact with one
another (see Figure 1).

4.2.2. Feature-related difficulty

A useful distinction made by the creators of this framework is that

feature-related difficulty can be seen as resulting from two types of
factors: firstly, feature-intrinsic properties, and secondly, the way in
which this particular feature appears in the input. Feature-related
complexity can be either formal or functional.
The intrinsic formal (structural) complexity can be assessed
according to a variety of different linguistic theories, and the assess-
ment of the degree of such complexity need not be exactly the same
according to, for example, a generative, typological or a cogni-
tive account of language. However, regardless of which theory is
applied, the level of complexity is a feature of the specific struc-
ture in question, regardless of its distribution in the input, which is
a separate criterion. Functional (also called semantic, or conceptual)
complexity refers to the number and character of the meanings and
functions expressed, and the mapping of forms and functions onto
each other. This aspect is reminiscent of the “one to one principle”
(Andersen, 1984), which postulates that the acquisition of a new form
is greatly facilitated when there is a clear and unique correspond-
ence between form and meaning. Input-related properties, in turn,
such as regularity, redundancy, and frequency, all contribute to how
transparent and how salient the feature will be to the learners. The
more salient and transparent a feature appears to the learner, the eas-
ier it is to learn.
4.2. Theoretical approaches to difficulty 69

4.2.3. Context-related difficulty

The second type of difficulty is determined by the context, that is, by

the learning conditions. It is important to distinguish between nat-
uralistic (or unguided) learning contexts, and instructed (or guided)
ones. The former are said to encourage implicit processing, while
the latter involve explicit attempts at guiding the learner’s cognitive
resources. This may involve “directing learners’ cognitive resources
to the targeted form-meaning mappings by enhancing their trans-
parency, modifying their input frequency, or otherwise increasing
their salience so that, ultimately, their learning difficulty is miti-
gated” (Housen & Simoens, 2016, p. 169). One more crucial role of
instruction (apart from modifying or enhancing the presentation of
form-meaning mappings in the structured input which is offered to
the learners) is to provide learners with metalinguistic propositions,
commonly known as grammar rules as instantiated in pedagogical
books. Both the implicit-explicit distinction and the issue of meta-
linguistic propositions have important implications for the learning
and use of articles, and they will be discussed in more depth in the
sections which follow.

4.2.4. Learner-related difficulty

The third type of difficulty is related to the learner, and, in this cat-
egory, we finally find a place (though, admittedly, not a prominent
one) for crosslinguistic aspects. The most important learner factors
are “individual differences in cognitive abilities, particularly language
aptitude (…), working memory, and implicit and procedural learn-
ing ability” (Housen & Simoens, 2016, p. 167), as well as socio-affec-
tive and personality factors, but there is also a place for “the learn-
er’s previous knowledge” which encompasses the first language or
70 Chapter 4. Articles as a source of difficulty in SLA

any other language already known by the learner. The learner-related

factors also include overall L2 proficiency and stage of L2 develop-
ment. Taken together, these learner differences are considered to be
“the core ingredients of L2 learning difficulty” (Housen & Simoens,
2016, p. 167).

4.2.5. Some final words on objectivity

Perhaps the most problematic element of this framework is the treat-

ment of functional (semantic, conceptual) complexity as objective.
All-feature related aspects of difficulty are believed to be “objective,”
because “some language features are more cognitively demanding for
all language learners, irrespective of their individual learner charac-
teristics” (Housen & Simoens, 2016, p. 167). Commonsensically, this is
hard to accept, since even the most complicated form-meaning map-
ping in L2 will be relatively easy for certain learners if it happens to
be very similar to the one in their L1.
To give a very simple example, beginners’ courses in German
for speakers of English usually devote a note or a section in one of
the very first lessons on the German word bitte, because in Ger-
man it is used with a number of meanings (situations) in which dif-
ferent English expressions would be used (though different into-
nations may accompany the different meanings of bitte):

Please (when asking for something, or asking someone to do some-

thing)
You’re welcome (when responding to thanks)
Here you go (when handing something over)
May I help you? / What would you like? (from a person who wants
to take your order)
Pardon? (when asking for a repetition)
4.2. Theoretical approaches to difficulty 71

Even though the difficulty here is minor, because the target form is
a single, simple word, there is some remapping of form to meaning
required; whereas for a Polish learner of German, this particular
issue does not arise, as the Polish word “proszę” mostly overlaps with
“bitte” in the range of situations in which it can be used.
The challenge one faces with trying to think of meaning-related
difficulty as separate from crosslinguistic aspects can be illustrated
with this passage from an article by DeKeyser:

Linguistic difficulty is due to the characteristics of the L2 target:

the meanings to be expressed, the forms used for expressing them,
and most crucially, the form-meaning mappings. (…) Meanings can
be difficult because they are both novel (not represented in L1) and
abstract, and hence hard to infer from the input. The semantics of
articles and of grammatical aspect marking on the verb are notorious
examples of this. Classifiers in languages like Chinese, Japanese, or
Korean are somewhat less abstract but are still novel for many learn-
ers; conversely, person, number, and tense marking on the verb are
far from obvious for speakers of these languages.
(DeKeyser, 2016, p. 354)

This passage strongly suggests that it is impossible to consider the

difficulty of the “L2 target” without considering it through the lens
of the L1 (or other languages the learner knows).

4.2.6. Articles and feature-related difficulty

However, assuming this problematic aspect can be overcome, the

above framework can be used to assess the difficulty level of the
English articles. First of all, articles score very low on form com-
plexity, so in this sense they are easy. Their functional complexity,
72 Chapter 4. Articles as a source of difficulty in SLA

however, is high. In a taxonomy for assessing learning difficulty,

Roehr-Brackin and her colleagues (Roehr & Gánem-Gutiérrez,
2009; Silva & Roehr-Brackin, 2016) distinguish between opacity
of form-meaning mapping (one form, X meanings) and opacity of
meaning-form mapping (one meaning, X forms). In the case of arti-
cles, it is the first of the two which is the source of difficulty, as each
of the L2 forms (the and a/an) maps onto multiple meanings or func-
tions (see also Shintani & Ellis, 2013). Additionally, one of the cate-
gories involved, the distinction between countable and uncounta-
ble nouns, is difficult in itself, because (like grammatical gender) it
is not predictable (see 2.4).
As far as the input properties are concerned, articles are very
frequent, which should contribute to noticing, and thus lower the
level of difficulty. It is interesting to note here that this characteris-
tic (high frequency) is a facilitating feature only when input prop-
erties are concerned. When it comes to language production, the
situation looks radically different. As Master notes (2002: 332), in
the case of articles, their high frequency adds a significant bur-
den to the learner’s language production, as it makes it necessary
to continuously apply a set of complex rules. However, this view of
frequency as a source of difficulty does not fit within the present
framework, which is reception-oriented, and does not account for
language production.
As far as articles are concerned, there is a characteristic which is
missing from the framework under discussion and from most other
sources which list difficulty-inducing factors. Among the input-re-
lated properties, an important difficulty-inducing factor which should
be included is something that DeKeyser (2005) refers to as “optional-
ity,” and I will adopt this term. In his words:

Optionality of certain elements, such as null subjects in Spanish or

Italian (…) or case marking in Korean, only makes matters worse.
4.2. Theoretical approaches to difficulty 73

Not only does the optional character of the case marking or the overt
subject pronoun suggest it is redundant, but its alternating presence
or absence in the presence of the same meaning, except for subtle
aspects of pragmatics, makes the form-meaning link even harder
to establish.
(DeKeyser, 2005, p. 8)

Optionality is an important article-related characteristic. Exam-

ples abound, but let me illustrate this phenomenon with an example
from the pilot phase of the study which is presented in Chapter 8 of
this book, in which native speakers of English were asked to supply
missing articles in the sentence:

Airbags are supposed to inflate automatically in case of collision.

The following corrections were made by the respondents, in

roughly equal proportions:

Airbags are supposed to inflate automatically in the case of collision.

Airbags are supposed to inflate automatically in case of a collision.
Airbags are supposed to inflate automatically in the case of a collision.

This lack of consistency in article use, along with the complex

form-meaning mappings, results in low transparency of the input,
which is a factor predicted to raise the difficulty level. Another dif-
ficulty-inducing factor is communicative redundancy, which in the
case of articles is high, as in many cases utterances deprived of arti-
cles would be understandable.
74 Chapter 4. Articles as a source of difficulty in SLA

4.2.7. Salience

An element which appears in the framework proposed by Housen and

Simoens (as a consequence of various input properties), and which is
often evoked in comments about the learning of the English articles,
is salience. Generally speaking, salience refers to the ability of a stim-
ulus to stand out from the rest (N. C. Ellis, 2016). Items which are sali-
ent are more likely to be noticed, cognitively processed, and learned,
whereas linguistic forms of low salience are more difficult both to
perceive and to learn (N. C. Ellis, 2006). Importantly for the case of
articles, low salience can override high frequency (N. C. Ellis, 2016).
It is very commonplace to read that articles in English are not very
salient, or do not appear salient to the learners. However, any discussion
of the concept of salience10 in the literature has to be treated with cau-
tion, as the term seems to be used with whatever meaning that comes
to the writer’s mind, and the available publications, instead of working
towards establishing a shared definition and a usable taxonomy of dif-
ferent types of salience, appear to take delight in the fact that the term
cannot be defined with any clarity; as illustrated, for example, by the
introduction to the very recent book Salience in Second Language Acqui-
sition (Gass, Spinner, & Behney, 2018). In fact, somewhat amusingly, the
different meanings are so different from each other that it is possible to
find both statements that articles in English are salient and that they
are not salient – not because there is an actual difference of opinion
between the writers, but because what seems to be meant by “salience”
is completely different, as illustrated by the following two quotations:

… the SLA literature has clearly shown that English articles are
considered to be a nonsalient feature. They are not phonologically

10 Or saliency, with the two terms often being used as synonyms in the literature.
4.2. Theoretical approaches to difficulty 75

salient, and misuse of articles rarely leads to communication break-

down …
(Sheen, 2007b, p. 262)

The L1 also provides a platform for expectations and, hence, sali-

ence. For example, a learner who does not expect articles by virtue
of not having them in their L1 will not expect them in L2. When the
learner encounters them in the L2, the unexpectedness of articles
creates a space for their salience.
(Gass et al., 2018)

Rather than list the different understandings of salience that can

be inferred from various sources (they are very numerous, and it
is unlikely that one could arrive at an exhaustive list), let us focus
on the two that are important for discussing articles. The first one
refers to the phonetic features of the item. For example, Goldschnei-
der and DeKeyser (2005) measure the “perceptual salience” of an item
by looking at its phonetic substance (number of phones), syllabic-
ity, and total relative sonority. In their meta-analysis of morpheme
acquisition studies, they found that perceptual salience, thus defined,
was the best predictor of the sequence of acquisition for morphemes,
stronger than frequency or morphophonological regularity. It seems
that in this sense (understood in terms of phonetic qualities), sali-
ence would not be the outcome of various input-related properties,
as represented in the Housen and Simoens framework; rather, sali-
ence would be a combination of intrinsic features of the morpheme.
Goldschneider and DeKeyser (2005) thus measure perceptual sali-
ence by computing “the number of phones in the functor (phonetic
substance), the presence/absence of a vowel in the surface form (syl-
labicity), and the total relative sonority of the functor” (p. 48). When
compared to most words found in an English text, it becomes evident
that the and a/an possess fewer phones than most words, that they
76 Chapter 4. Articles as a source of difficulty in SLA

contain only one syllable and vowel, and that as a consequence their
total sonority (as calculated by the total number of points of each
phone on the sonority hierarchy) is below average.
It should be noted that the salience of articles is even lower in
spoken than in written language. Parrott makes the observation that
because articles are pronounced in very weak form, learners may fail
to notice them, and that learners who are exposed only to spoken Eng-
lish, but read relatively little, may be at a disadvantage when it comes
to the acquisition of articles (Parrott, 2000, p. 50).
The other meaning of salience, which is often evoked in comments
about the article system, seems to be more general, in the sense of
“noticeable.” This is, however, not really an objective characteristic
of articles, but rather a combination of factors. As Nick Ellis explains
(2016), ultimately salience depends not only on the features of the
stimulus, but on the context and the learner. In this view, salience is
ultimately in the eye of the beholder, as it is an outcome of certain
expectations.
Statements of this kind – that articles are not salient in the sense
of “noticeable” – are related to a number of observations made in the
literature that articles tend to be overlooked by learners in the input.
This is could be caused by both a lack of salience in the narrow, pho-
netic sense (e.g. Master, 2002), or to the fact that articles, like function
words, tend to be overlooked when L2 learners process the language
for meaning (Ekiert, 2004).

4.2.8. Context- and learner-related difficulty

The middle section of the diagram in Figure 1 covers context-re-

lated aspects, that is, learning conditions, which may constitute yet
another important source of difficulty. Although the graph provided
by the creators of the framework does not indicate this, it seems
4.2. Theoretical approaches to difficulty 77

reasonable to assume that the input-related properties mostly affect

the naturalistic (implicit) mode of the two modes which we find
under “learning conditions,” whereas the feature-specific conditions
affect the explicit mode, as they shape the metalinguistic descrip-
tions of language rules. For implicit learning, the already-mentioned
lack of transparency and low salience are predictors of difficulty.
For explicit learning, the form-related aspects are relatively unprob-
lematic, but the functional complexity of articles is high. Therefore,
it comes as no surprise that the metalinguistic propositions (rule
descriptions) used in the explicit teaching of article use are numer-
ous and complicated. Another issue is how useful the propositions
are and to what extent learners can profit from instruction, in other
words, the question concerns the teachability of articles. This is an
important issue, and will be taken up and examined in more detail
later in this book. The implicit-explicit distinction is also impor-
tant, and will be discussed later in this chapter.
Among individual learner characteristics, pride of place must go
to the languages already known by the learner, as evidenced by the
numerous studies which show that speakers of article-less languages
have less success with article use (see section 5.2). Other learner
characteristics are likely to play a role, but there is an apparent lack
of studies that connect any specific individual learner differences
with success in article use. Because some learners display persistent
problems with articles, while other individuals apparently overcome
this difficulty, it stands to reason that the ability to internalize the
English article system is related to aptitude, which is a combination
of different cognitive abilities which have been found to correlate
with success in language learning (Carroll, 1981; Dörnyei & Ske-
han, 2003). To the best of my knowledge, the use of articles by L2
learners has not been investigated with respect to aptitude, which
is not very surprising, given the limited number of studies on apti-
tude and language learning (see Yalçın & Spada, 2016 for a review).
78 Chapter 4. Articles as a source of difficulty in SLA

However, we do know that certain subcomponents of aptitude cor-

relate with success in learning specific features of a language. For
example, in one study the success in learning the passive correlated
with the component of grammatical inferencing, whereas the com-
ponent of associative memory contributed to learners’ gains on the
past progressive (Yalçın & Spada, 2016). It would be very interest-
ing to find out which of the components of language aptitude, if
any, can be linked to success in using English articles by speakers
of article-less L1s.

4.2.9. L anguage patterns versus metalinguistic

propositions and rule difficulty

A number of comments have been formulated in the earlier chapters

about the complexity of the “rules” that govern article use. To analyse
this in more detail, one needs to take a closer look at what the rules
are and consider the specific qualities that make them difficult. We
should distinguish between different senses in which the term “rule”
may be used. Though various such distinctions have been proposed
(e.g. Dietz, 2002), for my present purposes the most useful appears to
be a simple distinction between a) regularity in language (I will call it
a “pattern”), and b) the description of such a pattern which attempts
to capture that regularity by means of a metalinguistic description.
Within the second category, various descriptions are offered within
the frameworks of different linguistic theories, for example struc-
tural, generative, functional or cognitive. Among the many different
linguistic descriptions, there is a special sub-category of descriptions
that are typically pedagogical, aimed at L2 learners. These pedagogical
rules may be based on any linguistic theory; however, most pedagog-
ical grammars rely on descriptive grammars (R. Ellis, 2006). Under-
standably, pedagogical rules are usually simplified when compared
4.2. Theoretical approaches to difficulty 79

to full descriptive accounts of a given language feature. As Housen

and Simoens note, pedagogical rules may be very different from the
actual target language features which they are attempting to describe,
as they often are “perceived or intuited ‘rule-like’ patterns of covar-
iance in the input” (Housen & Simoens, 2016, p. 169), and the same
pattern may have very different metalinguistic descriptions.
Although rules and the reality can be presented as separate enti-
ties, and despite the usefulness of distinguishing between the ped-
agogical rules and the actual reality of the language that such rules
try to capture, there must be some degree of correspondence between
the level of complication of one and the other. It would be unusual
to encounter an exceedingly complicated metalinguistic proposition
delineating the rules for the use of the third person singular –s end-
ing in English verbs in the present simple tense. The pattern is sim-
ple and works without exceptions. In contrast, a set of metalinguistic
propositions explaining how to form conditional sentences, no matter
how concise, must be more extensive.
In some studies, the distinction between a “rule” as a pattern in
language, and a “rule” as a metalinguistic proposition is not imple-
mented. For example, in a study by Scheffler (2011), “rule difficulty”
(that the teachers in the study were supposed to rate) is a label that
encompasses the structural and functional complexity of a gram-
matical pattern, as well as the difficulty of the pedagogical rule as
such (the metalinguistic proposition in itself). No distinction is made
between the two, which makes sense because the study aims at assess-
ing teachers’ judgment of rule difficulty, and it is quite likely that
in teachers’ and students’ minds, the two are conflated. However,
for the purpose of analysing the difficulty of the article system, the
distinction between the actual language patterns and the linguis-
tic descriptions of those is crucial, because there appears to be less
overlap between one and the other in the case of articles than in the
case of most syntactic patterns. I will therefore retain the distinction
80 Chapter 4. Articles as a source of difficulty in SLA

between language patterns and metalinguistic propositions in the

discussion that follows.

4.3. Implicit versus explicit learning

and knowledge

Rod Ellis argues that “what constitutes learning difficulty needs to be

considered separately for implicit and explicit knowledge” (2008, p. 4).
The terms “explicit” and “implicit” have been used in SLA research
for more than 35 years,11 and many studies are based on the assump-
tion that “explicit and implicit knowledge of language are represented
differently and that explicit and implicit learning of language involve
distinct cognitive processes” (Silva & Roehr-Brackin, 2016, p. 318).
The core meaning of the implicit-explicit distinction appears
to revolve around the presence or absence of awareness (Andringa
& Rebuschat, 2015; Hulstijn, 2005). Explicit learning, is “character-
ized by the learner’s conscious and deliberate attempt to master some
material or solve a problem” (Dörnyei, 2009, p. 136). It has also been
defined as “learning to think and talk about the language system in
symbolic terms – in terms of rules and their exceptions – by com-
mitting these to memory through practice and rehearsal” (Andringa
& Rebuschat, 2015, p. 188). The learners may be presented with explicit
rules which explain how a particular language feature works, or they
may themselves form hypothetical rules on the basis of available evi-
dence, but if they do it consciously, it also constitutes explicit learn-
ing. Explicit learning logically leads to explicit knowledge, which
is also often described as the kind of knowledge that can be called

11 The distinction is related to, and can be seen as going back to, Stephen Krashen’s
“learning” versus “acquisition” dichotomy (see DeKeyser, 2003).
4.3. Implicit versus explicit learning and knowledge 81

up on demand or articulated in a verbal statement (Silva & Roehr-

Brackin, 2016).
Implicit knowledge, in turn, is usually described as intuitive, and
one that cannot be brought into awareness or articulated (Dörnyei,
2009; Hulstijn, 2005). That type of knowledge seems to derive from
implicit learning, which DeKeyser defines as “learning without
awareness of what is being learned” (2003, p. 314). Whereas drawing
on explicit learning is considered to be effortful, implicit knowledge
is often associated with automatic processing (Hulstijn, 2005).
It has to be noted that the distinction between explicit and implicit
knowledge, and explicit and implicit learning, though accepted by
many scholars, is not universally recognized as valid. It also gener-
ates disagreement between researchers when those notions are ana-
lysed in finer detail.12 Explicit learning involves a number of differ-
ent constructs, which are themselves challenging and complicated to
investigate, such as awareness, attention and noticing (see Leow, 2015).
Many controversies remain as to the exact nature of implicit learning
(see Godfroid, 2016 for a review).
It has been argued that implicit learning results from an uncon-
scious, automatic, frequency-based process of induction (N. C. Ellis,
2005). N. C. Ellis gives the following extensive description of implicit
learning:

(T)he bulk of language acquisition is implicit learning from usage.

Most knowledge is tacit knowledge; most learning is implicit; the
vast majority of our cognitive processing is unconscious. Implicit
learning supplies a distributional analysis of the problem space: Fre-
quency of usage determines availability of representation according
to the power law of learning, and this process tallies the likelihoods

12 For a review of those issues, see DeKeyser (2003).

82 Chapter 4. Articles as a source of difficulty in SLA

of occurrence of constructions and the relative probabilities of their

mappings between aspects of form and interpretations, with gener-
alizations arising from conspiracies of memorized utterances col-
laborating in productive schematic linguistic constructions. (…)
Implicit learning also forges serial associations, synthesizing collo-
cations, larger formulas, and composite constructions by chunking
together contiguous components, thus creating hierarchical organ-
izational structures. (…) Once associated, the components stimu-
late each other, via these connections, in the spreading activation
of the cognitive unconscious. Related exemplars thus work together
in implicit memory, their likenesses harmonizing into an attractor
state, and it is by these means that linguistic prototypes and catego-
ries emerge. (…) Implicit learning, operating throughout primary
and secondary neocortical sensory and motor areas, collates the evi-
dence of language, and the results of this tallying provide an opti-
mal solution to the problem space of form-function mappings and
their contextualized use. The representational systems modularize
over thousands of hours on task. In these ways, unconscious learn-
ing processes, which occur automatically during language usage, are
necessary in developing the rationality of fluency.
(N. C. Ellis, 2005, pp. 306–307)

However, one should note that implicit learning has its limitations,
since it exploits resources other than those involves in explicit earn-
ing, especially in relation to metacognition (Dienes & Perner, 2002).
To that effect, Nick Ellis mentions in the same source that “many
aspects of a second language are unlearnable – or at best are acquired
very slowly – from implicit processes alone” (N. C. Ellis, 2005, p. 307).
Unfortunately, the author does not provide information as to which
aspects of SLA are alluded to by that remark.
The view that learning is unconscious, automatic, and frequen-
cy-based is consistent with the position which holds that the primary
4.3. Implicit versus explicit learning and knowledge 83

language learning mechanism is statistical learning (Rebuschat &

Williams, 2012; Romberg & Saffran, 2010), which can be defined as “a
gradual process of accumulating linguistic knowledge based on the
distributional properties of the input” (Andringa & Rebuschat, 2015,
p. 188). In this view, humans are sensitive to frequencies of (co-)occur-
rence, and, without being consciously aware of it, they accumulate
this frequency information until patterns of grammatical structure
emerge. This phenomenon is of central importance to phraseological
aspects of language learning, which will be discussed in Chapter 7.
Even if most learning is indeed implicit, for the learning of spe-
cific L2 features over a limited period of time, explicit instruction is
by far more efficient. De Keyser (2003) provides a review of studies
which directly compared explicit and implicit instruction (the lat-
ter may involve input flooding, or presenting learners with a large
amount of input containing the target pattern). Those studies, con-
ducted in the form of experiments with artificial languages, as well
as classroom experiments, invariably show that the explicit instruc-
tion groups strongly outperform implicit groups. This confirms the
status of metalinguistic instruction as an invaluable way to speed up
the learning process.
However, it is believed that, as opposed to the use of implicit
knowledge, which is automatic and effortless, the use of expli-
cit knowledge, i.e., relying on the conscious application of explicit rules
in language comprehension and production, must be more cognitively
demanding for the language user, as it taxes working memory resources
(Silva & Roehr-Brackin, 2016). Therefore, it would be ideal if the learner
could move beyond explicit knowledge to the faster, effortless implicit
knowledge. In other words, hopefully the development of explicit knowl-
edge in an L2 learner is not the final state, an end in itself, but an aid in
the development of implicit knowledge. This brings us to an issue of cru-
cial importance, namely, the relationship between implicit and explicit
learning and knowledge. Does explicit knowledge help in the formation
84 Chapter 4. Articles as a source of difficulty in SLA

of implicit knowledge? Anyone involved in either learning or teach-

ing an L2 must certainly hope that it does, otherwise the prospects for
instructed L2 learning would be rather bleak: instructed learners would
forever have to rely on the effortful process of applying explicit rules
(which are never entirely accurate) when producing language output.
This pessimistic view, known in the literature as the no-inter-
face position,13 is generally considered unlikely (see e.g. Andringa
& Rebuschat, 2015 for an overview), especially as it goes contrary to
the bulk of research on automaticity that has been carried out with
respect to language since the late 1970s. For example, Segalowitz inter-
preted neurophysiological evidence “to mean that explicit knowledge
forms a prerequisite for implicit knowledge to come into existence”
(2005, p. 378) and conducted a review of studies which suggest that
“already in the initial phases of learning, implicit knowledge is spon-
taneously formed, and that explicit processes are simply not used any
longer in the later phases” (2005, p. 378).
The discussion centres instead on which of the two other views,
known as the strong- and weak-interface positions, is true. The for-
mer holds that explicit knowledge can turn into implicit knowledge
as the result of practice; in the latter view, this is possible only to some
extent, or under some conditions. Explicit instruction and metalin-
guistic information can be seen as a way of building up the target
associations faster, a kind of shortcut to achieving a structure which
would take much longer to emerge from the statistical learning pro-
cess. DeKeyser’s view, for example, is that explicit knowledge can
become so automated with time and practice that it becomes impos-
sible to distinguish it from implicit knowledge. Also, he points out
that some learners, as a result of rule automatization and extensive

13 The no-interface position can be seen (DeKeyser, 2003) as a continuation of

Krashen’s view that “learned competence does not become acquired competence”
(Krashen, 1985, p. 42).
4.3. Implicit versus explicit learning and knowledge 85

communicative use, may gradually lose their awareness of the explicit

rules they once held in memory (DeKeyser, 2003). A weaker position
is supported by Nick Ellis, who sees explicit instruction as a way to
direct attention towards particular formal features of the input, which
in turn affects the statistical, frequency-based uptake of those features
(N. C. Ellis, 2015).
Another consideration which is important for the present discus-
sion is the fact that no linguistic description, and – even more so – no
pedagogical rule (because they have to be simple) can fully capture
or depict the language patterns which they try to describe. However,
I would argue that some language features are “depictable” (for lack
of a better term). This “depictability” is low in the case of articles, as
evidenced by the multiplicity of metalinguistic propositions needed
to describe the way they are used, the numerous exceptions, and a fair
amount of optionality.
To take this thought further, the easily “depictable” language pat-
terns are the ones for which there is extensive (maybe even complete?)
overlap between the language pattern and the explicit rule. It stands
to reason that in such cases, highly proceduralized explicit knowl-
edge may be virtually impossible to distinguish, in a learner’s perfor-
mance, from implicit knowledge. However, with poorly “depictable”
patterns – such as the English articles – even a highly proceduralized
knowledge of an extensive list of rules will not result in article use
that is comparable to native speaker use.
This fact can be explained by differences in the nature of explicit
and implicit knowledge. Certain inherent qualities of explicit and
implicit knowledge make them a better match for some linguistic
patterns and not for others. We find a pertinent observation on this
topic in a publication by Roehr-Brackin:

… it has been argued that explicit knowledge is characterized

by stable, discrete, and context-independent category structure.
86 Chapter 4. Articles as a source of difficulty in SLA

Correspondingly, linguistic constructions which show compara-

tively systematic, stable, and context-independent usage patterns
can be described more easily by means of metalinguistic rules. Such
linguistic constructions should be more amenable to explicit learning
than linguistic constructions with less systematic and more context­-
-dependent usage patterns that require a greater number of, or more
complex metalinguistic rules to describe them adequately.
(Roehr-Brackin, 2015, p. 126)

Thinking of English articles, this implies they are less amenable

to explicit learning, as they require a great number of metalinguistic
rules. They are more likely to be successfully used if learnt implicitly,
as “implicit representations with flexible, context-dependent cate-
gory structure can fully capture prototype effects and distributional
effects” (Silva & Roehr-Brackin, 2016, p. 319), so they result in knowl-
edge that is both reliable and accurate.

4.4. Conclusions

There are three main conclusions which emerge if the theoretical

considerations on difficulty in SLA presented in the above review
of literature are applied to the specific characteristics of the English
article system.

4.4.1. High difficulty level

The first conclusion is that both implicit and explicit learning of arti-
cles is highly difficult. An examination, from a theoretical perspec-
tive, of the degree of difficulty posed by the article system in the
4.4. Conclusions 87

learning of L2 English, as presented in this chapter, confirms that there

are indeed very sound reasons to regard the difficulty level as high.
Three main factors contribute to the difficulty posed by articles. Two
of these are feature-related characteristics that have a negative impact
on implicit learning. The first element is intrinsic to the articles them-
selves: Despite the fact that English articles are of low formal complex-
ity, their functional complexity is high, which leads to high form-func-
tion opacity (one form may serve many functions). The second element
is related to language input: despite their high frequency, articles are
characterized by communicative redundancy and optionality, which
means that articles are seldom needed to ensure understanding and
sometimes more than one option is acceptable in the same context.
The combination of form-function opacity and optionality con-
tributes to low transparency for articles, while their morpho-pho-
nological composition, communicative redundancy, and optionality
contribute to their low salience.
The third factor that makes articles difficult to master is especially
relevant for explicit learning. Grammar books and pedagogical man-
uals contain a wealth of highly difficult descriptions to account for
how English articles work (metalinguistic propositions), mostly due
to a multiplicity of rules that govern article use.

4.4.2. Implicit learning

The second conclusion is that articles have to be learned implicitly.

Due to the fact that article use is not well “capturable” by means
of metalinguistic propositions, the explicit knowledge of article use
(even highly proceduralized) does not enable the learner to use arti-
cles in a fully target-like way. Therefore, articles have to be learned
implicitly to make the use fully native-like, even though this process
is difficult, as was just stated in the first conclusion.
88 Chapter 4. Articles as a source of difficulty in SLA

Another reason that the knowledge of article use must be highly

automatized is related to frequency. Articles are very frequent in the
input, which normally lowers the difficulty level for acquisition. How-
ever, their frequency brings about difficulties in the production of L2
utterances: consistent rule application, given that the rules are very
complicated, is effortful and puts a strain on the learner’s cognitive
resources.

4.4.3. L imited usefulness of metalinguistic

propositions

The third conclusion is that metalinguistic rules show little promise for
the teaching of articles. Chapter 6 will look at specific research findings
concerning the teaching of articles. However, this chapter allows us to
make some predictions on this topic. On the basis of theoretical con-
siderations only, it seems that the chances of learning to use articles in
a completely target-like manner as the result of explicit instruction are
not very high, not because explicit instruction is generally ineffective
(as was mentioned above, it has proven effective) but because articles
do not lend themselves easily to explicit instruction. This indeed seems
to be the prevailing view among language teachers. For example, Sheen
(in a study which will be discussed later in this book) gives the follow-
ing description of the situation in ESL classes at an American college:

… a series of discussions with the participating faculty members

at the college (…) revealed that (a) participating students are not
explicitly taught articles during the semester, and (b) articles, though
constituting a structure where students commonly make errors, are
infrequently corrected because they are nonsalient and they require
complicated rule explanations.
(Sheen, 2007b, p. 262)
4.4. Conclusions 89

However, as was said above, explicit rules are valuable and

time-saving in instructional settings. If they are of somewhat lim-
ited use as far as articles go, and fully implicit learning of the articles is
not realistic given the time constraints of L2 learning14 (no input flood
has a chance of making a difference to the learner’s implicit knowl-
edge), it appears that the only potential for helping learners resides
in exploring the possibility of awareness raising. Noticing tends to
be hindered by the relatively non-salient character of articles and, as
suggested by Yoo (2009), articles have suffered as a result of commu-
nicative approaches to language teaching, which emphasise fluency
over accuracy. Encouraging noticing may lead to some improvement
of the normally very slow processes of implicit learning.

14 And given possible age effects, as there is also the possibility that implicit learn-
ing, even with a sufficient amount of input, may be constrained by maturational
constraints (see DeKeyser, 2003); but it needs to be remembered that the specific
nature and extent of maturational constraints is a matter of much controversy (see
Muñoz & Singleton, 2011). At this moment there seems to be no reliable informa-
tion regarding how age impacts the capacity for implicit vs explicit learning (see
Andringa & Rebuschat, 2015; DeKeyser, 2003; Verneau, van der Kamp, Savels-
bergh, & de Looze, 2014).
5.1. Introduction 91

Chapter 5

Articles in SLA research

5.1. Introduction

In view of the immense complexity of the English article system, as

well as its high level of difficulty, as discussed in the preceding chap-
ters, the challenge facing learners of English as a second language
seems considerable. The previous chapter has shown the numerous
reasons why articles may cause difficulty for ESL learners, especially
for those whose L1 contains no articles. As has already been men-
tioned, research findings confirm the significant difficulties that
learners encounter with articles, but also provide information on
other aspects of article acquisition and use. The overview below pro-
vides the most important facts about articles in L2 acquisition that
have been established by researchers.
It should be noted here that information about the acquisition and
use of articles is available from studies with a wide range of research
topics, since any study dealing with learner language may gather
information on the learners’ use of articles, among many other lan-
guage features. This makes it impossible to consider every single study
92 Chapter 5. Articles in SLA research

which mentions articles, and the review of literature presented in this

chapter is necessarily selective, discussing mostly studies which are
concerned primarily with articles, but also selected ones in which
articles were not the main focus of inquiry.

5.2. Crosslinguistic aspects

From the vast body of research on articles, the finding about articles
in L2 English which emerges with by far the greatest robustness and
clarity is the observation that learners who do not have an article sys-
tem in their L1 find it more difficult to acquire articles in an L2. The
first observations about the crosslinguistic effects of learners’ article
use were made already in the early days of second language acquisi-
tion research. An important paradigm in the emerging field of applied
linguistics, derived from Bloomfieldian linguistics, was that of con-
trastive analysis. Its basic assumption was that areas of difficulty in
language learning can be identified and anticipated for speakers of
a specific L1 on the basis of a comparison of the learner’s language
with the target language. The more different the rules were, the more
problematic their learning was expected to be in an L2. It is not sur-
prising that crosslinguistic aspects of the use of articles attracted con-
siderable interest, since this topic was perfectly suited for contrastive
analysis and the other dominant approaches of the day, error anal-
ysis and transfer studies. In one of the early studies on articles in
a crosslinguistic context, Dušková (1969) attributed the article errors
of Czech learners of English to the absence of articles in Czech. Pol-
ish linguists were among the first ones to publish observations on
the problems caused by articles in the process of learning English by
speakers of an article-less L1 (Arabski, 1968, 1990; Kałuża, 1963); it
appears that already in the early days of English language instruction
5.2. Crosslinguistic aspects 93

in Poland it was a recognized fact that “since articles do not exist in

Polish, learners misuse them in many possible ways” (Arabski, 1968,
p. 84).
The topic seems to have attracted less interest in the following
decades, but evidence continued to accumulate on the importance of
crosslinguistic factors in article acquisition and use. Larsen-Freeman,
in a study of morpheme acquisition orders (1975), found out that the
rank occupied by articles in morpheme accuracy studies was lower
for speakers of article-less languages (e.g. Japanese) than for speak-
ers of L1 that contain articles. A number of studies on Swedish and
Finnish learners (reviewed by Ringbom, 1985) showed that Swedes,
whose L1 has articles, make fewer errors than Finns, whose L1 does
not have them.
More recently, a very interesting study by Jarvis (2002) yielded
overwhelming evidence supporting the hypothesis that the use of arti-
cles in L2 is influenced by the article system of the L1. The study design
involved eliciting narratives in English from native speakers of Amer-
ican English, Swedish, and Finnish. The Swedes and the Finns addi-
tionally wrote the same narratives in their L1. The author commented
that, in terms of article use, the English texts written by Americans
and Swedes “are, in fact, so conspicuously similar that they leave little
doubt concerning the (facilitating) influence of L1 Swedish in this area
of interlanguage performance” (Jarvis, 2002, p. 401). The Finnish data
show a striking underuse of articles, and also points to the possibility
of L3 influence, because those speakers of L1 Finnish who also studied
Swedish displayed some advantage over non-Swedish-speaking Finns.
Success in article use is thus proven to be strongly dependent on
crosslinguistic factors, with speakers of article-less languages finding
English articles much more problematic than speakers of languages
which feature independent prenominal article-like morphemes, who
profit from facilitative effects of crosslinguistic similarity (Chra-
baszcz & Jiang, 2014; R. Hawkins et al., 2006; Ionin, Zubizarreta,
94 Chapter 5. Articles in SLA research

& Maldonado, 2008; Snape, 2008; Zdorenko & Paradis, 2008).15 This
is in keeping with the growing recognition of the ubiquitous charac-
ter of crosslinguistic factors in contemporary second language acqui-
sition studies (Arabski & Wojtaszek, 2016).

5.3. S emantic universals and the acquisition

of articles

While there is plenty of evidence that speakers of languages which do

not feature articles have greater difficulty learning how to use them
in English, it is less clear how exactly that difficulty is to be accounted
for in theoretical terms. The simplest approach would be to think that
the semantic and discoursal roles played by articles (the marking
of specificity and assumed hearer knowledge, the latter overlapping
with the discoursal reference to new or reintroduced referents) are
unknown to the language learner with an article-less L1. The learner
has to learn new semantic and discourse knowledge, instead of just
mapping a new form onto a meaning that is already known from the
L1 (as in, for example, the past tense marking). However, this would
be an oversimplification, as it is true that (as was mentioned above in
the section on information flow) referentiality and definiteness are
also known in article-less languages, at least to some extent. However,
they are expressed through linguistic means so different that they are
not normally apparent to language users (see e.g. Firbas, 1992).

15 Even so, the acquisition of the English article system is not entirely unproblem-
atic for speakers of languages which feature articles, as there are always some
differences in article use between languages. For example, the French language
has no zero-article, but it has an additional category of articles called partitives,
which have no equivalent in English.
96 Chapter 5. Articles in SLA research

A vast majority of studies on articles are based on Bickerton’s con-

ceptualization of NP reference and discourse function, or subsequent
developments of that conceptualization (e.g. Chaudron & Parker,
1990; Díez-Bedmar & Papp, 2008; Tarone & Parrish, 1988). Those arti-
cles uses which fall outside of the semantic wheel were labelled by sub-
sequent researchers as “idiomatic uses” (Butler, 2002; Thomas, 1989).
This fifth category of article use includes expressions such as “in the
1970s,” “all of a sudden” and “living hand to mouth,” and contains all
three article choices: the, a, and Ø.
A typical taxonomy used in research on ESL articles is presented in
Table 1; it is a shortened version of the classification used by Díez-Bed-
mar and Papp (2008), who adapted it from earlier researchers (Butler,
2002; Ekiert, 2004; Huebner, 1985; Thomas, 1989). Most studies that
look into selected, specific features of NP reference follow this tax-
onomy and thus focus on learners’ sensitivity to the features of defi-
niteness and specificity.
Features Environment Articles Examples
Type 1 generic nouns a, the, Ø An elephant never
[–SR,+HK] forgets.
They say that the
elephant never forgets.
Elephants have
trunks.
Type 2 referential definites the Pass me the pen.
[+SR,+HK], previous mentions I found a book. The
specified by entailment book was …
or definition
The first person to
unique in all contexts
walk on the moon…
/ in a given context
Type 3 referential indefinites a, Ø A man phoned.
[+SR,–HK] first-mention nouns I keep sending
messages to him.
5.3. Semantic universals and the acquisition of articles 97

Features Environment Articles Examples

Type 4 non-referential nouns a, Ø Alice is an accountant.
[–SR,–HK] attributive indefinites I need a new car.
non-specific indefinites
Foreigners would
come up with a better
solution.
Type 5 idioms and conventional a, the, Ø all of a sudden,
uses in the 1950s,
living hand to mouth

Table 1. A typical taxonomy used in research on ESL articles (based on Díez-Bedmar & Papp,
2008).

The central question in this kind of research is whether the learners

seem to have an innate – and therefore universal – sensitivity to cer-
tain semantic, syntactic, and discourse distinctions and, if so, how
and when they mark such distinctions. The rationale behind investi-
gating “universals” is the assumption that all languages have means
of indicating reference, achieving topic continuity, and thus making
sense of connected discourse. The marking of specificity and hearer
knowledge may be present in other languages, even though they may
be realized in other ways than through articles. Thus, on a general
level, all learners of English “know” the function of the article sys-
tem, as it is an alleged part of the Universal Grammar. They will,
therefore, tend to mark it in one way or another, possibly using dif-
ferent formal means than native speakers of English. For example,
researchers may be interested in whether, at the beginning stages
of ESL learning, the functions of the article system are realized in
interlanguage by means of other features transferred from the L1,
for example, as hypothesized in the functional sentence perspective
(Firbas, 1992), which requires that new information be positioned
toward the end of the sentence.
98 Chapter 5. Articles in SLA research

As was mentioned in section 3.4, in generative studies, the acquisi-

tion of a new article system by language learners is seen as a process of
parameter resetting. Various parameter-based hypotheses have been
put forward to explain the mastery of articles, or persistent problems
with this language feature. Providing an overview of such studies,
Roger Hawkins (2001) offers a syntactical explanation of that defi-
nite-then-indefinite sequence in terms of complement specification
and co-indexation for both article categories, denying any notable
impact of the learner’s L1 on the observed sequence of emergence in
L2 English. However, some other researchers in Generative Gram-
mar do acknowledge the impact of the learner’s L1 when they inter-
pret their results. Among those, Ionin, Ko & Wexler (2004) suggest
their “fluctuation hypothesis,” where the binary parameter for arti-
cles is definiteness/specificity. They argue that languages like English
possess an article system based on definiteness, whereas in other lan-
guages, like Samoan, it is based on specificity.16 Therefore, a learner of
English whose L1 is French (also based on definiteness) would benefit
from assigning the [+definite] to English, while a learner whose L1
contains no articles – such as Polish – would fluctuate for a certain
amount of time between definiteness and specificity, thus mixing
more a/an and the and producing a higher number of omitted arti-
cles (see Zdorenko & Paradis, 2008).
The fluctuation hypothesis was supported by Ionin, Ko & Wex-
ler’s study (2004). The authors gave a multiple-choice task to Korean
and Russian learners of English, in which the learners were expected
to choose one of the three possibilities (a, the, zero) in a given slot.

16 For example, in English the definite article can be used even when referring to
a specific entity (as was discussed in section 1.3). For example, it is possible to say
“I want to talk to the manager” without knowing who that person is. In Samoan,
the article would vary depending on whether the article refers or not to a specific
entity.
5.4. The sequence of acquisition of articles in L1 and L2 English 99

Accuracy in the use of articles was better when definiteness and speci-
ficity had the same value. In other words, the [+definite, +specific] and
[–definite, –specific] contexts yielded the highest accuracy rates. The
options were available only with singular nouns. The learners were
less successful with items which were [+definite, –specific] or [–defi-
nite, +specific]. The authors interpreted this to mean that the learn-
ers had yet to settle on the feature of definiteness. Similarly, Ionin,
Zubizarreta, and Maldonado (2008) obtained results supporting the
fluctuation hypothesis. Such fluctuation was also found by Zdorenko
and Paradis (2008), whose child participants from both types of L1s
(with and without articles) used both the definite and the indefinite
article for the [–definite, +specific] context.

5.4. The sequence of acquisition of articles

in L1 and L2 English

The acquisition of articles in L1 English is, in contrast to the L2 situa-

tion, relatively unproblematic, although it takes some time to develop.
Brown (1973) was the first to point out that articles are not very sim-
ple to master in comparison to other grammatical morphemes, since
they ranked 8th out of the 14 morphemes he examined for the three
children whose first language acquisition he followed. This means that
articles appear in children’s speech even after irregular past endings
and the possessive ’s. Subsequent studies investigating the order of
acquisition of grammatical morphemes have confirmed that articles
tend to be acquired quite late in comparison to other morphemes (for
an overview, see Pawlak, 2006, pp. 103–107). Regarding the moment
they become mastered, Warden (1976) showed that, at the age of three,
children still use the definite article about half the time to introduce
a noun in telling a story, instead of an indefinite article, which adults
100 Chapter 5. Articles in SLA research

would use. The bulk of research conducted since then suggests that
children seem to acquire the article system somewhere between 2;8
and 3;8 years of age, at which time they hardly make any article errors.
If there are errors, it is mostly the overuse of the definite article when
the hearer does not in fact have any knowledge of the referent (i.e.,
the [+SR, –HR] case), but they tend not to make errors when the ref-
erent is nonspecific and assumed unknown to the hearer, that is, in
the [–SR, –HK] case (Brown, 1973; Maratsos, 1976).
This seems to make sense – children may not yet be able to make
correct assumptions about the listener’s knowledge, as this ability
develops with experience and knowledge of the surrounding world.
They do, however, seem to distinguish without any difficulty between
specificity and nonspecificity, even though there are no physical fea-
tures of the referents that make them specific or nonspecific as such
(Maratsos, 1976, p. 94). It has been suggested that children have an
innate sensitivity to specificity and nonspecificity, as assumed by
Bickerton’s (1981) bioprogram hypothesis.
According to a synthesis of studies on article acquisition in L1
(Cziko, 1986), four stages of this process can be distinguished:
• at Stage 1, all referential nouns, both [+HK] and [–HK], are
marked with a or the, but articles are not used with nonrefer-
ential nouns,
• at Stage 2, the is used in [+SR] contexts and a in [–SR] contexts,
• at Stage 3, sensitivity to the feature [±HK] appears,
• at Stage 4, the target system of articles is used, that is, both fea-
tures [±SR] and [±HK] are taken into consideration.
As far as the early stages of L2 acquisition are concerned, the
most common finding that emerges from the available literature
is that learners commonly overuse the definite article (Bitchener
& Knoch, 2010; Chaudron & Parker, 1990; Huebner, 1983; Master,
1997; Young, 1996). This phenomenon is sometimes referred to as
“the-flooding.”
5.4. The sequence of acquisition of articles in L1 and L2 English 101

For speakers of L1s which feature articles, the findings appear to

be mixed as to the sequence of a/an and the (see, e.g. Díez-Bedmar
and Papp (2008) versus Thomas (1989)), while both the definite and
the indefinite article seem to be mastered earlier than the zero article.
For learners with article-less L1s, studies on the acquisition of the
definite article the and the indefinite article a also yield somewhat
conflicting results. It seems that in the majority of studies, the definite
article is used correctly earlier than the indefinite one (Master, 1997;
Parrish, 1987; Sun, 2016; Thomas, 1989), but there are studies which
suggest the contrary (Díez-Bedmar & Papp, 2008; Liu & Gleason,
2002; Young, 1996). Park (1996) examined ESL learners whose L1 had
articles (French, German) and others who did not (Korean and Jap-
anese). His results showed that definite-first sequence of acquisition
(definite first, then indefinite) is observed whether or not the L1 pos-
sessed an article system. Of special interest is a study by Ekiert (2004),
as the participants were speakers of L1 Polish. Ekiert’s results show
that the first article to be used by learners was a (in nonreferential and
first-mention contexts). Additionally, the learner’s output was charac-
terized by a very substantial overuse of the zero article, which means
underuse of the overt articles. Finally, of all the contexts of article, it
was the generic uses, as well as the idiomatic ones, that caused Polish
learners the greatest difficulty.
Similarly to Ekiert’s study, the overuse of the zero article has been
observed in many other studies with learners from article-less lan-
guage backgrounds (Díez-Bedmar & Papp, 2008; Master, 1997; Par-
rish, 1987). This underuse of articles which have an overt form persists
in the case of learners from article-less languages, sometimes even at
high levels of proficiency (Ekiert, 2004; Kharma, 1981; Master, 1997).
Even the most advanced Japanese learners in Butler’s study, remained
well below target norms with respect to article use. She notes that “the
finer complexities of English article usage continue to challenge learn-
ers’ abilities to correctly use articles” (Butler, 2002, p. 472). In Sun’s
102 Chapter 5. Articles in SLA research

(2016) study with 18 adult ESL learners from various language back-
grounds doing a cloze test on articles, the main difference between
speakers of L1s with and without articles resided in the zero arti-
cle, which turned out to be the most challenging of article types. At
beginner and intermediate levels, the participants with article-less L1s
scored much lower for zero articles than their counterparts with arti-
cles in their L1. At advanced levels, the means for zero articles were
comparable across groups (62% for each language type), suggesting
that the zero article still poses challenges for advanced learners, due
mostly to the aspect of countability, on which more will be said in
the next section.
The acquisition of articles by Polish learners of English was also
investigated by Świątek (2013, 2014). Participants at three proficiency
levels had to supply missing articles in sentences containing obliga-
tory contexts for the use of an article (the sentences were adapted from
earlier studies). The results were interpreted by the author as showing
that, generally, the sequence of L2 article acquisition is mostly similar
to the order in L1, as evidenced by the success rates in the supplying
of the correct articles by the various groups.
A considerable amount of attention is paid in the literature to the
fact that beginning learners seem to overuse the definite article in the
[–HK, +SR] contexts, that is, they tend to incorrectly associate the use
of the definite article with specificity (Chaudron & Parker, 1990; Tar-
one & Parrish, 1988; Thomas, 1989; Young, 1996).
In Thomas’s study (1989) L2 learners overgeneralized the in
first-mention contexts [+SR, –HK] but not in [–SR, –HK] contexts,
which mirrors what happens in the case of L1 learners of English. The
author concluded that both L1 and L2 learners tend to initially asso-
ciate the use of the definite article with the feature [+SR].
Butler (2002) gave excellent examples illustrating the two trends
mentioned in this and the previous section: the incorrect use of the in
contexts perceived to be specific, and problems with article use based
5.5. Studies on countability 103

on wrong assumptions about countability. Examples of the learners’

article choices are presented below along with the explanations (pro-
vided by the learners in an interview) of why they decided to use those
articles (Butler, 2002, p. 459):

“Japan has the old culture, stretching “…this is the, because this means Japa-
back in time more than 2,500 years…” nese culture, which is specific”
“Japan has Ø old culture, stretching “I did not insert anything here because
back in time more than 2,500 years…” I thought culture is not countable”

The first example illustrated the overuse of the resulting from the
overgeneralization of the rule of specificity. The other trend is the
overuse of the zero article because of a misjudgement concerning
countability. Butler’s learners often assumed that abstract nouns are
not countable.

5.5. Studies on countability

Some fine-tuned analyses of the problems learners face with articles

point to the connection between errors in article use and the counta-
bility of nouns (Butler, 2002; B. White, 2009; Yoon, 1993). Error analy-
sis studies reveal that some article errors made by learners are actu-
ally very likely to be due to another error, namely a misspecification
of the countability or the number of a noun, as the assignment of the
count/non-count feature to nouns in interlanguage may differ from
the target (Young, 1996). Errors regarding countability may have to
do with the fact that learners at beginner levels tend to assume that
nouns are either countable or not, and may try to rely on memorized
lists of such nouns in making decisions on whether to use an article
or not. In fact, nouns can commonly be used as either countable or
104 Chapter 5. Articles in SLA research

not, depending on the intended meaning (see the discussion in 2.4),

which limits the usefulness of memorizing noun status with respect to
countability. Apparently, learners tend to treat countability as a fixed
property of specific nouns, so they are less likely to be able to use
nouns which are usually uncountable with a countable meaning in
a given context (Yoon, 1993). Butler also comments on this tendency,
stating that, at the beginning stages of language acquisition, count-
ability is treated as a “fixed or static entity” (Butler, 2002, p. 466).
In other words, there is insufficient awareness of nouns having the
capacity to function both as countable as an uncountable, depending
on the intended meaning.
It might seem that this is a problem restricted to beginners, but
Butler (2002) provides an excellent example of this problem occurring
in the language production of advanced learners. The specific type of
problem which caused a large number of errors in Butler’s study was
the use of the indefinite article with abstract nouns denoting enti-
ties which are normally perceived as indivisible, but can sometimes
be conceptually divided into different kinds, for example, a warm
environment, a cold environment. The learners in Butler’s study were
unaware of this possibility when it existed, or alternately, they over-
generalized this principle and produced word combinations such as
a warm water, a cold water.
Similar findings were obtained by White (2009). In this study,
speakers of article-less L1s (mostly Korean), advanced ESL learners,
filled in a forced-choice elicitation test. They were also asked to pro-
vide confidence ratings and explanations for their article choices.
White’s results indicate that with the definite article, the learners
were mostly sensitive to the appropriate feature – definiteness – and
were not influenced by noun type in their decisions. With the indefi-
nite article, they were sensitive to both indefiniteness and noun type.
However, with the zero article, the participants were the least sensitive
to semantic context; rather, they seemed to be sensitive to countability
5.6. Abstractness 105

and noun type. The learners’ explanations revealed that often their
wrong choices of articles were due to misidentifications of count sta-
tus for abstract count nouns and non-count ones. Also, in some cases
the learners seemed to choose articles on the basis of the nouns’ count
status alone, which resulted in the frequent use of the zero article for
nouns which were believed (wrongly) to be uncountable. White con-
cluded that some learners pay “disproportionate attention to counta-
bility,” which leads to “insensitivity to semantic context” (2009, p. 28).
Thus, a major problem for some ESL learners is that: “they consider
the count status of the noun without considering definiteness in the
semantic context. Once they perceive a noun to be uncountable, they
resort to the choice of Ø” (B. White, 2009, p. 28). White also observes
that in some studies, it is not always possible to tease apart count-
ability and the effects of specificity and definiteness. For example,
a learner’s use or non-use of the definite article may be attributed by
researchers to sensitivity to definiteness or the lack of it, whereas it
may in fact be due to mistaken assumptions concerning countability.
Because the learner has to take into consideration both countability
and the context of the NP, he attends to countability first, which over-
rides any consideration of the semantic context.

5.6. Abstractness

Another factor at play which has been identified by studies investi-

gating learners’ article use is abstractness, which in turn is related
to countability. Apparently, more errors in article use occur with
abstract nouns than with concrete nouns (Hua & Lee, 2005; Ogawa,
2008; B. White, 2009). In a study by Amuzie and Spinner (2013), the
level of accuracy in article use was attributed not only to abstractness,
but also to more nuanced aspects of abstractness, such as the nouns’
106 Chapter 5. Articles in SLA research

degree of boundedness. Similarly, in a study by Hiki (1991), learners

of English were particularly likely to misjudge the count status of
nouns in the case of abstract nouns, and were least accurate when
using articles with abstract nouns. In White’s study (2009), the par-
ticipants had significantly lower confidence ratings for their article
choices in the case of abstract nouns.

5.7. Articles and learners’ interim rules

Some light is thrown on the difficulties which learners have to confront

when using articles in English by studies which investigate the learn-
ers’ metalinguistic knowledge of the English article system, that is, the
interim rules of the learners’ interlanguage which are relied on at a given
moment. For example, Butler (2002) analysed the process of “making
sense” of the English article system by Japanese ESL learners at differ-
ent levels of proficiency, who had to complete an article-fill-in-test and
immediately afterwards participated in a structured interview. Such
studies provide a valuable addition to the more common quantitative
investigations of article use, by providing qualitative data that reveal
very interesting facts about how learners struggle to make sense of the
English article system by forming numerous ad-hoc working hypoth-
eses, then discarding them, and forming new ones. Such studies also
attest to the frustrating nature of the English article system, recording
the learners’ confusion, lack of confidence and hesitation. To quote just
one of the learners who were interviewed in Butler’s study (2002, p. 468):

I know there are some words that take the, garden, for example.
Wait, garden might not take the. I’m not confident about garden.
But I learned that park takes the. To be honest with you, I don’t know
whether this is really true or not. But I use the anyway.
5.8. Type of task 107

Another observation by Butler (2002) is that the learners tend

to rely on syntactic or structural cues which are interpreted in an
oversimplified way. For example, the learners assumed that a refer-
ence was specific if the noun was modified, which led to the produc-
tion of errors such as “I made the terrible mistake” (first-mention).
As a result of these tendencies, the indefinite article is bound to be
underused.
However, the main observation that emerges from Butler’s study is
that learners at lower to intermediate proficiency levels generate large
numbers of working hypotheses on which they try to rely to make cor-
rect article choices. It is clear that the learners can provide an expla-
nation for a large number of their article-related choices; however, in
most cases, they overgeneralize a “rule” or misapply one. The over-
all picture is that the system of rules for article use in English is such
that it is actually almost impossible for lower-level learners to apply
those rules, which means that attempts at teaching those rules may
result in more confusion than profit.

5.8. Type of task

It has been observed that different tasks yield different frequencies

of article errors (Kharma, 1981; Tarone & Parrish, 1988). Production
tasks, for example essays, or responses in interview tasks, have pro-
duced lower error rates than objective tasks, such as cloze tests. Tar-
one and Parrish (1988) investigated the use of articles by ESL learn-
ers in two different tasks – interviews and narratives – and found
a number of differences. For example, referential definites were used
more often in the narrative; they were also used more accurately.
The authors explained this observation by means of the narrator’s
need to communicate effectively: in a narrative, it is important for
108 Chapter 5. Articles in SLA research

the speaker to mark NPs in a way that makes it possible for the hearer
to keep track of the story, and this is done mostly by means of clear
anaphoric marking.

5.9. Lexical chunks

Although the studies of article use overviewed in this chapter did

not address the issue of the use of articles in lexical chunks, they
sometimes make brief references to such uses. However, such men-
tions are extremely rare; usually in the context of the limitations of
a given elicitation test or a given study. For example, White (2012)
used a forced-choice elicitation test of articles, designed specifically
to measure the learners’ sensitivity to noun type, countability, and
other characteristics. The test was accompanied by an explanations
sheet, on which the learners explained their choices. The learners
sometimes gave unexpected answers, on which the author com-
mented:

It is possible that participants’ article decisions for individual dia-

logues were influenced not solely by semantic features and noun
types, but also by previous exposure to lexical chunks. For example,
Participant 18 wrote in explanation of Item 14 that she often heard
her roommate say “I’m not in a mood.”
(B. White, 2012, p. 29)

In Young’s study (Young, 1996), all NPs were examined, but some
were not coded and thus excluded from the analysis, for a number of
reasons. Interestingly, one of the reasons for the exclusions of some
NPs is “the likelihood that they had been learned as chunks” (Young,
1996, p. 152). Unfortunately, no more details are provided on such NPs.
5.10. Conclusions 109

5.10. Conclusions

It appears that learners of English as a second language have to deal

with the fact that there are no one-to-one form to meaning mappings
in the system of English articles, they have to learn the properties
of nouns related to countability and understand that those proper-
ties are context-dependent, they have to find out how to mark spe-
cific versus nonspecific reference, and they have to learn to take into
account whether or not the context is shared between the speaker
and the hearer. This sounds like a major challenge, but is not yet all,
as learners will face a body of idiomatic uses which may not conform
to these complex and intricate rules. Indeed, it comes as no surprise
that articles are a major source of difficulty for learners of L2 English.
As could be seen from the above review, it has been established
beyond any doubt that articles are more difficult to learn for speakers
of article-less languages, than for speakers of mother tongues which
feature similar independent prenominal morphemes. The bulk of
research on articles in SLA has been done on semantic universals,
and it has yielded mixed results as on how learners become sensitive
to the features of definiteness and specificity.
It is unfortunate that research on language universals has so far
failed to provide a clear picture of the use of articles by learners from
various language backgrounds; the findings are sometimes inconclu-
sive, and at times contradictory. Most importantly, however, interest-
ing as they are, studies on language universals and the acquisition of
articles have limited applicability to language instruction. Genera-
tive grammar consists of fundamental research rather than applied
research, and it aims at modeling human language competence. There
is usually little concern for how those hypotheses may serve peda-
gogical purposes. In fact, one of main tenets of generative grammar
is that under normal circumstances, minimal exposure to human
110 Chapter 5. Articles in SLA research

language is sufficient to trigger normal language development, sug-

gesting limited effects of pedagogical intervention.
Another important fact is that the count/mass distinction and its
current status in a learner’s interlanguage affects the ability to cor-
rectly use articles. Interesting findings have been published concern-
ing the fact that as learners struggle to make sense of the article sys-
tem in English, they generate a number of hypotheses and interim
rules which are often inadequate. The above overview has also shown
that relatively little is known about the use of articles in idiomatic
expressions and fixed phrases.
6.1. Introduction 111

Chapter 6

Articles and ESL teaching

6.1. Introduction

Articles are usually considered unteachable. In the words of Gass and

Selinker: “the English article system… appears to be virtually imper-
meable to instruction” (2008, p. 383). Testimonies to that effect can
easily be found in the literature, such as a teacher’s comment to the
effect that the way his students use articles “bears little or no resem-
blance to established English practice; the students seem to use arti-
cles almost randomly” (Yamada & Matsuura, 1982, p. 50).
There are good theoretical grounds for articles to be difficult for
learners, as was shown in Chapter 4, and much has been written
about the various difficulties that learners face with articles. How-
ever, despite the acquisition of articles by ESL learners being quite
a well-researched area (as could be seen in Chapter 5), relatively little
has been said in terms of specific teaching recommendations. One
problem is the low level of success with direct instruction, which is
most likely due to the huge complexity of metalinguistic rules and
the poor usability of such rules in language production. Articles are
112 Chapter 6. Articles and ESL teaching

commonly believed to be best learnt from context, and therefore large

amounts of exposure to English should facilitate acquisition. This
does not mean, however, that learners exposed to a lot of input do not
have problems with articles. In Ekiert’s study (2004), learners from
an EFL context were compared with those in an ESL context. While
the sequence of the acquisition of articles was similar, the EFL learn-
ers were more accurate in their article choices. This appears to con-
tradict the belief that exposure to natural language is the best way to
learn the “unteachable” articles. However, due to small group size and
differences in learner profiles, those conclusions are only tentative.

6.2. Articles and the efficacy of corrective

feedback

The bulk of research on the teaching of articles comes from studies on

the efficacy of corrective feedback, as articles are often selected as the
target language feature in such studies. Corrective feedback is defined
as the information provided to L2 learners about the ill-formedness
of their production (Stefanou & Révész, 2015). In the studies which
involve articles, the corrective feedback tends to be written. Correc-
tive feedback is usually classified as either indirect (when only an
indication that an error has been made is provided) or direct (when
the correct form is provided as well). Additionally, corrective feed-
back may be accompanied by metalinguistic explanations, that is, the
rules behind the use of the selected language feature.
Very interesting pedagogical considerations arise from a study by
Sheen (2007a). The aim of that study was to compare the effects of dif-
ferent types of corrective feedback on the learning of articles, and to
see if the effects would be mediated by learner aptitude and attitude.
One type of treatment involved the provision of feedback in the form
6.2. Articles and the efficacy of corrective feedback 113

of teacher’s recasts of the learners’ article errors, while the other type
was metalinguistic explanation (accompanied by the corrected form).
Sheen found that the latter type of feedback was clearly far superior
to the former: learners who received the “metalinguistic” feedback
showed gains on both the immediate post-test and the delayed post-
test, whereas the learners who received feedback in the form of recasts
performed on the same level as the control group.
In another, similar study by the same author (Sheen, 2007b), the
two types of feedback were investigated when used in dealing with
errors in texts written by learners of mixed L1 backgrounds (includ-
ing Polish). The target feature was article use for first mention and
anaphoric reference. As in the case of the oral task study, direct cor-
rective feedback which incorporated metalinguistic explanations and
provided the correct form proved to be the superior form of feedback,
resulting in significant gains both immediately after treatment and
on a delayed post-test. However, while the recasts used in the oral
task study did not lead to any gains, in the study of written feedback,
corrective feedback without metalinguistic explanations produced
some positive results. They were inferior to the metalinguistic expla-
nations feedback, but still superior to the control group. The study
thus shows that the learners clearly benefited more from the provi-
sion of feedback than when given no feedback. Also, the superiority
of metalinguistic explanations was mostly observed on the delayed
post-test. This finding is explained by the author by referring to the
distinction between two levels of awareness: noticing and under-
standing. While noticing involves giving attention to specific lan-
guage forms, understanding involves the awareness of the rule that
underlies the choice of those language forms. While understanding
entails noticing, the reverse is not necessarily true. Noticing is a cru-
cial step towards acquisition, but understanding leads to greater and
deeper learning, and facilitates future learning, hence the gains on
the delayed post-test.
114 Chapter 6. Articles and ESL teaching

Another interesting study investigating direct corrective feedback

and metalinguistic explanation was carried out by Shintani and Ellis
(2013). The authors found that only the latter type of feedback had
any effect on the learner’s use of the English indefinite article. They
examined the learners’ indefinite article use in a free writing task
(which, they argue, is more likely to reflect implicit knowledge) and an
error correction task (presented as a measurement of explicit knowl-
edge). Direct corrective feedback had no impact on the learners’ use
of articles: learners who had their articles corrected did not improve
as a result. On the contrary, learners who had the rules behind arti-
cle use explained to them benefited from the treatment insofar as the
error correction task after treatment and the first texts they wrote
after treatment showed improvement; however, in subsequent writing
tasks this effect wore off. Shintani and Ellis interpret this finding as
meaning that neither type of treatment benefited implicit knowledge,
because one did not yield any results, and the other only registered
briefly, so it only temporarily affected explicit knowledge.
Another study by the same authors (Shintani, Ellis, & Suzuki,
2014) yielded even more pessimistic results as far as the teaching of
articles is concerned. In that study, adult pre-intermediate Japanese
learners were provided with direct corrective feedback and metalin-
guistic explanations with respect to two target features: the indefinite
article (for first-mention contexts) and the hypothetical conditional.
Interestingly, the study led to gains in the use of the hypothetical con-
ditional, but not to an improved command of the indefinite article.
In other words, in that study, the (relatively complex) grammatical
structure of the hypothetical conditional proved to be more “teach-
able” than the use of the indefinite article. The authors explain this
by saying that the learners are more likely to attend to features that
contribute more to the global meaning of texts.
The effectiveness of feedback with respect to articles was also inves-
tigated in a study with Polish participants. Zabor and Rychlewska
6.2. Articles and the efficacy of corrective feedback 115

(2015) looked at written error correction and found that all types of
corrective feedback under investigation were superior to the results
obtained for the control group, which confirmed that corrective feed-
back does bring some improvements to the learners’ use of articles.
What is particularly interesting in their study is that the feedback was
useful especially when it included metalinguistic information and
when it was combined with the inductive type of formal instruction.
However, there are also studies which show that corrective feed-
back results in improvement in article use, but the inclusion of meta-
linguistic knowledge does not make a difference. Research involving
advanced learners (Bitchener & Knoch, 2010) confirmed that even at
an advanced level, learners can make further improvement in their
use of articles as the result of corrective feedback, with or without
metalinguistic explanation. In this study, it was again the use of indef-
inite articles with first mention nouns that was the subject of inves-
tigation. A study that looked at the use of articles for specific and
generic plural reference (Stefanou & Révész, 2015) also found that the
provision of feedback is superior to providing no feedback, but the
inclusion of metalinguistic information does not make a difference.
Summing up, existing studies on the efficacy of corrective feed-
back bring mixed results. In some studies (Shintani et al., 2014) feed-
back did not result in improvement, or resulted only in temporary
gains (Shintani & Ellis, 2013). In other studies, feedback resulted in
improved article use (Bitchener & Knoch, 2010; Sheen, 2007b, 2007a;
Stefanou & Révész, 2015; Zabor & Rychlewska, 2015) but the results
were mixed with respect to the inclusion of metalinguistic explana-
tions, which were found to enhance learning only in some of these
studies.
Even though the findings are mixed, they seem to give a gen-
eral picture of the superiority of written corrective feedback over no
feedback, and of direct over indirect feedback. This seems to con-
tradict the remarks made earlier in this book about the seemingly
116 Chapter 6. Articles and ESL teaching

unteachable nature of the English article system. If articles are as

resistant to teaching as common wisdom holds, one would expect
their use not to improve much as the result of corrective feedback,
and one would expect metalinguistic explanations to be of little use,
given the high level of complexity of the rules needed to explain arti-
cle use. This apparent contradiction warrants an explanation.
It should be noted that the above studies investigated article use,
but they selected the specific types of article use which are the easiest
ones to teach, that is, ones involving a correspondence between the
language pattern and the metalinguistic proposition (to use the termi-
nology introduced in Chapter 4). For example, Sheen’s results (2007a,
2007b) leave no doubt that learners benefited from corrective feedback,
especially with metalinguistic explanations, but it should be noted that
only two specific uses of articles were investigated: the indefinite article
for first-mention referents, and the definite article for anaphoric men-
tion, with the two presented in combination. This is the kind of article
use which is often selected for research studies, of the type “I saw a man.
The man was tall.” This particular use of the indefinite and definite
article, especially in combination, is easy to explain. Potentially end-
less new combinations of words can be generated using this principle,
which makes this pattern more teachable than other uses of articles. In
fact, this particular use of articles may be the most unproblematic of all.
Because of this consideration, the studies on corrective feedback
cannot provide conclusive information on the teachability of articles.
This is partly because such studies are not designed to investigate the
learning of English articles as such. Their main aim tends to be inves-
tigating what types of corrective feedback lead to learning gains, and
articles are chosen simply as a convenient sample target language
feature, with only a specific type of article use selected for inves-
tigation (for example, the use of the indefinite article in first-men-
tion contexts). To find out to what extent the use of articles actually
improves as the result of corrective feedback, and to what extent their
6.3. Other treatments and recommendations concerning the teaching of articles 117

use improves after treatment with metalinguistic information, we

need a study which would measure the improvement over time in the
learners’ use of articles in all types of contexts, not just in the more
“rule-based” contexts such as first-mention nouns.

6.3. O
 ther treatments and recommendations
concerning the teaching of articles

Apart from studies which investigate the efficacy of corrective feed-

back, there are some other accounts of successful pedagogical inter-
ventions in the area of articles. Miller (2005) describes an improvement
in her ESL learners at an American university, after a pedagogical
treatment which involved discovery techniques combined with an
explanation regarding the role of countability and definiteness.
Gillian (2017) designed an e-learning tool for the practice of Eng-
lish articles, specifically designed to help Polish learners aged 11–14,
which has been shown to give superior teaching results compared
to traditional instruction. This is a very innovative approach in that
it utilizes the potential of an e-learning platform to practice articles
specifically, and it adopts a “game based learning” approach – defined
by the author as “a learning approach where learning and interactive
entertainment are combined to create a fun and engaging experience
for students in learning contexts designed by teachers” (Gillian, 2017,
p. 82). At the same time, the tool is traditional in the sense that it is
organized around the typical pedagogical rules that are usually used
for the teaching of articles.
Some other publications make specific recommendations concern-
ing the teaching of articles, though this is generally rare. White (2009)
recommends that article instruction should involve determination of
definiteness before determination of count status. Because (misguided)
118 Chapter 6. Articles and ESL teaching

consideration of countability often interferes with learners’ considera-

tion of the semantic context of a NP, White argues that:

[L]earners should be encouraged to contemplate definiteness before

countability. A definite context (with common nouns) will yield the
definite article irrespective of the count status of the NP. Thus, teach-
ers might guide their students to consider countability only after
indefinite contexts have been determined. To help learners more
effectively determine the definiteness or indefiniteness of a NP, teach-
ers may want to emphasize the discourse context, especially the per-
spectives of the interlocutors (i.e., the speaker and the hearer). Learn-
ers can be guided to consider the speaker’s presumptions about the
hearer’s knowledge.
(B. White, 2009, p. 28)

White also recommends that learners should be trained to con-

sider discourse context to determine definiteness, and finally, that
they should be made aware of the need for a flexible approach to noun
countability, by considering the different conceptualizations of count-
ability. White recommends drawing on the work of Wierzbicka (1988)
and the concept of individuation as presented by Yule (1998). A zero
article before a noun means that its referent lacks clear boundaries
and thus resists individuation. An indefinite article before a noun
suggests that its referent possesses boundaries and can be construed
as an individual entity. Appropriate activities would involve making
the learners consider the same noun across different contexts.
The majority of Polish scholars who have devoted attention to the
teaching and learning of articles seem to share the belief in the poten-
tial of explicit instruction (Gozdawa-Gołębiowski, 2003; Paradowski,
2006, 2008; Scheffler, 2007). Such instruction should focus on making
the learners aware of crosslinguistic differences and similarities with
respect to the structural properties of language.
6.3. Other treatments and recommendations concerning the teaching of articles 119

Paradowski recommends the “language interface” method intro-

duced by Gozdawa-Gołębiowski (2003), in which the teaching of new
L2 structures is preceded by awareness-raising explications concern-
ing the working of equivalent or comparable mechanisms in the L1. As
Paradowski explains,

language-awareness tasks sensitize the learner to language phenom-

ena which are present in both his/her L1 and the TL [target language],
but whose overt realization in the two languages may differ. Learners
discover whether the L1 rules are operative in the L2 and vice versa.
The teacher’s task is to demonstrate to the learners through compar-
ative analysis that they already know something which they have so
far regarded as mysterious. This eases the burden and is greatly facil-
itative in lowering the affective filter.
(Para­dowski, 2008, pp. 232–233)

Moreover, the new structures or forms may be practiced first in the L1,
and only then in the L2. In the case of articles, the relevant L1 feature
would be the sequencing of information in Polish sentences.
In a similar vein, Scheffler (2007) recommends translation activi-
ties as a way of sensitizing learners to contrastive aspects. In his opin-
ion, translation activities are useful, because they enable the teacher to
point out to the students that some of the meanings conveyed by the
English articles are rendered in Polish by other means, for example
by means of the sequence of elements in a sentence. To illustrate this,
Scheffler (2007, pp. 61–62) quotes examples given by Szwedek (1976,
p. 57). In the following two sets of sentences, only the first one can be
understood as pertaining to one and the same woman:

Widziałem w oknie kobietę. Po chwili kobieta wyszła z domu.

[I saw a woman in the window. After a moment, the woman came
out of the house.]
120 Chapter 6. Articles and ESL teaching

Widziałem w oknie kobietę. Po chwili z domu wyszła kobieta.

[I saw a woman in the window. After a moment, a woman came out
of the house.]

Presenting such sentences to learners of English should make it possi-

ble for them to realize that they are already familiar with the concepts
of new versus given information. Unfortunately, not all uses and mean-
ings of the English articles (not even a majority) yield themselves to this
type of instruction. Actually, it is mostly the indefiniteness of first-men-
tion nouns that compares neatly with the information flow structure in
Polish. Other features of Polish which share some meanings and uses
with the English articles, such as the Polish definite (demonstrative) and
indefinite pronouns (ten, ta, to, jakiś, jakaś, jakieś, etc.), as well as the
numeral jeden („one”), can be only presented as equivalents of the arti-
cles in certain specific cases, and could result in confusion if the learn-
ers tried to rely on the assumed equivalence in a systematic manner.
A critical view of the idea to use contrastive data to teach the English
articles to Polish learners has been expressed by Król-Markefka (2008),
who conducted an exploratory empirical study which showed that Pol-
ish learners of English are not very sensitive to the feature of definite-
ness as encoded in the word order in their L1. The author concluded that:

… tracing the ways in which [definiteness] may be encoded in both

languages would place a double cognitive burden on the learner’s
mental processing. … making students aware of the devices which
they automatically use to express (in)definiteness in Polish is likely
to introduce unnecessary complications and, consequently, hinder
rather than facilitate language acquisition.
(Król-Markefka, 2008, p. 111)

An important trend in the teaching of articles comes from cog-

nitive linguistics. This field of linguistics has generally been a source
6.3. Other treatments and recommendations concerning the teaching of articles 121

of inspiration for language teaching for some time now (see Bielak,
2011; Holme, 2007; Littlemore, 2001; Tyler, 2012).
The work of Epstein (see section 3.3) was applied in a study con-
ducted by Hinenoya and Lyster (2015) with Japanese ESL learners.
They compared the efficacy of two different instructional treatments,
both computer-assisted, devoted to teaching the definite article. While
one treatment focussed on the concept of identifiability, the other was
based on the concept of mental spaces (MS), employing the notion of
accessibility as formulated by Epstein (2002).
The authors analysed a large number of randomly extracted
instances of definite article use, and grouped the non-generic uses
into four types: (1) structural use, i.e., direct structural anaphoras (e.g.
A ball was on the table. The ball suddenly rolled and fell off ); (2) the
“visible in the situation” usage (e.g. Here is the bathroom); (3) the “vis-
ible in the mind” category (e.g. Where is the bathroom?); 4) the inclu-
sive/associative usage (e.g. I’ve just been to a wedding. The bride wore
blue). While in the case of the first two types, the explanations offered
by traditional accounts and the ones put forwards by cognitive lin-
guists almost entirely coincide, the other two types can profit from
a cognitive interpretation, in which accessibility is operationalized as
the activation of schemata. A schema is “a body of knowledge that is
acquired through experiences in life and is stored (to be accessed) in
our mental dictionary” (Hinenoya & Lyster, 2015, p. 400). In 3), it is
unlikely that the speaker has a specific bathroom in mind, as the loca-
tion of the bathroom is unknown to him/her, and it is even possible
that there may be no bathroom or that there may be more than one. In
such cases, according to Hinenoya and Lyster, traditional accounts of
definite article use are not very successful, whereas the MS approach
can be used to activate the appropriate mental space, in this case,
the schema of a house, which typically includes a bathroom. In 4),
the mention of a wedding activates the schema of a wedding, which
involves the presence of a bride.
122 Chapter 6. Articles and ESL teaching

The findings of the study confirmed the hypothesis that the MS

approach is particularly helpful with types 3) and 4) of article use
when it comes to Japanese learners. The authors see the pedagogical
potential of the concept of accessibility, and they say that, “despite
the absence of linguistic equivalents in their first language, (…) such
learners are nonetheless equipped with a cognitive capacity for inter-
preting article usages through mental space configurations” (Hine-
noya & Lyster, 2015, p. 400).
Attempts to apply concepts taken from cognitive linguistics to the
teaching of articles to speakers of L1 Polish have been made by Król-
Markefka (Król-Markefka, 2010; Król, 2005), who found them prom-
ising. Król-Markefka argues that cognitive-based approaches make
it possible to provide students with general, conceptually-motivated
rules for the use of articles.
It seems that the pedagogy of articles has not been affected much
by the recent trends which emphasise the role of phraseology in
language learning (more on this topic will be said in the following
chapter). Only recently, a study by Shin and Kim (2017) explored the
connection between formulaic sequences (more specifically, lexical
bundles, which will be discussed in the next chapter) and the acqui-
sition of articles by learners of L2 English at different proficiency lev-
els. In view of the shortcomings of pedagogical rules on article use,
and the high level of difficulty that articles present to learners, the
authors decided to investigate whether lexical bundles can help L2
learners to improve their article use. The pedagogical tool they devel-
oped resulted in significant improvement in article use for learners of
various language backgrounds and proficiency levels.
6.4. The treatment of articles in teaching materials and pedagogical grammars 123

6.4. T
 he treatment of articles in teaching
materials and pedagogical grammars

The way English articles are presented, explained and practiced

shows some similarity across the wide range of textbooks cur-
rently available. Despite numerous differences, articles in pedagog-
ical materials are presented as essentially rule-governed, in keeping
with the well-established status of articles as a “grammatical” topic
(Holmes, 1988; Hsu, 2008). Exercises on articles are always found
in the “grammar” sections of textbooks. The main rules (differently
named in various textbooks) revolve around countability, definite-
ness and specificity (Ekiert, 2007). For example, the is introduced
as indicating common ground / shared knowledge / referents iden-
tifiable to both interlocutors, etc. The textbooks usually contain the
most basic rules about the indefinite article, concerning the impos-
sibility to use it with a plural (*a tables) and with non-count nouns
(*a water). Another frequent rule found in textbooks concerns direct
anaphora: the indefinite article subsequently becomes definite upon
the second mention (e.g. I saw a player leaving the game. The player
was hurt.). Textbooks also usually provide an explanation for the use
of the zero article, i.e. the possible omission of articles for generic
meanings, as in the sentence English parsley has curly leaves (Par-
rott, 2000, p. 46).
Because of the challenging character of the main rules for arti-
cle use, most textbooks resort to a set of practical rules, often called
“rules of thumb” (Parrott, 2000, p. 48). These rules of thumb seek to
account for a certain number of exceptions to basic rules by group-
ing those exceptions and showing that they follow the same pattern.
A frequent rule of thumb aims at accounting for the fact that the
article is dropped before some nouns that denote institutions (e.g.
school, church, prison, hospital). Although countable, they will be
124 Chapter 6. Articles and ESL teaching

used without a definite article, provided the primary function of those

places is being referred to.
In addition, most pedagogical grammars and textbooks make
a connection between article use and idiomaticity, by acknowledg-
ing that there are some “fixed” or “idiomatic” expressions in which
articles are used differently than the available rules would suggest
(for example, living hand to mouth, all of a sudden, in front vs. in the
back, or game of cat and mouse). This strengthens the impression that
the unpredictable use of articles is limited to a handful of idiomatic
expressions, which constitute exceptions to an otherwise rule-gov-
erned system.
It has been pointed out that the importance assigned to those
rules is often out of touch with descriptive accounts of English
grammar and corpus findings. Generally, the pedagogical treatment
does not correspond well to what corpus analyses reveal about arti-
cle use (Yoo, 2009). For example, textbooks tend to overemphasize
direct anaphoric use of the definite article compared to the actual
frequency of this type of article in corpora (Yoo, 2009). As Hine-
noya and Lyster point out, the usage of the definite article which
they call “inclusive/associative” (of the type “I’ve just been to a wed-
ding. The bride wore blue,” i.e., indirectly anaphorical,) is very fre-
quently used in reality, whereas it is rarely presented in textbooks
(Hinenoya & Lyster, 2015).
An analysis of nine commonly used grammar textbooks with
exercises with respect to articles was conducted by Król-Markefka
(2012), who pointed out considerable differences in how the usage
of articles is presented, and noted a significant amount of confusion
and lack of clarity. According to the author, the rules presented in
the analysed textbooks are inadequate especially in terms of their
predictivity and discriminatory power. The main shortcomings of
the textbook rules are: “1) lack of consistency, which is confusing
and hinders meaningful learning; 2) numerous ‘uses’ and exceptions,
6.4. The treatment of articles in teaching materials and pedagogical grammars 125

which makes the rules less memorable; 3) lack of unity and meaning-
ful justifications, which makes the rules and their functions not so
easy to understand; in consequence, the learners cannot make their
own contextually-dependent decisions when using articles” (Król-
Markefka, 2012, p. 105).
It seems that the large number of rules presented in textbooks,
along with their lack of predictivity, is the main source of difficulty
in applying those rules. A learner, faced with the need to communi-
cate a specific message, can easily think of a number of rules, some-
times contradictory ones, that all seem to apply. I have explained this
in an earlier publication:

The entire point of having rules is that they are generalizable, that
is, productive. A learner of English, knowing the rule for the third
person singular ending of verbs, being familiar with sentences such
as Alice likes flowers, or Chris loves fast cars, and knowing the verb
hate can come up with a correct sentence such as Alice hates Chris by
means of applying that rule. With articles, a similar action is much
more difficult, if not entirely impossible, because very often all of the
rules seem to apply. For example, a learner may say “I need to go to
a bathroom,” because there are several bathrooms in the house he
happens to be in, bathrooms are clearly countable, and he does not
need to use a specific one. He may equally well think it is acceptable
to say, “I need to go to bathroom,” because he intends to make use
of the primary function of the bathroom.
(Leśniewska, 2017, p. 68)
126 Chapter 6. Articles and ESL teaching

6.5. Conclusions

As this chapter has shown, accounts of successful pedagogical inter-

ventions with respect to articles are not numerous, but they do exist.
The teaching of articles has been studied mostly within the frame-
work of studies on the efficacy of corrective feedback. While these
studies tend to show the superiority of feedback over no feedback,
and they usually report better results when metalinguistic explana-
tions are provided, it has to be remembered that such studies focus
on selected types of article use (most commonly, the use of the indefi-
nite article for first mention, or the definite article for anaphoric use),
that is, the type of uses of articles which is are relatively easy to iso-
late and straightforward to teach by means of rule explanation. This
leads to the tentative conclusion that, among all types of article uses,
there are some that may be very resistant to teaching, and are more
likely to be produced correctly mostly through collocational compe-
tence, and some that are graspable for learners at the level of a gen-
eral “grammar” rule.
Studies on the efficacy of corrective feedback also tell us, indi-
rectly, something interesting about the reason for the problematic
character of articles in ESL acquisition. They show that simple recasts
or corrections (without metalinguistic explanations) bring negligible,
if any, improvement in the learners’ performance (as in, most notably,
Sheen, 2007a). In Sheen’s study, this lack of improvement occurred
even though the recasts were focused solely on articles. Even more
striking is the fact that the learners did not even notice that articles
were the focus of the recasts, as evidenced by a questionnaire incor-
porated in the study. This persistent failure to notice article-related
corrections strongly suggests that one of the reasons for their seem-
ing “unteachability” may be that they are largely ignored by learners,
as they are not perceived as salient.
6.5. Conclusions 127

Among the innovative approaches to the teaching of articles are

explorations of the pedagogical potential of concepts taken from cog-
nitive linguistics, as well as methods based on the application of con-
trastive studies to the teaching of articles. However, opinions on the
usefulness of those approaches are mixed, and empirical evidence
remains scarce.
7.1. Introduction 129

Chapter 7

Formulaicity

7.1. Introduction

In our understanding of how language works, the word is a crucial ele-

ment. The lexicon is usually seen as collection of individual words, which
form the building blocks of larger utterances. One important character-
istic of words is that they seem to participate in different kinds of “lexi-
cal partnerships” (Singleton, 2000, p. 47). Because some words are often
found in the company of certain other words, lexical partnerships of
various kinds have long been recognized in language research.
As the ubiquity of such language phenomena is increasingly
noticed, the term “formulaicity” seems to be emerging as the most
popular “umbrella” term used to refer to the quality of language that
makes it impossible to reduce it simply to individual words and syn-
tactic rules, and to refer to the many different kinds of connections
between words in a language (e.g. Meunier, 2012).17

17 Obviously, having a term with that general a meaning renders it useless for
any kind of precise denotation of any specific phenomenon. In fact, it has been
130 Chapter 7. Formulaicity

To provide the broadest possible definition of formulaicity, it is

useful to take as the starting point the traditional “slot and filler”
approach to language. In this view, for each “slot” along the syntag-
matic axis, one word is chosen from a number of options available
along the paradigmatic axis, depending on the intended meaning.
The choices for each slot, however, are restricted to certain classes
of words by syntagmatic restraints which result from the syntactic
rules of the language. In its broadest sense, formulaicity can be taken
to mean all the phenomena which affect how words are combined
together which fall outside the “slot and filler” model. Boers and
Lindstromberg offer a good practical definition of formulaicity: “the
use of word strings that have become conventionalized in a given lan-
guage as attested by native-speaker judgment and/or corpus data”
(2012, p. 83). Many authors use definitions of formulaic expressions
which are based on the assumed holistic processing, such as: “fre-
quent multiword combinations that are stored and retrieved holis-
tically from the mental lexicon at the moment of speech” (Nekras-
ova, 2009, p. 647). However, such definitions are better avoided, since
proving that multiword combinations are stored/retrieved holisti-
cally is in itself problematic, as will be explained in more detail below.

argued that using formulaicity as an umbrella term brings more confusion than
good, as it creates the impression that different researchers are discussing the
same phenomenon, whereas they are in fact talking about very different con-
structs (see Myles & Cordier, 2016). However, once the term has been appro-
priated by researchers working with very different theoretical frameworks and
methodologies, there is no going back. It would be nothing short of a miracle if
a consensus was achieved across academia as to which narrow sense the term
“formulaic” should be reserved for. In this situation, it makes sense to use it
as a much-needed hold-all term. The term will also be used in this study in its
most general meaning. Another popular umbrella term is phraseology, usually
defined as an area of academic inquiry which deals with the formulaic nature
of language (see eg. Granger & Meunier, 2008).
7.1. Introduction 131

It is usually assumed that the rationale for the existence of formu-

laicity in language has to do with advantages it offers in terms of ease
and speed of processing, as “formulaic sequences have been argued
to minimize encoding work for both the speaker and addressee, thus
allowing for the construction of fluent spoken discourse” (Nekrasova,
2009, p. 647). In the words of Conklin and Schmitt (2012, p. 45):

It makes sense that our brains would make use of a relatively abun-
dant resource (long-term memory) to compensate for a relative lack
in another (working memory) by storing frequently occurring for-
mulaic sequences. These could then be easily retrieved and used
without the need to compose them online through word selection
and grammatical sequencing.

The implications of this widely accepted assumption, as well as empir-

ical research supporting it, will be presented and discussed later in
this chapter.
Myles and Cordier (2016) argue that studies on formulaicity suf-
fer from the confusion between what they call “speaker-external”
and “speaker-internal” discussions of formulaicity, which should
be clearly distinguished. Speaker-external formulaicity can be
observed in language. For example, formulaic sequences may be
defined on the basis of their frequency of occurrence in a corpus.
When claims are made about the psycholinguistic reality of for-
mulaic language, the approach is speaker-internal. In the overview
below, I first look at what could be called by Myles and Cordier
the speaker-external approaches to formulaic language, and then
move on to the speaker-internal ones, before discussing the aspect
which is of crucial importance to this study, namely the connection
between the two.
132 Chapter 7. Formulaicity

7.2. Speaker-external formulaicity

As has been said above, words enter into various lexical partnerships
with one another. However, the lack of a universally accepted tax-
onomy, as well as the fact that interest in word combinations comes
from many different theoretical perspectives, result in a substantial
amount of confusion. Among the most frequently discussed phenom-
ena related to the lexical partnerships of words are “idioms,” “col-
locations,” “restricted collocations,” “collostructions,” “binomials,”
“formulas,” “chunks,” “prefabs,” “routines,” “lexical phrases,” “fixed
phrases,” “lexical bundles” and “n-grams.” Except for the last two,
these terms are sometimes used with different meanings by different
authors, and their meaning is not always rigorously defined. All of
these concepts are related to the fact that certain words co-occur, but
there are important differences among them. Perhaps the best-studied
one of the above categories, idioms, are often opaque in meaning and
are non-compositional. For example, the meaning of the expression
on/over the grapevine cannot be deduced from its component parts.
However, this quality is far from being a defining feature of idioms,
as their meaning may be deducible, as in the case of, for example, the
expression a drop in the ocean. To further complicate the situation, the
distinctions involved in classifying such expressions are not binary.
One could argue, for instance, that there are degrees of opacity, and
the phrase a fair-weather friend is more comprehensible than a dif-
ferent kettle of fish.
The term “multi-word units” (MWUs) seems to be emerging as
the most universally accepted general name for idioms, collocations,
binomials, phrasal verbs, etc. (e.g. Wolter & Yamashita, 2018). This
usage will be followed here, even though it should be noted that the
term is misleading, as the name “unit” overemphasises the holistic
nature of such expressions, which is debatable in view of the fact that
7.2. Speaker-external formulaicity 133

most MWUs can be decomposed into parts, and that the co-occur-
rence of words is a matter of degrees.
The wide range of terminology used to refer to MWUs reflects the
fact that interest in lexical co-occurrence comes from various disci-
plines: linguistics, lexicography, corpus linguistics, language process-
ing, discourse analysis, second language acquisition research, and sec-
ond language pedagogy. The terms “collocation,” “lexical bundles,”
“n-grams,” etc. are used in corpus linguistics in reference to the fact
that words co-occur in corpora. “Prefabricated units,” “prefabs” and
“chunks” are discussed in studies of language acquisition and pro-
cessing to indicate that certain word combinations are acquired and/
or processed as wholes; “formulas” or “formulaic language” are terms
which emphasize the fact that many phrases can be used repeatedly
in certain contexts instead of having to be generated anew each time;
the terms “institutionalized phrases,” “fixed expressions” and “rou-
tines” suggest that many phrases can be used to achieve particular
functions in predictable social situations – the emphasis here is on
the conventionality of language.
What is important to note is that the various types of MWUs may
display very different properties. They may be defined on the basis of
intuitive criteria, corpus data, or pragmatic function, or any combina-
tion of those. They may range in opacity from opaque to transparent,
in frequency from very frequent to relatively rare (as is the case with
some idioms), and in the degree of fixedness, from completely “fro-
zen” phrases to ones which allow for a certain degree of manipulation.
A very important concept in the study of formulaic language is
that of collocation. The term does not have a single, widely accepted
definition,18 but – generally speaking – it refers to the habitual co-oc-

18 In keeping with the origin of the word collocation, i.e. the “placing together” of
words (from the Latin collocare: co (together) + locare (to place)). In this sense
the term has been used since the mid-20th century, when it was popularized by
134 Chapter 7. Formulaicity

currence of words (Cruse, 1986, p. 40). Collocations are often defined

as words that co-occur more frequently than their respective frequen-
cies would predict (e.g. Jones & Sinclair, 1974). For that reason, a word
“predicts” the word(s) which typically follow(s). Laufer and Waldman
(2011, pp. 648–649) provide a definition of collocations which corre-
sponds very well to what seems to be the most common understand-
ing of the term in contemporary literature:

We regard collocations as habitually occurring lexical combinations

that are characterized by restricted co-occurrence of elements and
relative transparency of meaning. Restricted co-occurrence distin-
guishes collocations from free combinations in which the individual
words are easily replaceable following the rules of grammar. Relative
semantic transparency of collocations, on the other hand, distin-
guishes them from idioms whose meaning is much less transparent
than that of collocations and is very often opaque because it cannot
be understood from the words that compose them.

This is also the understanding of “collocation” which is of most rele-

vance to second language acquisition and language teaching contexts,
as, from a language acquisition perspective, the most important fea-
ture of collocations is their lack of predictability from a “words-and-
rules” point of view. Apart from knowing the meaning of do and home-
work, a L2 learner of English must also know that homework collocates
with do and not with make. This is why collocations are challenging
for learners in production, and not so much in comprehension.

Firth and his followers. However, there is a wide array of definitions of the term
which are completely different. Firth originally used the term in a broader sense,
referring to all kinds of co-occurrences at various levels of language, for exam-
ple to alliteration. Nation uses the term “collocation” to describe “any generally
accepted grouping of words into phrases or clauses” (Nation, 2001, p. 317).
7.2. Speaker-external formulaicity 135

Collocations can be variously defined and classified with refer-

ence to frequency of co-occurrence, restrictions on commutability,
semantic characteristics, the notion of familiarity (the fact that cer-
tain phrases are recognized by native speakers as familiar), or habit-
ual use. Because of the availability of large corpora and the ease with
which frequency information can be extracted from them, a very
common approach to collocation is a frequency-based one, where
collocations are identified on the basis of automated searches only.
By researching recurrent combinations of words in corpora, it is pos-
sible to easily obtain information on which words typically collo-
cate, either side-by-side or within a certain span of words. A basic
measure of how common a collocation is in a language is its fre-
quency in the corpus. However, one can also measure the exclusivity
of co­-occurrence, that is,

the extent to which the two words appear solely or predominantly in

each other’s company, usually expressed in terms of the relationship
between the number of times when they are seen together as opposed
to the number of times when they are seen separately in the corpus.
(Gablasova, Brezina, & McEnery, 2017, p. 160)

This measure is known as the Mutual Information [MI] score. For

example, torrential is very often found with rain, glaringly with obvious.
Another important term – this used only in corpus-based stud-
ies – is lexical bundle. Introduced by Biber and colleagues (Biber,
2009; Biber et al., 1999; Conrad & Biber, 2004), the term denotes the
most frequently recurring sequences of words in a collection of texts.
Because bundles are identified using corpus analysis software, which
automatically retrieves word sequences which occur most frequently,
they do not necessarily coincide with multiword units defined accord-
ing to other criteria, for example pragmatic or discursive ones, nor
do they tend to be complete structural units (see e.g. Conrad & Biber,
136 Chapter 7. Formulaicity

2004). For example, they may incorporate clausal boundaries, as in

the bundle I know that I, or they may contain incomplete nominal
chunks, for example, as the result of. Lexical bundles are also known
as “n-grams,” with n standing for the number of words in the bun-
dle. The lexical bundle (n-gram) approach identifies strings of adja-
cent words of a given length. In contrast, collocations can be found
in the corpus with a predetermined degree of proximity, as specified
by the window span set for a specific search.
To what extent is language formulaic? Conklin and Schmitt (2012)
discuss seven studies which used very different methods to answer this
question. For example, researchers count the percentage of identical
strings in stretches of talk, or they look at the percentage of text that
is made up of lexical bundles. Of course, the exact amount of formu-
laic language will inevitably depend on the method used to identify
formulaic elements. However, the average from this group of studies
gives a certain rough idea, and Conklin and Schmitt (2012) conclude
that between one third and one half of discourse is formulaic in nature.

7.3. Speaker-internal formulaicity

7.3.1. The idiom principle

Studies on large corpora of texts have proven beyond doubt that cer-
tain word combinations tend to reoccur, but as such they tell us noth-
ing about the reasons why this happens. It is commonly assumed
that this tendency is a reflection of a psycholinguistic phenomenon
which makes it easier for humans to retrieve and process familiar
combinations of words, and that is why they use them more fre-
quently. Already in the 1980s, in a much-cited paper, Pawley and
Syder sought to explain what these authors saw as “two puzzles for
7.3. Speaker-internal formulaicity 137

linguistic theory.” The two puzzles were two intriguing characteristics

of the speech of native language users, namely “native-like selection”
and “native-like fluency.” The first one denotes the ability to produce
phrases which are the natural form of expression, rather than “odd”
collocations. The other characteristic – “fluency”– was described as
puzzling because “human capacities for encoding novel speech in
advance or while speaking appear to be severely limited, yet speakers
commonly produce fluent multi-clause utterances which exceed these
limits” (Pawley & Syder, 1983, p. 191). Comparing recordings of sponta-
neous speech, Pawley and Syder observed that some expressions – e.g.
I don’t need anyone to tell me what to do – were relatively pause-free
and delivered faster than the normal rate for language production.
The authors argued that both native-like selection and fluency pro-
vide the evidence for the fact that “fluent and idiomatic control of
a language rests to a considerable extent on knowledge of a body of
‘sentence stems’ which are ‘institutionalized’ or ‘lexicalized’” (Pawley
& Syder, 1983, p. 191). The fact that chunks can be recalled from mem-
ory does not mean that they are not analysable into segments. The
speakers can always use their ability to construct an infinite number
of new sentences on the basis of a finite set of grammatical rules as
one of the modes which is available to them, but they “do not exer-
cise the creative potential of syntactic rules to anything like their full
extent, and that, indeed, if they did so they would not be accepted as
exhibiting native-like control of the language” (1983, p. 193).
It should be noted here that Pawley and Syder actually combined
two elements in their model: a pragmatic one and a psycholinguistic
one. “Lexicalised sentence stems” play a particular pragmatic role,
they have a specialized function in discourse, and this is connected
to their pragmatic non-compositionality. In the words of Weinert,
“certain language sequences have conventionalized meanings which
are used in certain predictable situations” (Weinert, 1995, p. 196). That
meaning may not be clear from the literal meaning of the expression.
138 Chapter 7. Formulaicity

Many of the examples given by Pawley and Syder refer to the kind of
language choices which one would normally call “choosing the right
expression for the right situation.” The other aspect is psycholinguis-
tic, as they claim that “memorized sentences” and “lexicalised sen-
tence stems” are to some extent processed as wholes, as ready-made
chunks of language.
Another linguist who is often credited with bringing the concept
of those two modes of processing into linguistics is John Sinclair.
Because of his ground-breaking research on large samples of lan-
guage, Sinclair is brought the study of long authentic texts into the
centre of attention in linguistics (Stubbs, 2009). His well-known dis-
tinction between the “open-choice principle” and the “idiom princi-
ple” was based on the idea of dual, simultaneous access to two modes
of processing. On the one hand, any speaker can use language in the
“open-choice” mode, that is, in the “slot and filler” system mentioned
above, with a number of syntagmatic choices available for each slot
along the paradigm, with the only restriction being imposed by syn-
tactic rules. On the other hand, the speaker also has access to “a large
number of semi-preconstructed phrases that constitute single choices,
even though they might appear to be analysable into segments” (Sin-
clair, 1991, p. 110), which makes it possible to produce language accord-
ing to the “idiom principle.” Both of those modes are at the disposal
of any speaker, and they are not necessarily exclusive; to the contrary,
it seems more likely that they are used at the same time, in different
proportions. Sinclair thus changed the emphasis in linguistics from
the previously overestimated role of paradigmatic choice to syntag-
matic constraints in linear sequences (Stubbs, 2009).
A similar distinction was made by Skehan (1998), who wrote about
two modes of processing available to language users. One operates
on the level of grammatical rules, which make it possible to gener-
ate novel utterances by putting individual words together, for exam-
ple when precision of expression or creativity is needed. The second
7.3. Speaker-internal formulaicity 139

mode is based on memorized multi-word items, which can be quickly

retrieved, enabling fluency.
It should be noted that the above approaches propose a certain
duality: there are two modes of processing, they can be used inter-
changeably, and they complement each other. It stands to reason that
both modes of processing co-exist, and speakers have access to both
individual items and entire chunks. One could at this point ask the
same question which was asked at the end of the previous section
with respect to language processing: to what extent is language for-
mulaic? If the two modes co-exist, are they used to the same extent?
The “dual” models seem to imply a fifty-fifty share. However, some
scholars strongly emphasize the role of the formulaic mode of pro-
cessing, as if giving it priority. According to Wray (Wray, 2002a,
2002b), “formulaic processing is the default,” and “construction out
of, and reduction into, smaller units by rule occurs only as necessary”
(Wray, 2002b, p. 119). According to Wray, evidence for this formulaic
nature of language storage and processing can be found in the mul-
tiple irregularities of all kinds which are so commonly observed in
all natural languages:

if we only create and understand utterances by applying rules to

words and morphemes, it is difficult to see why irregularity should
be tolerated, let alone why an item or construction should progress
from regular, to marked, to antiquated, to a fossilized historical relic.
(Wray, 2002b, p. 118)

Indeed, the existence of such irregularities is bound to be connected

with the formulaic character of language; however, they do not prove
that formulaic processing is primary or that it takes some kind of
precedence over the “open-choice” mode.
To answer this question, one would need a way of knowing to what
extent a speaker relies on prefabricated formulas when producing
140 Chapter 7. Formulaicity

language. At the moment there is no answer to this question. Of

course, we could say that formulaic language is produced by recourse
to the idiom principle, in which case the answer to the question would
be the same: between one third and a half. However, maybe in the
future other methods will provide an answer. One promising direc-
tion of research is the exploration of the prosody of formulaic lan-
guage. As was mentioned above, Pawley and Syder in their discus-
sion of “native-like fluency” noted that certain phrases were delivered
faster than the normal speed and were relatively pause free. This aspect
of formulaic language does not appear to be at the centre of attention,
but there is some research evidence showing that prosodic evidence
may provide a verification for the formulaicity of word sequences (Lin,
2012; Lin & Adolphs, 2009). There is also the possibility that eventu-
ally research into brain activity will throw more light on the picture,
as the processing of formulaic and non-formulaic language has been
connected to different parts of the brain.

7.3.2. Theoretical support for the formulaic nature

of language processing

At this point one could ask if the dual models presented above are
based on pure speculation, or if there is empirical evidence to sup-
port them. The criticism made by Myles and Cordier about the con-
fusion of speaker-internal and external aspects of formulaicity has to
do with the fact that many researchers “take as their basis what usu-
ally happens in the language surrounding the speaker, extrapolating
that this preferential status has consequences for the storage of these
sequences in the speakers of that language” (Myles & Cordier, 2016,
p. 4). In other words, it is somehow taken for granted that manifesta-
tions of formulaicity in language production do have an underlying
psycholinguistic reality, as in the “dual” models mentioned above.
7.3. Speaker-internal formulaicity 141

Even though they do have a point in that such a connection is

indeed often assumed or taken for granted without sufficient proof,
there is substantial theoretical and empirical evidence to support
the connection between formulaicity observed in language and the
postulated formulaic nature of language processing.
First of all, the formulaic nature of language processing is sup-
ported by theories of chunking. The process of chunking, a phe-
nomenon described by psychologists (Chase, W. G., & Simon, 1973;
Gobet et al., 2001), is believed to take place at all levels of language
(Nation, 2001, p. 319), enabling the grouping of smaller units into
larger wholes. For example, morphemes are processed as units, not
as sequences of phonemes, and complex words are likewise pro-
cessed in their entirety, rather than as combinations of individ-
ual morphemes. The fact that chunking takes place does not pre-
clude the possibility of the analysis of wholes into parts if the
need arises. Thus the purpose of chunking is to give a processing
advantage, enabling language comprehension and production at
a greater speed (N. C. Ellis, 2003). At the same time, chunks need
to be separately stored, alongside individual items (Nation, 2001,
pp. 320–321). The concept of a “chunk” implies that a multi-word
combination is stored just like an individual word, or, at least, that
the constituent items of the chunk are recalled and produced in
a certain linear order. However, as Weinert (1995, p. 197) suggests,
there is no reason to believe that all kinds of formulaic chunks are
stored in the same way. Indeed, there is no clarity as to what chunk-
ing is, when it comes to language, technically speaking. Wray (2012,
pp. 233–234) asks the important question

… of whether a processing advantage in terms of speed indicates

holistic storage or simply the faster mapping of components – and
indeed whether these are really two different things or just differ-
ent ways of conceptualizing the same thing. The notion of holistic
142 Chapter 7. Formulaicity

storage could be viewed as a device for talking about linguistic units

that are ring-fenced for speedy processing – at the level of articula-
tion, this must presumably be the case. On the other hand, there is
potentially a different quality to holistic storage and access. First,
it could entail the synchronous access of all components (…). Sec-
ond, it could entail accessing (albeit in sequence) larger base com-
ponents, through a direct mapping from the meaning of the entire
expression to the phonological form (of those parts that are reliably
fixed – as construction models might favour). Third, in the case of
word strings that have become fused through repeated use, it could
entail the creation of a new access pathway that bypasses the origi-
nal componential route – as the extensive neurological research into
automatic processing indicates.

Hoey’s concept of “lexical priming” (2005) is based on the idea that

lexical patterns are responsible for the structure of language: “as
a word is acquired through encounters with it in speech and writ-
ing, it becomes cumulatively loaded with the contexts and co-texts
in which it is encountered, and our knowledge of it includes the fact
that it co-occurs with certain other words in certain kinds of con-
text” (Hoey, 2005, p. 8). It is priming, then, that enables recurrent
co-occurrence of words; syntax, therefore, is merely an outcome of
the pervasiveness of collocation.

7.3.3. Empirical evidence supporting the formulaic

nature of language processing in native speakers

Even though some scholars argue that there is no sufficient evidence

of the psychological reality of the idiom principle (Siyanova-Chan-
turia & Martinez, 2015), there is in fact substantial evidence confirm-
ing the existence of the formulaic mode of language processing.
7.3. Speaker-internal formulaicity 143

Empirical evidence which suggests that multi-word units are stored

as wholes comes mostly, though not exclusively, from research on the
processing of idioms by native speakers. Conklin and Schmitt provide
a review of studies on idioms which provide empirical evidence that
“formulaic language is processed both more quickly and potentially
differently from nonformulaic language” (2012, p. 47). Such studies are
very numerous (too numerous to be listed here) but the general con-
clusion that emerges from this vast body of research is that idioms in
general tend to be processed faster than comparable non-idiomatic
sequences. This phenomenon has been observed both in studies which
compared the processing times for the figurative and the literal inter-
pretations of the same idioms, as well as ones which compared the
processing of idioms and comparable non-idiomatic sequences. For
example, in one of the early such studies, Van Lancker and colleagues
(Van Lancker, Canter, & Terbeek, 1981) had participants read aloud
sentences with phrases which can be read either literally or figuratively,
such as “he didn’t know he was skating on thin ice,” and observed
that the participants inserted longer pauses into the phrase when the
sentence context required a literal interpretation, while running the
words together if the context supported the idiomatic interpretation.
Whether the idiom is decomposable or oblique – for example, pop the
question vs. kick the bucket – does not make a difference, both of these
types are processed faster than their non-idiomatic counterparts, for
example, ask the question and fill the bucket, respectively (Tabossi,
Fanari, & Wolf, 2009). This suggests that formulaicity in itself speeds
up processing. However, as rightly pointed out by Conklin & Schmitt
(2012), studies on idioms are not ideal for drawing conclusions about
formulaic language in general, because idioms are special in many
ways – they are marked, and they are sometimes relatively rare. This
is why of even more interest are studies which examine the processing
of formulaic language by native speakers in which the target expres-
sions are less idiomatic, but simply frequent.
144 Chapter 7. Formulaicity

In a study by Bod from 2000 (quoted in Bod, Hay, & Jannedy,

2003) participants responded faster to three-word subject-verb-object
sequences when they were frequent (for example, I like it) than when
they had a lower frequency (as in the case of, for instance, I keep it).
The sequences were otherwise matched for structure, level of com-
plexity, as well as the lexical frequencies of the individual items.
This was interpreted as suggesting holistic storage of the more fre-
quent ones. A similar study by Arnon and Snider (2010) employed
four-word phrases which are completely compositional and com-
prehensible, such as don’t have to worry. Again, the components of
those sequences were matched for length and frequency; it was only
the frequencies of the entire phrases that were different. The results
confirmed faster processing time for the more frequent sequences.
This was interpreted as meaning that language users notice and store
frequency information with respect not only to individual lexical
items, but also with regard to multiword phrases, even entirely com-
positional ones. Another interesting study (Tremblay, Derwing, Lib-
ben, & Westbury, 2011) compared the speed with which the partici-
pants read two versions of a passage: one containing lexical bundles
(such as don’t worry about it), and the other comparable non-bundle
strings. The results showed that the passages with the bundles were
read faster, and they were also better remembered and recalled more
correctly than the passages which contained ordinary word combi-
nations. Those findings were corroborated by a study using a dif-
ferent methodology, namely phrase recall and electrophysiological
(ERP) measures (Tremblay & Baayen, 2010). They focused on phrases
such as in the middle of, and found that their frequencies correlated
with the participants’ recall. Sosa and MacFarlane (2002) tested reac-
tion times to phrases with different frequencies, all containing the
word of. The authors observed a clear correlation between the fre-
quencies and the reaction times, which they interpreted as indicat-
ing holistic storage.
7.3. Speaker-internal formulaicity 145

There are also studies in which lexical bundles are the specific type
of formulaic language under investigation (Nekrasova, 2009; Schmitt,
Grandage, & Adolphs, 2004). Schmitt and his colleagues used a dicta-
tion task in which the participants are presented with longer stretches
of dictated language than normally one’s short-term memory can
hold. Some of the dictated stretches had formulaic items embedded
in them, some did not. The idea was that the participants would have
to reconstruct the strings, so they would recall them imperfectly, but
the embedded formulaic phrases were expected to be rendered more
accurately than the rest of the dictated text, possibly without pauses
and hesitations, which would be interpreted as evidence of their holis-
tic storage. This hypothesis was only partly supported, with some for-
mulaic sequences being recalled intact by some speakers. This raises
the possibility that either the frequency of occurrence in language is
not directly related to which phrases are stored holistically, or there
may be variability from speaker to speaker with respect to which
phrases are stored as chunks. However, as Nekrasova (2009) pointed
out, the formulaic phrases embedded in the study by Schmitt and
colleagues were of many different types. Some did and some did not
have a pragmatic function, some were more typical of conversation,
some were more academic, and they varied greatly in terms of fre-
quency, which may explain the inconclusive results.
A very important fact which should be noted at this stage is that
all the studies mentioned above, except for the ones involving idioms,
used the frequencies of the entire phrases as the measure of formulai-
city. However, it has been suggested that formulaicity may be related
not only to the raw frequencies of strings, but also to measures of
the exclusivity of the co-occurrence of the constituent elements of
a string. As was mentioned above, one such measure is the MI (mutual
information) score. Indeed, it seems reasonable that MI may be related
to the psycholinguistic status of some phrases. As Gablasova, Brezina
and McEnery (2017, p. 160) explain:
146 Chapter 7. Formulaicity

Exclusivity is likely to be strongly linked to predictability of co-oc-

currence, when the appearance of one part of the collocation brings
to mind the other part. For example, collocations such as zig zag,
okey dokey, and annus mirrabilis are fairly exclusively associated. We
could hypothesize that words that are likely to be seen in each oth-
er’s company may be more easily recognized, acquired, and stored
as a unit.

Another source of support for the existence of formulaic processing

of language comes from studies involving patients with brain dam-
age (Van Lancker-Sidtis & Postman, 2006; Van Lancker & Kempler,
1987; Van Lancker Sidtis, 2012). It appears that the representation of
familiar phrases (idioms, collocations, routines, etc.) in the brain may
be different from that of non-formulaic language, and that the two
types of language may differ with respect to lateralization: the results
suggest that novel language is processed in the left hemisphere, while
formulaic language in the right one.
Another line of inquiry which offers support for the co-existence
of formulaic and non-formulaic processing of language can be found
in studies on the prosody of formulaic language. Lin (2012) argues
for the potential of prosodic features to identify formulaic phrases,
and mentions a number of studies which point to the association of
formulaic phrases with certain prosodic features, specific stress pat-
terns, specific alignment of pauses, and the faster speed and rhythm
of speech (Kuiper, 2004). Further support for this position is provided
by Hallin and Van Lancker Sidtis (2017), who compared the prosodic
characteristics of naturally spoken Swedish sentences with proverbs
and matched control sentences. The sentences with proverbs were
spoken differently, both in terms of speed and tonal pattern. The dis-
tinctive prosodic characteristics of the formulaic parts of sentences
were interpreted by the authors as suggesting holistic storage for for-
mulaic sequences, supporting in general the dual process of language
7.4. Formulaicity, frequency and recent trends in linguistics 147

competence. The authors also provide a tentative suggestion concern-

ing the neurolinguistic reality behind formulaic language, suggest-
ing that formulaic expressions are stored and processed as complex
motor gestures, modulated by subcortical motor systems, in which
a key role is played by the basal ganglia (Hallin & Van Lancker Sidtis,
2017). As they say, “the dual-process model proposes that formulaic
expressions are produced using holistically stored motoric gestures
modulated by a right hemisphere-subcortical system. A holistically
stored motoric gesture can be expected to be produced with a faster
rate than a composed one” (Hallin & Van Lancker Sidtis, 2017, p. 85).
This would explain why patients with damage to the basal ganglia
suffer an impoverishment of formulaic language.
All in all, the available research seems to strongly support the view
that there exists a formulaic mode for the processing of language,
and, very importantly, for formulaic language other than idioms, the
researchers have established a connection between the frequency of
phrases and the speed or ease of processing. In other words, the more
frequent a phrase is, the more likely it is to be processed as a chunk.
At the same time, processing advantages are also observed for for-
mulaic language which is not necessarily frequent, but fixed (such as
idioms), which means that the token frequency of certain phrases is
not the only predictor of formulaic processing.

7.4. Formulaicity, frequency and recent

trends in linguistics

The research reviewed above – especially the fact that frequency

effects are observed for decomposable, transparent but frequent
sequences – ties in with a number of linguistic theories which blur the
distinction between grammar and lexis. As far as English linguistics
148 Chapter 7. Formulaicity

is concerned, the concepts of lexis and grammar being interconnected

can be seen as going back to Halliday (1961). Later on, corpus linguis-
tics provided extensive support for the recognition of the ubiquitous
nature of formulaicity in language. More recently, the recognition of
the role of frequency has chimed very well with usage-based models
of language (Bybee, 2006, 2007; Bybee & Hopper, 2001; Goldberg,
2006; Langacker, 1987, 2000; MacWhinney, 2006; Tomasello, 2003).
Those models emphasise the data-driven nature of learning, see-
ing it as a process driven by experience. In section 4.3, it was said that
learning can be seen as an unconscious, automatic, frequency-based
process of induction (N. C. Ellis, 2005). As the result of coming in
contact with samples of language, certain form-meaning mappings
become gradually automatic for a given speaker. Related terms used to
refer to this process are entrenchment (Langacker, 1987) and automa-
ticity (Segalowitz, 2003). In usage-based models, learning is inductive.
Since exposure to input is the reason why people acquire language,
it naturally follows that the more frequently a certain structure is
encountered, the more likely it is to become entrenched. Entrench-
ment is thus related to the frequency of occurrence of specific items
in the language.
In the school of linguistic thought known as construction gram-
mar (Goldberg, 2006; Tomasello, 2003), the language learner acquires
a set of constructions as a result of repeated experiences. These con-
structions may be of different sizes as well as levels of abstractness.
Many different factors may interplay with the acquisition of con-
structions; however, the most important variable at play is frequency
(see e.g. Gries & Ellis, 2015). Depending on the frequency with which
a particular linguistic construction is encountered, it becomes rep-
resented in the lexicon with weaker or stronger connections. Again,
the language appears to be an accumulation of experiences, which
changes every time (even if only to a minute extent) every time a new
sample of language is encountered.
7.4. Formulaicity, frequency and recent trends in linguistics 149

Those models tie in with the theory that the primary language
learning mechanism is statistical learning (which was already briefly
mentioned in section 4.3). This view holds that humans are sensi-
tive to frequencies of co-occurrence, and, without being consciously
aware of it, they accumulate this frequency information until pat-
terns of grammatical structure emerge (Rebuschat & Williams, 2012;
Rebuschat et al., 2012). Learning is thus seen as “a gradual process
of accumulating linguistic knowledge based on the distributional
properties of the input” (Andringa & Rebuschat, 2015, p. 188). It
has to be noted, though, that the discovery of patterns in the input,
such as the transitional probability between adjacent elements, has
been investigated mostly in infants, in L1 contexts (see Romberg
& Saffran, 2010), and the exact nature of statistical learning remains
unknown, with one likely possibility being chunk formation (Ham-
rick, 2014).
Those theories are compatible with connectionist approaches
(Rumelhart & McClelland, 1986) to language processing (for an illu-
minating overview and discussion, see Singleton, 1999). The connec-
tionist paradigm “sees knowledge in terms of soft connection strength
rather than rules or patterns” (Singleton, 1999, p. 123). The concept
of spreading activation and parallel processing are compatible with
views which emphasize statistical properties of the input in language
learning (Chater & Manning, 2006).
A related line of inquiry is the application of probabilistic models
to language acquisition and processing. Chater and Manning (2006)
provide an in-depth discussion of those models and argue that, with
recent advances in the field, probabilistic models can now account for
the learning and processing of language, and that learning a language
involves learning probabilistic models (Chater & Manning, 2006).
Those models can thus provide a theoretical base for explaining how
learners acquire language through sensitivity to the frequencies of
patterns in the input.
150 Chapter 7. Formulaicity

7.5. Formulaic language in L2 speakers

All of the studies mentioned in the above section involved native

speakers, and most of them involved only native speakers. This sec-
tion looks at the available information concerning formulaic language
in L2 learners. For the sake of clarity, it will first look at “speaker-ex-
ternal” data, and will then look at attempts to investigate the “speaker­-
internal” aspects.

7.5.1. Speaker-external aspects

The best-known fact about formulaic language and L2 learn-

ers is that it usually appears to be insufficiently developed. There
seems to be a consensus as to the fact that “formulaicity (…) is an
area where second language (L2) learners only very slowly close
the gap on native speakers” (Boers & Lindstromberg, 2012, p. 83).
In a study that investigated the use of one type of formulaic lan-
guage – namely collocations – by L2 learners, Laufer and Waldman
(2011) found that L2 learners at all proficiency levels produce fewer
collocations than native speakers, errors in collocation use persist
even at high level of advancement, and it is only at the relatively
advanced stages of L2 learning that the number of collocations
increases in learners’ language. Collocations are well documented
as a problem area even for advanced learners: “even advanced learn-
ers have considerable difficulties in the production of collocations”
(Nesselhauf, 2003, p. 237); “Thus, for advanced students colloca-
tions present a major problem in the production of correct Eng-
lish” (Bahns & Eldaw, 1993, p. 101).
A number of studies provide the finding that the use of formulaic
language by language learners is different from that of native speakers
7.5. Formulaic language in L2 speakers 151

(e.g. Granger, 1998). L2 learners are known to be unaware of some

typical phrases, and tend to put words together on the basis of their
individual meaning, which accounts for the “foreign-soundingness”
of L2 learners’ language. Learners’ collocational choices are also likely
to be affected by their L1 (Leśniewska, 2003; Singleton, Leśniewska,
& Witalisz, 2007; Yamashita & Jiang, 2010).
The use of collocations in a second language also tends to be very
limited, and influenced by frequency in the language. To that effect,
a study of collocations in L2 learners (Durrant & Schmitt, 2009)
showed that L2 learners indeed use collocations, but only certain
types; namely, they rely heavily on high-frequency collocations, while
underusing certain collocations that would better express the mean-
ing intended, but that are less frequent in English.

7.5.2. Speaker-internal aspects

Two very different possibilities are entertained in the literature as far

as the use of formulaic language by L2 learners is concerned. One
view is that L2 learners (as opposed to child L1 learners) may be taking
a non-formulaic approach to learning a language (Wray, 2002a). In
this view, learners are much more likely to focus on individual words.

The adult language learner, on encountering major catastrophe,

would break it down into a word meaning “big” and a word mean-
ing “disaster” and store the words separately, without any informa-
tion about the fact that they went together. When the need arose
in the future to express the idea again, they would have no mem-
ory of major catastrophe as the pairing originally encountered, and
any pairing of words with the right meaning would seem equally
plausible.
(Wray, 2002a, p. 209)
152 Chapter 7. Formulaicity

This stands to reason, and is supported both by research (such as that

mentioned in the above section) and plenty of anecdotal evidence.
Differences in the use of formulaic expressions between L1 and L2
speakers are well documented, and language learners often sound
unidiomatic. In this view, efforts to learn idiomatic expressions made
at higher level of advancement are just the icing on the cake, rather
than being a basis of all learning, as is the case with L1 learners, espe-
cially as seen by usage-based models.
Another view is that the mechanisms central to the usage-based
models of learning, statistical, probability-driven learning, deter-
mined by the frequency of forms in the input, may also be applicable
to L2 acquisition (N. C. Ellis, 2003). Some authors like to view those
two possibilities as contradictory and mutually incompatible (see e.g.
Durrant & Schmitt, 2009). For example, Conklin and Schmitt ask:

If native speakers are able to decrease demands on cognitive capacity

because formulaic sequences are, in a sense, ready to go, are non-na-
tive speakers able to do the same? This is an important issue, as some
evidence seems to show that second language (L2) learners neglect
phrases, focusing instead on individual words.
(Conklin & Schmitt, 2012, pp. 45–46)

The authors thus address the potential use or non-use of formulaic

sequences by non-native speakers as a dichotomous issue, which may
not be entirely justified. Cognitive behaviour usually shows high var-
iation across individuals, and within the same individual across lev-
els of L2 proficiency. Consequently, the amount of reliance on for-
mulaicity may vary greatly from learner to learner, and within each
individual as well.
Two main reasons could explain why the research on the process-
ing of formulaic language by L2 learners is not likely to make signif-
icant advances: (1) various processes in L2 learners may well coexist;
7.5. Formulaic language in L2 speakers 153

and (2) it makes little sense to consider language learners as a group

with shared characteristics, because of the wealth of different fac-
tors involved, among which one of the most important ones is the
level of proficiency. This would explain why there is both evidence
pointing to the non-formulaic nature of learners’ production, as was
briefly presented above, and why there is quite a lot of empirical sup-
port for the fact that formulaicity does play a role in L2 use, as will
be shown below.

Research findings supporting formulaic processing in L2

While it is known that language users are sensitive to frequency effects

in their L1, it is only recently that this issue has been investigated with
respect to L2 acquisition and use. The most common means of investi-
gating frequency effects is by testing whether response times in online
lexical decision tasks vary depending on the frequency of the stimulus
word in the language. Durrant and Schmitt found that adult learners
do retain information about the co-occurrence of specific words in
the input (Durrant & Schmitt, 2010), which they interpreted as under-
mining the belief that the L2 use is “non-formulaic.”
A study by Siyanova-Chanturia, Conklin, and van Heuven (2011)
employed eye-tracking methodology to investigate processing by
native and non-native English speakers of sentences in­cluding for-
mulaic sequences with different frequencies, more specifically, three-
word binomials such as bride and groom and their inverted versions
(groom and bride), which provided excellent controls because they
were identical in all respects (syntax, meaning, frequency of constit-
uent items) except the frequency of the entire binomial. The more
frequent versions of the binomials were read faster, both by native
and non-native speakers. The impact of input frequency on process-
ing time was also demonstrated by Jiang and Nekrasova (2007) in
154 Chapter 7. Formulaicity

a study involving grammaticality judgements made by L2 speak-

ers. The researchers observed that those judgements were faster and
more accurate than for control strings which were not formulaic.
Likewise, Wolter and Gyllstad (2013) found that advanced learners
of English show sensitivity to frequency effects with respect to col-
locations.

Level of proficiency

Interestingly, it seems that proficiency level of the learners plays

a major role in their sensitivity to frequency effects, which is perfectly
in keeping with commonsensical expectations, since sensitivity to fre-
quency effects comes from exposure to language, and that exposure
increases as the learners make progress. In the above-mentioned study
by Siyanova-Chanturia, Conklin, and van Heuven (2011) observed
sensitivity to frequency effects in both native speakers and advanced
non-native speakers, but not for intermediate learners. Wolter and
Yamashita (2018) found that in both native speakers, as well as lower
level and higher level learners of English, the speed of processing of
collocations depends on both the frequency of individual components
and the frequencies of the whole collocation, but the proportions to
which those factors influence the response times varies depending on
the proficiency level. The authors observed “a shift away from reliance
on word-level frequency to collocational frequency that occurred with
gains in proficiency” (Wolter & Yamashita, 2018, p. 441). The two stud-
ies mentioned earlier in and Schmitt and colleagues (Schmitt et al.,
2004), apart from native speakers also included non-native speakers,
and found that non-natives become better at accurately producing
lexical bundles as their proficiency increases.
7.5. Formulaic language in L2 speakers 155

Formulas as an aid

Formulaic language may be seen as a learning strategy in L2 acquisi-

tion: learners may memorise chunks, use them as wholes, and even-
tually derive rules for productive use. Learners may also use chunks
which enable them to function in a specific situation, in which case
the chunks play a specific pragmatic role and enable communication.
There is a certain similarity in the role of chunks in L1 and L2 acqui-
sition: in both cases the chunks may contain language which is more
syntactically complex than the current ability of the learner allows
for in free production, or it may contain vocabulary items which the
learner is unable to use correctly outside of the chunk. Apart from
this similarity, however, making comparisons between L1 and L2 use
of chunks is risky, nor is it justified to draw any conclusions about
the role of chunks in L2 acquisition on the basis of L1 acquisition, as
the two processes may be essentially different, due to differences in
cognitive processing between infants acquiring the L1 and L2 learn-
ers, who may be of any age.
Learners “always feel pressure to produce more than they can”
(Nattinger & DeCarrico, 1992, p. 27), and prefabricated patterns ena-
ble them “to express functions which they are yet unable to construct
from their linguistic system” (Hakuta, 1976, p. 333). The role of for-
mulaic language in language production is discussed, among others,
by Raupach (1984), who shows that advanced L2 learners make use
of formulas in their speech, mostly as fluency devices and discourse
organizers. Raupach also demonstrates that L2 learners use their own
formulas, which may or may not correspond to target language for-
mulas. Some target language formulas are overused, and some are
underused by learners. Towell, Hawkins and Bazergui (1996) carried
out a longitudinal study comparing learners of French before and after
they stayed in a French-speaking country, and attributed the resulting
increase in the fluency to the proceduralisation of knowledge, that is,
156 Chapter 7. Formulaicity

to the fact that the learners stored memorized sequences of language.

The observation was made on the basis of counting the mean length
of run (number of successive syllables unbroken by a pause) and other
characteristics of the learners’ speech. Dechert (1983) examined the
language production of a German learner of English, and that learner
proved more fluent when formulaic language was used.
This kind of facilitating effect of collocational knowledge in
learners (making it easier to use articles, prepositions, etc.) has also
been treated as a compensatory strategy, even though this view
is not common. An example is provided by a study by Mueller
(2011), who tested advanced adult English learners with different
L1s on their ability to fill in a gapped test in which prepositions had
been removed. The study is carried out on the basis of the assump-
tion that “collocational knowledge may serve a compensatory role”
(Mueller, 2011, p. 480). The author observes that “second language
learners’ interlanguage relies on collocational knowledge in lieu
of precise semantic knowledge” (p. 480). This position sees an ele-
ment of the target language competence (in this particular case,
the command of prepositions) as being produced by native speak-
ers thanks to “precise semantic knowledge.” Learners, who lack
that knowledge, may rely on the compensatory strategy of relying
on remembered collocations. This view does not seem particularly
convincing: the whole power of collocational associations between
words is seen as merely an intermediate step on the way to full mas-
tery of language.

Do chunks feed into rules?

There exists an important difference between L1 and L2 learners with

respect to acquisition. Chunks are known to play an important role in
child L1 acquisition, being acquired as wholes and only later broken
7.5. Formulaic language in L2 speakers 157

up into constituent parts (Brown, 1973; Peters, 1983; Wong-Fillmore,

1976).
As to whether or not this happens in L2 learning, opinions and
research findings are mixed. Some researchers say that the roles for-
mulas play in the development of rules is negligible. An early review
by Weinert (1995) arrived at this conclusion: the use of formulas does
not contribute to the subsequent grammatical development in the L2.
Yorio (1989, p. 68) concluded that adult L2 learners “do not appear to
make extensive early use of prefabricated, formulaic language, and
when they do, they do not appear to be able to use it to further their
grammatical development.” Similarly, Wray (2002b, p. 116) suggests
that the formulaic mode of processing “tends to be avoided – with
some disadvantageous consequences – during post-childhood sec-
ond-language learning.”
In contrast, Bolander (1989) gave evidence of the presence of for-
mulaic speech in the language production of tutored adult learners of
Swedish as a second language. The use of “formula-like clauses” was
found to promote the application of syntactic rules, such as inver-
sion and the placement of the negative particle in Swedish. Formulas
were identified as

different types of frequent combinations of words constituting various

parts of sentences, that appear to be used as ready-made units in pro-
cessing… The criterion for regarding a sequence as a chunk/prefab is
the learners’ manifestation of them as such in their speech, as shown
either by a more frequent than average correct production of a cer-
tain structure, or by errors in structures that are otherwise correct.
(Bolander, 1989, p. 73)

Bolander also gave one suggestion as to how prefabs could lead to the
formulation of rules: “… when the number of prefabs stored in mem-
ory is large enough, syntactic rules are derived as help for the memory
158 Chapter 7. Formulaicity

to economize and rationalize processing” (Bolander, 1989, p. 85). Sim-

ilar conclusions were drawn by Myles, Hooper and Mitchell (1998),
who found that formulas in the acquisition of French as L2 facilitated
communication, accelerated production in the early stages, and also
contributed to the acquisition of grammatical rules.

Prosody

Going back to the connection between formulaic language and pros-

ody (discussed in 7.3.3.), Lin (2012) argues that prosody might be
more than just a characteristic feature of formulaic language. It
may be that “prosody underlies the mechanism by which we learn
and remember formulaic language” (Lin, 2012, p. 343). The main
point raised by Lin (2012) is that prosody is likely to play a key role
in formulaic language.
Lin (2012) looked at a study by Peters (Peters, 1977), who – accord-
ing to Lin – was the first researcher to notice that children can produce
chunks that are more complex than their current stage of develop-
ment would allow for. Also, importantly, some utterances produced
by young children, despite their low phonemic accuracy, have a dis-
tinctive melody which resembles the sound of those chunks in adult
speech. Lin interprets this as meaning that “prosody takes precedence
in children’s acquisition of formulaic language” (2012, p. 344), which
means that it is possible that “the prosodic form of the chunks is the
primary form that is stored in every entry of the formulaic chunks”
(2012, p. 344).
Lin writes that this is essentially bad news for language learners,
because of the “relative poverty of spontaneous spoken input” (2012,
p. 344). She specifically has in mind the English language learners in
China. However, even if the input were just as rich as that of L1 learn-
ers, the presence of developed cognition and literacy in L2 learners
7.6. Formulaicity and second language teaching: A brief history 159

means that the processes have to be different. Literacy has a par-

ticularly important role, as traditionally words are represented as
individual entities separated by spaces, which encourages their per-
ception as units, whereas in speech word boundaries are not really
perceptible.

7.6. Formulaicity and second language

teaching: A brief history

Very broadly speaking, the beginnings of the field of second lan-

guage acquisition research were characterized by a focus on gram-
mar. For quite a long time, it seemed that vocabulary was a relatively
neglected area. What followed was a quick growth of interest in the
lexical aspects of L2 learning and teaching in the 1990s. At the turn
of the millennium, a new topic of interest began to emerge in second
language acquisition research: collocations. Personally, when I first
researched the topic of collocations in the first years of this century,
there was still only a handful of (mostly very recent) research studies
which referred to this concept. In the literature on language teaching,
there was no trace of that term, and a thorough search in the indices
of language teaching methodology handbooks available in the early
2000s revealed that the term was never discussed (Leśniewska, 2003).
The Oxford Collocations Dictionary (2002) intended for learners of
English, was first published in 2002. The beginnings of the 21st cen-
tury also saw a growing interest in advanced levels of L2 proficiency,
including near-native levels, which brought into focus the colloca-
tional aspects of language competence (Howarth, 1996; Nesselhauf,
2003).
For the sake of the historical record, it should be noted that,
while none of the major publishing houses of ELT textbooks and
160 Chapter 7. Formulaicity

dictionaries had any materials on collocations in the 20th century,

there were some notable exceptions. A set of very collocation-fo-
cused textbooks of English was published in London by Macmillan
by a Polish linguist, Brygida Rudzka, and her colleagues (Rudzka,
Channell, Putseys, & Ostyn, 1981, 1985). There was a small diction-
ary of collocations published by John Benjamins (Benson, Benson,
& Ilson, 1986). In Poland, a dictionary with selected English colloca-
tions was published at roughly the same time (Douglas Kozłowska
& Dzierżanowska, 1988).
In the last decade, the picture changed completely. Dictionaries
of English have begun to include increasing amounts of collocational
information, and activities explicitly addressing collocations are now
a standard feature of ELT materials.

7.7. Articles and formulaicity

As could be seen in Chapter 5, the assumption behind the vast major-

ity of studies on articles carried out so far is that articles are indeed
form-function relations, but very complex ones. Virtually the entire
body of research on articles deals with the issue from the perspec-
tive of grammar. It seems that, to date, researchers have not explored
the use of articles in L2 English from the perspective of formulaic-
ity, with one exception, which is very relevant for this study: a cor-
pus-based study by Leńko-Szymańska (2012). Leńko-Szymańska
looked at the use of the definite article in the 3-grams found in a cor-
pus of ESL learner writing, and compared it to that in a native corpus.
The 3-grams, that is, three-word lexical bundles, included items such
as one of the, go to the, part of the, there is a, he was a, there is a. One
very interesting finding which emerged from this study is that, in
3-grams, the definite article occurs much more often (ca. 30% of the
7.7. Articles and formulaicity 161

uses) than the indefinite article (ca. 17% of the uses). In the learner
data, the also occurs more often in bundles than a/an, across all levels
of proficiency. A corpus approach like this one has some limitations,
however: it does not take into consideration correctness – some, prob-
ably many, uses of articles in the learner corpus may be incorrect; it
does not provide any data about the use of the zero article. However,
this approach has the benefit of clearly showing how the use of articles
in lexical bundles becomes more frequent as proficiency increases. In
fact, at advanced levels, the frequency in the use of the definite arti-
cle in bundles by L2 learners reaches that of native speakers, and for
the indefinite article it is actually higher than the native norm. At the
same time, the rule-based uses of articles fall below the native norm,
even at the advanced level. This finding is extremely interesting, as it
suggests that there may be a phraseological effect at play affecting the
way learners of English use articles, since articles are more likely to be
used by learners if they are part of a frequent combination of words.
Of particular importance is the fact that most of the existing stud-
ies on articles not only do not deal with formulaic aspects, but also
actually exclude any conventional uses of articles from their analysis.
Such studies may acknowledge the fact that some uses of articles are
motivated by convention rather than by the referentiality and speci-
ficity of a given NP, but they actually take pains to omit such conven-
tional or idiomatic uses from their results (Butler, 2002; Díez-Bedmar
& Papp, 2008) or the discussion (Thomas, 1989), which is understand-
able, since they are not the focus of those analyses. There are some
rare exceptions to this general approach: for example, Thomas (1989)
remarks that a high level of accuracy observed for one particular use
of the indefinite article in the [+SR –HK] context could be attributed
to the learners acquiring the structure there is a/an as a single chunk.
Idiomatic uses of articles were included in the studies by Ekiert
(2004) and Li and Yang (2010) and the ability to use a/an develops
earlier than the ability to use “zero article” Ø. These findings help us
162 Chapter 7. Formulaicity

understand why lower-intermediate and intermediate learners tend

to overuse the and a/an and underuse Ø, while advanced learners often
overuse Ø. For intermediate Chinese learners of English, tasks that
require them to choose articles before generic nouns ([−SR, +HK], with
Polish and Chinese learners, respectively. The authors elicited article
choices by means of a gap-filling test. The tested items included such
expressions as game of Ø cat and Ø mouse, getting Ø cold feet, or thrown
out of Ø work. The results showed that the idiomatic use of articles,
understood in this way, is problematic for both groups of learners.
In view of the growing recognition of the fact that language process-
ing is formulaic to a considerable extent, as reviewed above, it stands to
reason that the use of articles in English may also to some extent rely
on formulaicity-related mechanisms. Even those uses of articles which
appear to be rule-governed (i.e., syntactically regular) may in fact be
aided by the mechanisms which are responsible for formulaicity in lan-
guage use. This possibility will be explored in Chapter 8.

7.8. Conclusions

The nature of language production can be accounted for in terms of

syntactic rules, on the one hand, and the retrieval of formulaic lan-
guage, which can also be broken down into component parts if there is
need to do so, on the other. The role of formulaic language is explained
by the economics of language production: it represents a compromise
between increased demands on storage space and gains in the speed
of production and ease of comprehension. It is now well established
that native speakers process various types of formulaic language in
a way that offers an advantage over individual word selection. Those
findings fit in well with current developments in recent developments
in linguistics and psychology, such as usage-based grammars, and
7.8. Conclusions 163

models which emphasise probabilistic and statistical learning. It is

known that frequency of occurrence of specific word strings corre-
sponds to the likelihood of the string being processed in a formulaic
manner, even though what formulaic processing technically means
exactly is not yet entirely clear.
The research findings are less clear with respect to L2 learners,
but all in all indicate that L2 learners are sensitive to frequency of
word combinations in the input, and that sensitivity improves with
the level of advancement.
Therefore, it seems that the conclusion of Conklin and Schmitt
seems fully justified:

[A]dult native speakers, and most likely children and non-native

speakers who have had enough exposure to a language, appear to
have representations not only for the words that make up formulaic
sequences (fish, and, chips) but also for the sequence itself (fish and
chips). Frequency seems to lead to a particular form being repre-
sented in the mental lexicon. However, if a form has not been encoun-
tered frequently enough, as in the case of lower proficiency non-na-
tive speakers or very young children, it appears that it may not be
well entrenched in memory.
(Conklin & Schmitt, 2012, p. 54)

Taking the above into consideration, it seems reasonable to assume

that formulaicity is related to the frequency of occurrence, even
though the exact nature of this connection has not been fully elu-
cidated yet.
8.1. Introduction 165

Chapter 8

Investigating article use

by advanced Polish learners of EFL:
The role of formulaicity

8.1. Introduction

This chapter presents an inquiry into the phraseological aspects of

article use. As has been shown so far, there is extensive theoretical
support in the literature for the idea that language use is to a consid-
erable degree phraseologically motivated, and some empirical evi-
dence has already accumulated which corroborates this view. It has
also been shown that linguistic attempts to capture the functioning
of the English articles in the form of rules are not only extremely
complex, but they have relatively little predictive power, which neg-
atively impacts their usability in English language teaching. In view
of the overwhelming number of rules pertaining to article use and
the apparent contradictions between them, it stands to reason that
some facilitating mechanism may be playing a role in the correct use
of articles, both for native and non-native speakers. As language users
166 Chapter 8. Investigating article use by advanced Polish learners of EFL…

rely to a significant extent on the partly automatized retrieval of some

word combinations, and articles are part of those word combinations,
it is possible that making correct article choices is aided by the “idiom
principle.” This would hold true even when the use of an article can
be explained by reference to a specific grammatical rule.
This chapter presents two investigations which put this hypothet-
ical possibility to the test, bringing together the research I have con-
ducted on the topic and showing the development of my methodo-
logical approach to investigating the issue. The first study reported
here has been published previously (Leśniewska, 2016). For that rea-
son, this chapter includes sections of text that are reproduced from
the original article: section 8.2.1 on the participants, section 8.2.2 on
the instruments and the procedure, and section 8.2.3 on the analysis,
results and discussion.

8.1.1. Purpose and rationale

As could be seen in Chapter 5, studies on articles in L2 English are

numerous, but the assumption which underlies most of them is that
articles are rule-governed. This perspective is largely justified, as the
English article system is indeed based on the key notions of definite-
ness, specificity and genericity, as well as on the countability and
number of nouns, and on the distinction between proper and com-
mon nouns. However, the rules provided never account for the total-
ity of article uses in natural language, as was demonstrated in Chap-
ters 2 and 3.
This chapter presents an enquiry into the possibility that there is
yet another perspective which may complement the existing picture of
article use, namely a phraseological one. In other words, the research
included here is an attempt at examining if the correct use of articles
by language learners is to some extent aided by the mechanisms that
8.1. Introduction 167

underlie the formulaic character of language. It is important to stress

that the focus here is on the possibility that formulaic “bonds” com-
plement, or further strengthen, the use of articles in those contexts
which are compliant with the “rules.”
It is also important to note that the role of formulaicity which is
hypothesised in this study is crucially different from the role usually
assigned to phraseological aspects in descriptions of article use. For-
mulaicity is usually called upon to account for those uses of articles
which somehow do not conform to the rules. For example, the use of
articles in expressions such as make a start, in a hurry, on loan, out
of action are in ESL materials explained as “idiomatic,” which means
that the learner should memorise them rather than try to apply any
of the rules on article use. As we can read in a popular pedagogical
grammar: “In a lot of idiomatic expressions articles are used or left
out for no apparent reason other than they belong or don’t belong
in the expression. Learners need to learn these like items of vocabu-
lary, and have to remember the whole phrase, ignoring general rules
or sub-rules” (Parrott, 2000, p. 50). For example, if we take the Eng-
lish idiomatic expression dog eat dog and see it from the perspective
of a learner of English, it is understandable that the learner needs to
suppress the need to make the expression grammatically regular by
adding articles and the third person-s ending to the noun (a dog eats
a dog) or making the nouns plural (dogs eat dogs). It is thus a case of
a clash between grammatical rules and the overriding, phraseological
rule. In this study, however, I am concerned not with situations where
the two are in conflict, but with those when the grammatical rules
and the phraseological binding coincide, the idea being that the latter
aids the former. For example, in the phrase for the first time the use
of the is regular, but it may still be easier for the learner to make that
correct choice and use the definite article in this phrase because the
entire combination of words is to some extent familiar in its entirety
(that is, it is formulaic).
168 Chapter 8. Investigating article use by advanced Polish learners of EFL…

8.1.2. Approach

Since it is assumed that formulaicity is related to frequency of use (as

was discussed in Chapter 7), it follows that frequent word combina-
tions are more formulaic than rare ones, all other things being equal.
The studies presented below utilize this assumption in their design.
Both feature research tools which allow for a comparison of the par-
ticipants’ use of articles in two types of contexts, comparable in all
ways except for their level of formulaicity. This will allow to check
the prediction that the participants’ success in using articles will be
greater in the more formulaic of two otherwise comparable contexts.
The research tools used in both studies were designed specifically for
this purpose, as no similar research seems to have been carried out.

8.2. Study 1

8.2.1. Participants

The participants in that first study were 90 Polish university students

majoring in English language or linguistics. All were adult learners of
English placed at the B2/C1 level in the Common European Frame-
work of Reference for Languages (CEFR) (Council of Europe, 2011).
The learners’ L1 was Polish, an article-less language.
The learners had a specific profile: they had been learning Eng-
lish in a predominantly instructional context. As opposed to studies
which involve ESL in immigrant settings, the learners in this study
were unlikely to function socially or professionally in English in
their everyday lives. This is relevant insofar as the instruction they
had received as far as articles are concerned is extremely likely to
have included metalinguistic explanation and focus-on-form type
8.2. Study 1 169

of teaching, just because of the general characteristics of the English

courses taught at Polish schools.
The level of advancement of the students was assumed to be
reflected in their placement in groups for their practical English
classes. This general indication of the level of proficiency was con-
sidered sufficient for the purpose of this study. The first group (n = 44)
gathered students whose level of English corresponded to the B2
level on the CEFR. The level of ESL proficiency of the second, more
advanced group (n = 46) corresponded to the C1 level. All partici-
pants were between 20 and 22 years of age, with the mean age being
slightly lower in Group 1.

8.2.2. Instrument and procedure

The basic idea for the research instrument was to create a test in which
the participants would have to supply missing articles in contexts
which feature different levels of formulaicity. The test would thus
include a number of obligatory occasions for the use of the definite
article. Those occasions would be of two kinds: in some, the sequence
of words requiring an article would be a frequently occurring com-
bination, in some, it would be a novel combination of words, charac-
terized by low frequency of occurrence. In other respects, the pairs
of test items (high frequency item – low frequency item) would be as
comparable as possible.
The one-page test used in this study (see Appendix 1) consisted of
sentences in English from which all the articles had been removed.
The participants were asked to put in the missing articles in the right
places. The tests included a total of 12 pairs of target items (presented
in a mixed-up order) which included exactly the same structures with
articles: the definite, the indefinite, or the zero article. The pairs were
as follows:
170 Chapter 8. Investigating article use by advanced Polish learners of EFL…

Item A: Item B:
1. a friend of mine – an acquaintance of mine
2. what a shame – what a remarkable player
3. twice a day – five times a semester
4. the sooner the better – the smaller the pot, the more critical the problem
5. a cup of tea – a spoonful of syrup
6. the day I die – the food I brought
7. help the poor – open to the insured
8. hit (someone) in the face – cut in the hand
9. speak English – learn Kurdish
10. get a job – live in a luxury apartment
11. have kids – eat carbohydrates
12. the centre of attention – the ecology of waterways

Grammatically speaking, the reason for the use of the article was
identical in item A and B of each pair, that is, the same grammati-
cal “rule” applied in both cases. For example, items 5A and 5B both
represent partitive expressions (a type of phrasal quantifiers) used to
impose countability on non-count nouns (Quirk & Greenbaum, 1973,
p. 67). Items 10A and 10B are both examples of the use of the indefinite
article with referents that can be classified as countable, indefinite,
and non-specific (Downing & Locke, 1992, p. 429). However, the items
differed in one important aspect: the article in version A was con-
tained in a frequently occurring word combination, whereas version
B was relatively more of an “open choice” type of word combination.

Test preparation

Test preparation relied on a combination of researcher intuition,

native speaker judgements, and frequency measures from corpus
examination. The most challenging step in the process was to find
8.2. Study 1 171

suitable word combination pairs which would qualify as, respectively,

more “idiom-principle-driven” and more “open-choice” in character.
A brainstorm session between two linguistics researchers aimed
at identifying pairs of word combinations that were perceived by the
researchers as being more typical and frequent, versus more “open-
choice” combinations. Intuitive ratings thus formed the basis for the
initial selection of word combination pairs. Those pairs were then sub-
mitted to two colleagues who are native speakers of English, which led
to further elimination of pairs on which there was lack of inter-judge
agreement, the replacement and changes to some word combinations,
and a resulting group of 20 word-combination pairs.
The frequency of co-occurrence for those initial intuition-based
pairings was verified using two corpora: the British National Cor-
pus (Davies, 2004) and the Corpus of Contemporary American Eng-
lish (Davies, 2008), which will be referred to as BNC and COCA,
respectively. Those corpora were deemed adequate due to their size
(100 million and 450 million words, respectively) and representative
character. The BNC is made up of written (90%) and spoken (10%) lan-
guage, and contains texts from a wide range of sources (for example,
different kinds of journals, periodicals, newspapers, academic books,
popular fiction), in order to represent a wide cross-section of British
English. The COCA is also a balanced corpus, made up of texts rep-
resenting spoken language, fiction, popular magazines, newspapers,
and academic texts.
While both corpora were used in the initial search in order to
locate suitable pairs for the test, a specific threshold was set with ref-
erence to the COCA corpus, the frequency findings from which were
considered more reliable because of its larger size. The frequent com-
bination in each pair had to occur at least 40 times more often than its
rare counterpart in order to be included in the test. The rare items had
a frequency of 0.02 per million words or less, while the frequent items
had a frequency of 0.18 or more. While it was impossible to determine
172 Chapter 8. Investigating article use by advanced Polish learners of EFL…

a perfect set of criteria which could be applied if there was a way to

extract the items automatically, this frequency requirement was con-
sidered to provide sufficient support for the intuitive judgements.
Not all intuition-based pairs corresponded to corpus-based fre-
quency data, nor were frequency counts always similar in both cor-
pora. Consequently, the 12 pairs which showed the most convincing
difference in the frequency of occurrence were retained. The full list
of items and frequencies is provided in Appendix 2.
The 12 pairs (24 target items) were presented hidden among other
sentences in the test, which not only helped to provide more con-
text for the target items, but also to make the relationship between
the pairs of target items less noticeable. It should be noted that all
the articles were removed from all the sentences included in the
test. Only some of the missing articles in the test were the actual
target items. Since all articles were removed, the number of miss-
ing articles was larger than the number of the target items under
investigation.
For example, in the case of pair no. 3, which tested the use of the
indefinite article in expressions of frequency, the more frequent of
the two combinations, twice a day, appears in test item 7:

By midsummer, herbs and vegetables in containers may need water

twice day.

Whereas its counterpart, the open-choice combination five times

a semester, can be found in test item no. 4:

We meet regularly, five times semester, at departmental meeting.

The noun phrase departmental meeting also requires an article, but

whether the participants inserted it or not was not taken into con-
sideration, as this noun phrase was not one of the target items. Such
8.2. Study 1 173

missing articles outside the target items helped distract the test-tak-
ers from any pattern in the test design they might be able to discern.
An initial version of the test was piloted with three native and
three non-native speakers of English to ensure that the removal of
articles did not create ambiguous or incomprehensible sentences, as
well as to check if, for all the target items, all the native speakers
always provided the same response. The target items which did not
meet this criterion were replaced. Variation in the native speakers’
choices of articles in the test outside the target items was considered of
no importance. Rare or difficult lexical items were avoided in the test.
Care was taken to ensure that both the frequent and the rare com-
binations of words were composed of “ordinary,” relatively frequent
lexical items which are expected to be known to learners of English
at the intermediate and higher levels. Two experienced teachers of
English were consulted about the likelihood that all the words used
on the test would be known by our target audience. Teachers were
convinced that all items would be known by our test participants,
and post-test conversations with a few participants confirmed that
no lexical item on the test was new to them. Difficult words, due to
their greater length and other difficulty-inducing factors, could affect
the processing of the test sentences in ways which could not entirely
be controlled, and they could interfere somehow with article use. For
the same reason, the test was composed in such a way as to avoid false
cognates (for speakers of L1 Polish) or any ambiguity.
In contrast to most tests on article use, which tend to have the
classic format of a cloze test, the instrument used in this study elic-
ited article use in a slightly different way: the text did not have gaps
indicating where the participants needed to provide an article. The
rationale for choosing this test design was that it is more similar to the
actual use of articles than a cloze test. On a cloze test, the test taker
receives a signal that an article may be missing at a specific location.
In the case of those tests where the zero article is one of the options,
174 Chapter 8. Investigating article use by advanced Polish learners of EFL…

the difference between the two formats is admittedly minor, but is still
exists. In the gapped version the test taker is specifically prompted, or
encouraged, to consider using an article at a specific place, and in the
design employed in this study there is nothing in the test that suggests
the need for an article at a specific place.

Procedure

The participants took the test on a voluntary basis, in a classroom set-

ting. They were encouraged to take the test to see how well they could
use articles. The tests were coded with random three-digit combina-
tions. Each participant was asked to write down their test number so
that they would be able to retrieve their results in the future. The par-
ticipants were asked to put in the missing articles in the right places.
They were encouraged to follow their intuition about what sounds
right. The test took around ten minutes for most participants. A time
limit of 8 minutes was enforced.

8.2.3. Analysis, results and discussion

In the analysis of the data, dichotomous scores were compiled: one

point was awarded for inserting a correct article, and 0 points for
omitting to insert an article or for inserting an incorrect one. In the
case of test items with the zero article, 1 point was given for not pro-
viding an article, and 0 points for providing an indefinite or definite
article.
The mean item score was calculated for the frequent and for the
rare uses for all 90 participants, as well as for each group separately.
Those mean item scores are presented in Table 2. A t-test was per-
formed to compare means.
8.2. Study 1 175

rare frequent t-test

combinations combinations
All participants 0.68 0.85 t = 9.50;
(n = 90) p < 0.00001
Group 1 0.53 0.75 t = 7.44;
(less proficient, n = 44) p < 0.00001
Group 2 0.82 0.94 t = 6.64;
(more proficient, n = 46) p < 0.00001

Table 2. Mean items scores for frequent and rare uses.

As shown by the very low p value yielded by the t-test for all 90 par-
ticipants as a group, the mean for the frequent combinations is sig-
nificantly higher than that for the rare ones. In other words, for the
same articles and the same grammatical rule, the participants tend
to be more successful when using the articles in those combinations
that occur more frequently, and less likely to be correct with the less
frequent combinations.
When analysed separately, both groups show higher success rates
in the case of the frequent combinations than the rare ones. In both
cases the difference is statistically significant. However, in the case
of Group 2 (more proficient) the difference between rare and fre-
quent combinations was smaller: a difference in means of only 0.12,
compared to a difference of 0.22 in the case of the less proficient
Group 1.
The fact that the difference between the rare and frequent combi-
nations becomes smaller as the level of proficiency increases is under-
standable: ultimately, with very advanced language competence, there
would be very little difference, as articles would be used mostly cor-
rectly in all cases for both the frequent and rare word combinations.
It should be noted that the predicted higher means for frequent
items were not obtained in the case of all the pairs on the test, as
shown in Table 3.
Version Article Target item Mean SD t-value p (t)
Frequent: a a friend of mine 0.80 0.4
an acquaintance 4.75 <.0001
Rare: 0.48 0.5
of mine
Frequent: a what a shame 0.79 0.41
what a remarka- –0.37 .71
Rare: 0.81 0.39
ble player
Frequent: a twice a day 0.89 0.31
five times 1.83 .04
Rare: 0.79 0.41
a semester
the sooner the
Frequent: the 0.84 0.36
better
the smaller the 10.35 <.0001
Rare: pot, the more criti- 0.23 0.43
cal the problem
Frequent: a a cup of tea 0.99 0.11
a spoonful of 1.94 .02
Rare: 0.90 0.25
syrup
Frequent: the the day I die 0.94 0.23
4.83 <.0001
Rare: the food I brought 0.68 0.47
Frequent: the help the poor 0.73 0.45
open to the 2.99 <.01
Rare: 0.52 0.5
insured
hit (someone) in
Frequent: the 0.69 0.47
the face 4.04 <.0001
Rare: cut in the hand 0.40 0.49
Frequent: zero speak English 1 0
2.52 .01
Rare: learn Kurdish 0.93 0.25
Frequent: a get a job 0.89 0.32
live in a luxury –0.24 .81
Rare: 0.90 0.3
apartment
8.2. Study 1 177

Version Article Target item Mean SD t-value p (t)

Frequent: zero have kids 1 0
3.34 <.01
Rare: eat carbohydrates 0.89 0.32
the centre of
Frequent: the 0.61 0.49
attention
0.6 .55
the ecology of
Rare: 0.57 0.5
waterways

Table 3. Compared scores for item pairs (n = 90).

Out of the 12 pairs of frequent-vs-rare, the differences between

the means for the frequent items and their rare counterparts as
shown by a t-test is statistically significant (at p < .05) for nine item
pairs, and not significant for three items. The items for which the
effect was not observed include: what a shame – what a remark-
able player, get a job – live in a luxury apartment, and the cen-
tre of attention – the ecology of waterways. In the case of the first
pair it is relatively easy to come up with a possible explanation for
the observed lack of a formulaicity effect. While the phrase what
a shame is definitely much more frequent than the rather “open-
choice” word combination what a remarkable player, the nouns
player and shame differ in the degree to which they are countable.
First of all, player is a concrete and shame an abstract noun, and,
as has been noted (see e.g. Amuzie & Spinner, 2013), the degree of
success in article use depends on this distinction (with abstract
nouns being more difficult to use correctly with articles) but also
on other more nuanced distinctions which result from the degree
of boundedness of a given noun. It is, therefore, possible, that the
abstract and less countable character of shame reduced the phrase-
ological advantage which was expected on the basis of the phrase
what a shame being frequent. For the pair get a job – live in a lux-
ury apartment, one plausible explanation is that the combination
178 Chapter 8. Investigating article use by advanced Polish learners of EFL…

live in a luxury apartment was generally very easy for the test-tak-
ers; apartment being a clearly countable, concrete noun. The mean
scores for both items were very high (0.89 and 0.9, respectively),
which means that the effect of formulaicity, if any, may not have
registered because of a kind of ceiling effect with the rare combi-
nation. For the last pair which did not show a difference, the cen-
tre of attention – the ecology of waterways, it is difficult to provide
a plausible explanation of this fact.
Two item-related issues need to be addressed. One is the possible
effect of adjectival pre-modification on the use of articles with nouns.
In two of the test items the noun happened to be pre-modified by an
adjective (a remarkable player, a luxury apartment), which introduced
the possibility of another variable confounding the results, as there
are reasons to believe that adjectival pre-modification may somehow
interplay with article use by L2 English learners. This is possible in
view of the fact that Trenkic (2008) found that learners from arti-
cle-less language backgrounds tend to omit articles more in adjec-
tivally pre-modified (Art+Adj+N) than in non-modified contexts
(Art+N). She also offered a “syntactic misanalysis account” (Tren-
kic, 2007, 2008), which links the failure to use articles to the fact that
articles are treated as adjectives.
In this study, the two items which included pre-modified nouns
in the target items belonged to the “rare” category, thus potentially
contributing to the expected lower scores for those items because of
a variable that was not taken into consideration. However, for the two
pairs in which the two items occur, what a shame – what a remarkable
player and get a job – live in a luxury apartment, the expected effect was
not observed. In other words, the learners were similarly successful in
providing an article in both the rare and the frequent item, despite the
fact that the rare item was additionally more likely to be more difficult
due to the use of an adjective. Thus, in this study, the issue of adjectival
pre-modification did not appear to play a role in article use, at least as
8.2. Study 1 179

far as one can tell on the basis of the two target items which featured
adjectival pre-modification.
Another issue which needs to be addressed is the level of difficulty
of some of the words. It is true that some of the “rare” items feature
words of somewhat lower frequency than the “frequent” combina-
tions. However, all the lexical items in both types of expressions were
expected to be familiar to the learners, as explained in the instrument
and procedure section.
An interesting finding concerns the types of wrong answers
provided by the participants on the test. As stated above, in the pro-
cess of compiling dichotomous scores, one point was awarded for
a correctly supplied article, and 0 points were given for omitting
to insert an article or for inserting an incorrect one. Out of the 512
answers for which the score was zero, an overwhelming majority
(475 answers, which is almost 93%) were answers which were wrong
because no article was provided. Only 37 answers were cases in which
the wrong article was supplied. This indicates that, regarding arti-
cle use by learners from article-less L1 backgrounds, failing to pro-
vide an article is much more common than providing an incorrect
one. Of course, failing to use an article can also be seen as a case of
wrong article choice, namely, the choice of the zero article. However,
it is impossible to distinguish between the use of the zero article and
failing to use any article (cf. Leńko-Szymańska, 2012), nor is it cer-
tain that making a distinction of this kind is feasible (see section
1.5). It should be noted that the format of the test used in this study,
which did not provide a prompt to use an article in specific place in
the text (as, for instance, a gapped article test would do) may have
contributed to the notable underuse of articles. As far as the present
analysis goes, whatever the reason for failing to use an overt article,
it remains an interesting finding in its own right that the participants
were much more inclined not to use an article than to use the wrong
one of the two overt articles (a(n) or the).
180 Chapter 8. Investigating article use by advanced Polish learners of EFL…

8.3. Study 2

8.3.1. Participants

The participants were 200 undergraduate university students major-

ing in English studies. They were all native speakers of Polish. The stu-
dents’ placement in groups for their practical English classes reflected
their level of proficiency, which was considered a sufficient indica-
tion of their level of knowledge for the purposes of the present study.
The first group (n = 100) was at the B2 level on the CEFR (Council of
Europe, 2011), whereas the second, more advanced group (n = 100)
was at the C1 level. All participants were between 20 and 23 years of
age, with a mean age slightly lower for Group 1.

8.3.2. Instrument and procedure

This study investigates the use of articles by ESL learners by draw-

ing on the concept of corpus-derived lexical bundles (see section 7.2).
The main idea behind the design of the research tool was similar to
that of Study 1, which was to compare learners’ article use in pairs
of comparable contexts which differ only with respect to the level of
formulaicity. However, it adopted a different method of finding suit-
able test items; also, rather than have a limited number of items on
each type of article use, this study focuses only on the use of the
definite article.
To create an instrument that is easy to administer and does not
discourage or tire test takers, efforts were taken to avoid repetitiveness
while at the same time reaching a sufficient number of article occur-
rences, so that the target items would become “hidden” in a sufficient
amount of text. The final test consisted of two parts, with isolated
8.3. Study 2 181

sentences in one section and short texts in the other. The procedure
that led to the creation of the test is described below.

The selection of test items

Suitable target items were selected on a data-driven basis, using the

COCA, which was selected for its relatively large size (520 million
words), and the fact that it is balanced for genre.

A) high-frequency items

A list of the most frequent 4-grams containing the definite arti-

cle the was retrieved from the COCA. The decision was made to
study 4-grams, or 4-word bundles, as they are the most commonly
studied type of lexical bundle (see e.g. Chen & Baker, 2010). Hyland
gives the following explanation for this fact: “Three-word bundles
are extremely common, and tend not to be very interesting, while 5-
and 6-grams are comparatively rare and often subsume shorter ones.
Four-word bundles seem to be most often studied, perhaps because
they are over 10 times more frequent than five-word sequences and
offer a wider variety of structures and functions to analyse” (Hyland,
2012, p. 151).
As was said before, bundles are simply the most frequently occur-
ring word sequences of a given length (in this case, 4-word strings).
When a list of such sequences is extracted from a corpus and ranked
according to decreasing frequency, the question arises as to the fre-
quency beyond which the sequences lose their bundle status. Of
course, such a cut-off point must be arbitrary, as we are dealing with
not a dichotomy, but a continuum. However, for practical purposes
such a distinction is made, and some kind of consensus exists, as
most studies put that cut-off point between the frequencies 10 and
182 Chapter 8. Investigating article use by advanced Polish learners of EFL…

40 times per million words, depending on the purpose of the study.

Biber and colleagues (e.g. Biber et al., 1999; Conrad & Biber, 2004)
set the frequency minimum for 4-word sequences to be considered
bundles at ten times per million words. This limit was applied in the
present study, yielding a list of 42 bundles containing the definite
article. Of those 42 bundles, only the ten most frequent are shown
in Table 4, for the sake of brevity, but they are very representative of
the rest, as the entire group of 42 bundles shares many similarities.
When we apply the classification offered by Biber and his colleagues
(1999) to the 42 bundles with the, it turns out that a vast majority
belong to the category of “prepositional phrase expressions” (such as
at the end of, in the middle of, for a couple of ). Most of the remaining
4-grams belong to the type classified as “noun phrase expressions,”
which consist of the beginning of a noun phrase, very often includ-
ing of, or other noun phrases and noun phrase fragments. Examples
include the end of the, the rest of the, the other side of the, the one with
the, the last time I (Biber et al., 1999, p. 1012).
In order to classify the bundles on the list in notional terms, they
were compared with the taxonomy offered by Erman and Warren
(2000), according to which they fall mostly into the category “period
or point in time” and “places and positions.” This is consistent with
the observation made by Biber and colleagues that frequent noun
phrase bundles usually identify a physical location, while preposi-
tional phrase expressions are used primarily as time or place adverbi-
als, such as I get paid at the end of the week (Biber et al., 1999, p. 1013).
An exception to this trend is the bundle on the other hand, which is
classified by Erman and Warren as grammatical, that is, “intralin-
guistic text-forming” rather than a unit with “extralinguistic refer-
ence” (Erman & Warren, 2000, p. 41).
8.3. Study 2 183

1. in the united states 69.28

2. the end of the 59.17
3. at the same time 55.20
4. at the end of 47.82
5. the rest of the 45.40
6. for the first time 45.37
7. one of the most 38.80
8. on the other hand 36.69
9. of the united states 34.53
10. in the middle of 34.44

Table 4. The top ten most frequent 4-grams containing the definite article, extracted from the COCA.

For the purpose of this study, the decision was made to retain the bun-
dles belonging to the category of prepositional phrase expressions for
developing the research tool, for the following reasons:

1. Any 4-gram may happen to contain up to two articles, and the

article(s) can occur in any of the four slots, but not adjacent ones
(unless the four-word sequence comes from a transcript of speech
and includes an instance of repetition or self-correction). Preposi-
tional phrase bundles all feature an article in the second slot of the
4-gram. The other large group among the frequent bundles, noun
phrase expressions (such as the end of the, the back of the), usu-
ally feature articles in the initial or final position, or both. When
the article occurs in the initial or final position, the 4-gram only
captures its syntagmatic environment on one side. Since the study
aims at exploring the role of some kind of a phraseological “bind-
ing” in the use of articles, the article must not be on the edge of
the 4-gram, because we wish to consider its syntagmatic environ-
ment on both sides. Obviously, articles which are captured in the
final slot of a bundle are in a syntactic and semantic relationship
184 Chapter 8. Investigating article use by advanced Polish learners of EFL…

with the rest of the noun phrase which is not included, as in e.g.
the bundle is one of the.

2. The prepositional phrase bundles were the most numerous among

the frequent bundles (above 10 per million). Therefore, in order
to eliminate a potential variable that would affect a minority of
the items, it was thought better to focus on one type of bundle for
designing the research instrument.

3. This decision also provided a solution to the usual problem that

researchers working with lexical bundles have to deal with:
the need to eliminate or conflate overlapping bundles (see e.g.
Hyland, 2012, for an explanation). Such overlapping bundles can
be seen, for example, in Table 4: the end of the and at the end of.

4. Prepositional phrase bundles are the ones which structurally cor-

respond to traditionally identified bundles or prefabs (e.g. Erman
& Warren, 2000), such as in the context of, at the beginning of, etc.

Additionally, several other items were eliminated, which are listed

below.
• Two items which were idiomatic in character, in the sense of
having a non-compositional meaning: on the other hand or
in the face of. In those two cases, there is significant opacity
of meaning, which could introduce an additional variable. In
ESL instruction, such items are likely to be taught as separate,
non-compositional chunks, with the meaning assigned to the
whole chunk.
• A somewhat similar case was represented by in the midst of,
which contains a word which is rare and dated if used outside
of this specific word combination. For that reason, it was also
removed.
8.3. Study 2 185

Another important issue was that of the extent to which the target
items would constitute obligatory occasions for the use of the definite
article. Here, the situation is further complicated by the specific way
in which articles are used in English. In contrast to many other mor-
phemes, the use of articles in English is characterized by a fair amount
of optionality. This means that even for contexts where the article is
normally required, it may be omitted in certain situations, registers,
or contexts, for example, to indicate a shorthand style caused by the
speaker’s haste, or in humorous uses of language. In some text types
it is customary to leave out articles (for example, in culinary recipes),
or in some elements of certain texts, for example newspaper headlines.
There may also be other special uses, as illustrated below, such as the
names of court cases. Below, there are some examples of the same
word combinations which occur in our frequent bundles (in the mid-
dle of, at the same time, in the United States), but without the article,
as found in the COCA:

Place mixture into muffin tins sprayed with cooking spray; bake in
middle of oven for 30 minutes.
At same time, roast pork until instant-read thermometer inserted in
center reaches 145°F.
For example, in United States v. Wilson, the defendant was charged
and convicted…

This is not the only reason, however, why article-less versions of the
frequent bundles (for example, for the first time, in the Middle East,
in the United States) may be found in the corpus. The may happen to
be part of a sequence which pre-modifies a noun, as in:

A guide for first time researchers in education and social science…

…a senior lecturer in Middle East politics at the University of Exeter…
…a symphonic band like those found in United States high schools…
186 Chapter 8. Investigating article use by advanced Polish learners of EFL…

However, for most of the lexical bundles, there was a striking dispar-
ity between the number of the occurrences of the lexical bundle, and
its version without the article. For example, the bundle at the same
time appears almost 29,000 times in the corpus, whereas at same
time appears only 29 times, so the former is a thousand times more
frequent. For other bundles the proportions were similar. There were
only two exceptions: in the case of and of the things that. These two
bundles are frequent, of course, but their article-less versions are
also frequent: in case of appears 1478 times, and of things that 3484
times. Such a high number of occurrences of the article-less version
of the two bundles casts doubt on the suitability of these two items
for the present test. However, this conclusion needed to be further
verified, because of aforementioned issues with looking up article
use in a corpus, and also because the bundles had to be placed in
sentence contexts for the test, which would affect whether the arti-
cle would be needed or not. The sentences containing the high-fre-
quency bundles were thus piloted on a group of native speakers
(n = 15). A unanimous decision to put an article in a specific place
was taken to mean that the place is an obligatory occasion for the
use of the definite article. Indeed, the native speakers’ judgments
confirmed that the two items (in the case of, and of the things that)
had to be eliminated, as there was variability in the use of articles
in those two cases.

B) low-frequency items

For each of the 15 frequent 4-grams that remained after the elim-
ination process described above, a corresponding low-frequency
sequence was selected. The 15 bundles fall into two types, syntacti-
cally speaking. The two types are described below, with the pedagog-
ical rules that are relevant to the use of article they represent.
8.3. Study 2 187

I) Preposition + definite article + adjective + noun

a) Plural geographical names are one of the sub-classes of proper
nouns which are used with the definite article, for example, the
Shetlands (e.g. Quirk & Greenbaum, 1973, p. 80 section 4.30);
this rule is indicated in Table 5 as PL GEO NAME.
b) A definite article is needed because of adjectival pre-modifica-
tion, as in this is the main bedroom (Sinclair, 1990, p. 46); this
rule is indicated in Table 5 as DEF-MOD.

II) Preposition + definite article + noun + preposition

A definite article is needed because of postmodification by an
of-phrase (“postmodification by an of-phrase usually requires the
definite article with a head noun, which thus has limited generic (par-
titive) reference” (Quirk & Greenbaum, 1973, p. 71, section 4.19). This
rule is indicated in Table 5 as OF-PHRASE.
For both bundles of type I) and type II), the following procedure
was applied when searching for suitable low-frequency equivalents.
Strings analogous to the frequent ones were found in the corpus, and
one was selected that had the lowest possible frequency, while also
meeting other criteria. Suitable phrases had to be semantically compo-
sitional and represent instances of article use which are regular in the
grammatical sense. Care was taken not to introduce additional varia-
bles: nouns were replaced with nouns which had a comparable level of
concreteness, the same number, and when possible, pertaining to the
same semantic field. The same “rule” was supposed to be represented
by the high-frequency and the low-frequency item. This approached
resulted in the following replacements, listed by type of bundle:
• For type Ia: the adjective and the noun in a geographical name
were replaced by another name with the same make-up: the
United States – the Kuril Islands.
• For type Ib: The adjective and the noun were replaced: for
example, for the first time – for the last lap.
188 Chapter 8. Investigating article use by advanced Polish learners of EFL…

• For type II: The noun was replaced: e.g. at the top of – at the
underside of.
The final pairs (lexical bundle – corresponding low frequency
sequence) are listed in Table 5.

high- freq low- freq pedagogical

frequency per frequency per rule
sequences million sequence million
(bundles)
1. in the United 69.278 in the Kuril 0.002 DEF-MOD
States Islands / PL GEO
NAME
2. at the same 55.196 at the right level 0.029 DEF-MOD
time
3. at the end of 47.819 at the closure of 0.010 OF-PHRASE
4. for the first 45.373 for the last lap 0.006 DEF-MOD
time
5. in the middle of 34.436 in the nexus of 0.017 OF-PHRASE
6. on the other 15.380 on the far edge 0.031 DEF-MOD
side
7. on the basis of 13.988 on the merit of 0.017 OF-PHRASE
8. in the form of 13.590 in the phrasing of 0.015 OF-PHRASE
9. for the rest of 12.821 for the width of 0.023 OF-PHRASE
10. at the top of 12.706 at the underside of 0.059 OF-PHRASE
11. in the middle 11.404 in the distant 0.002 DEF-MOD
east north
12. on the part of 11.244 on the factor of 0.008 OF-PHRASE
13. at the time of 10.638 at the period of 0.008 OF-PHRASE
14. at the 10.085 at the dawning of 0.021 OF-PHRASE
beginning of
15. in the context of 10.046 in the chance of 0.015 OF-PHRASE
Table 5. The high-frequency and low-frequency pairs.
8.3. Study 2 189

The creation of sentences and texts

Like in the case of the frequent items, a sentence context was pro-
vided for the rare items. Some of the sentences containing the tar-
get items were combined into paragraphs, to avoid the monotony
of too many single sentences. The final shape of the test was 16 sin-
gle sentences and two paragraphs. Naturally, the test included more
instances of article use (both definite and indefinite) than just the
articles located in the target items (all definite). In the next step, all
the articles were removed from the text and the place where they
were missing was not indicated. The task of the test-taker was to
supply the missing articles.
The entire test was piloted again on a group of native speakers
of English (n = 15) and this time also on learners of English, pri-
marily to ensure that the removal of articles did not create ambig-
uous or incomprehensible sentences. A few problematic sentences
were modified with the target items remaining intact. Some of the
native speaker judges on whom the final test was piloted (who were
asked to supply articles, but they were not given an explanation as
to the purpose of the test) commented on the apparent strange-
ness of some of the low-frequency items, saying that they would
recommend replacing them with a “more idiomatic” expression,
or a phrase which is stylistically more felicitous. This was taken
as confirmation of the suitability of those particular sequences for
the purposes of this study, and this reaction was understandable,
given that sequences with very low frequencies convey a meaning
which is usually expressed by a more familiar phrase. Apart from
several such comments, the test was described by the participants
in the pilot study as easy and quick to complete. The final test can
be found in Appendix 1.
190 Chapter 8. Investigating article use by advanced Polish learners of EFL…

Procedure

The participants took the test on a voluntary basis, in a classroom set-

ting. They were encouraged to take the test to see how well they could
use articles. The tests were coded with random three-digit combina-
tions. Each participant was asked to write down their test number so
that they would be able to retrieve their results in the future. The par-
ticipants were asked to put in the missing articles in the right places
and were encouraged to follow their intuition about what sounds
right. A time limit of 15 minutes was enforced, but the test took only
around ten minutes for most participants.

8.3.3. Results and discussion

The data were compiled in the form of dichotomous scores, with

a score of 1 indicating the use of the definite article in the right
place in the target item, and a score of 0 indicating the absence of
the article or the insertion of an indefinite article. Given the nature
of the test, there were no missing data, and a data matrix of 6,000
scores (items X participants) was analysed. Mean scores for all the
participants and for the two categories of items – for frequent and
rare uses – are presented in Table 6. A t-test was performed for each
comparison, showing in each case that the scores obtained for fre-
quent bundles are significantly higher than those for low-frequency
bundles.
8.3. Study 2 191

frequent sequences low-frequency t-test:

(bundles) sequences t value
All participants
0.73 0.46 –22.21***
(n = 200)
Group 1
(less proficient, 0.63 0.32 –16.14***
level B2, n = 100)
Group 2
(more proficient, 0.83 0.59 –11.96***
level C1, n = 100)
***= p < 0.001
Table 6. Mean items scores for frequent and rare uses.

For all the participants taken together, the mean score for the articles in
the frequent bundles is 0.73, while the score for the use of articles in the
rare combinations of words is 0.43. When the groups are analysed sepa-
rately, a similar pattern can be observed in each case, with the success rate
being higher for bundles than for rare word combinations. However, the
gap between the scores for the frequent and the rare category is larger in
the case of the less proficient group. In fact, for the less proficient learners,
the score on frequent bundles was twice as high as that on the rare ones.
Table 7 presents the mean item scores for individual test items. As
can be seen, while there is a significant difference in the means for both
categories, this does not mean that all frequent items were characterised
by more successful article use than all the rare items. Most notably, the
frequent bundles at the end of, on the part of, and in the form of were sup-
plied with the correct article less often than other frequent bundles. By
far the greatest success at providing the correct article was witnessed in
the case of the bundle for the rest of. Among the rare items, for the last
lap scored very high for a rare word combination, with the average score
of 0.68. One possible explanation of the high rate of success of this item
is that the combination for the last lap, rare as a 4-gram, contains a fre-
quent 3-gram, for the last, with the article enclosed in the middle position.
 frequent sequences (bundles) low-frequency sequences
Group 1 Group 2 mean Group 1 Group 2 mean
1. in the united states 0.82 0.89 0.86 in the kuril islands 0.45 0.53 0.49
2. at the same time 0.66 0.86 0.76 at the right level 0.53 0.68 0.61
3. at the end of 0.44 0.68 0.56 at the closure of 0.19 0.43 0.31
4. for the first time 0.76 0.91 0.84 for the last lap 0.50 0.85 0.68
5. in the middle of 0.82 0.95 0.89 in the nexus of 0.35 0.60 0.48
6. on the other side 0.51 0.72 0.62 on the far edge 0.25 0.54 0.40
7. on the basis of 0.48 0.87 0.68 on the merit of 0.46 0.74 0.60
8. in the form of 0.25 0.43 0.34 in the phrasing of 0.15 0.31 0.23
9. for the rest of 0.93 0.94 0.94 for the width of 0.36 0.82 0.59
10. at the top of 0.85 0.91 0.88 at the backside of 0.40 0.64 0.52
11. in the middle east 0.89 0.81 0.85 in the distant north 0.36 0.64 0.50
12. on the part of 0.20 0.76 0.48 on the factor of 0.30 0.68 0.49
13. at the time of 0.72 0.96 0.84 at the period of 0.25 0.61 0.43
14. at the beginning of 0.72 0.89 0.81 at the dawning of 0.20 0.42 0.31
15. in the context of 0.47 0.88 0.68 in the chance of 0.10 0.38 0.24
Table 7. Mean items scores for individual test items.
8.3. Study 2 193

When the frequencies of all items were compared to the average score
for each item, it was found that there is a correlation of 0.53 (p = 0.01)
between the two. The existence of such a correlation further confirms
that there seems to be a tendency for the rate of success in article use
to correspond to the frequency of the bundles; this, however, cannot
be stated with much confidence given the weak correlation which was
obtained. The fact that this correlation is weak was to be expected,
and even a lack of a correlation in this case would be entirely under-
standable. The research instrument in this study was constructed in
such a way that there was a major difference – a difference of three or
four orders of magnitude – between the frequencies of the two cate-
gories of items – the rare and the frequent ones. It was because of this
major difference that one would expect any difference between article
use in the rare and frequent items to show. However, given the some-
times very small difference in the frequency of occurrence found
within each of the two groups of items, especially for the frequent
bundles, we could not realistically expect the same within-group
differences in the degree of success in article use. For example, if
the combination on the merit of had a frequency of 0.017, and in the
phrasing of had a frequency of 0.015, it would be unrealistic to expect
the rate of success in providing the correct article among the partic-
ipants to mirror this difference in frequency. For this to happen, not
only would the phraseological effect postulated here have to be very
strong, there would have to be no other factors affecting the learn-
ers’ performance on the test, and – most importantly – the items’
frequencies in the corpus used in test construction would have to be
identical to the frequencies of those items in the input the learners
had been exposed to. This brings us to a certain general limitation
to all studies which compare learners’ performance against corpus
data, namely, that the frequencies in the corpus never match those
in the language input of any individual person. This limitation is
widely acknowledged. In the words of Michael Hoey, “the personal
194 Chapter 8. Investigating article use by advanced Polish learners of EFL…

‘corpus’ that provides a language user with their lexical primings is

by definition irretrievable, unstudiable and unique” (Hoey, 2005,
p. 14). Therefore, all studies which use corpus data assume that the
corpus provides information on which words are “likely to be rep-
resentative of the types of input speakers are likely to have encoun-
tered” (Durrant & Doherty, 2010, p. 127).
The difference between the less proficient and the more profi-
cient group suggests that the less advanced learners rely more on
the phraseological “help” in article use which they get in the case
of word combinations which may to some extent be remembered in
their entirety, whereas they are less successful in applying the rules
of article use in new contexts. This latter observation is entirely to
be expected, since the success in using any language feature should
normally grow with the level of advancement. Also, when we con-
sider that if fully proficient, native-like users of English would pro-
vide all the correct articles for all the obligatory occasions, their
score on a test like the one used in this study would be 100% for
both the rare and the frequent cases. Therefore, at the ideal fully
proficient level, one would expect the scores to be identical (100%)
for both categories. As the learners approach this level, we should
expect the gap between the rare and the frequent uses to narrow
down.
There is also one more issue which deserves some thought.
This study assumes that there is some formulaicity-related effect
which helps learners of English choose the correct article in regu-
lar, rule-governed contexts. However, we do not know, and we are
in no position to make any speculations about, the nature of this
phenomenon. We know that fully proficient speakers use articles
correctly. But we do not know what processes help with this use.
Maybe the same formulaicity-related effect contributes to article
use in general, in all language users, while at the same time being
of a different nature at different stages of learning. It is theoretically
8.3. Study 2 195

possible to display the same kind of performance through the work-

ing of different mechanisms.
The participants of this study were native speakers of Polish, a lan-
guage which does not have a system of articles. It is likely that similar
findings would be obtained with speakers of other article-less lan-
guages, but more studies with learners of other L1 backgrounds are
needed to confirm this.
Conclusion 197

Conclusion

This book introduces and explores a phraseological perspective on

the use of articles in L2 English. It argues that the correct use of arti-
cles in English by non-native speakers is aided by cognitive processes
related to the formulaic character of language processing. However,
the dominant view during the last century was that articles are pri-
marily an element of the morphosyntax of English, and that the suc-
cessful learning of articles consists of mastering relevant morphosyn-
tactic rules. As a result, pedagogical grammars and teaching materials
traditionally reflect the misconception that a limited set of rules will
lead to using articles correctly in English, with the exception of some
idiomatic or fixed uses, which have to be memorized. The first chap-
ters of this book give an overview of how articles are conceptualized
and their usage accounted for within various descriptive frameworks
and theoretical approaches. This overview not only brings home the
extent of the complexity of the principles that underlie the use of
articles, but it also makes it clear that articles to a great extent elude
description and explanation. Most importantly, “rules” turn out to be
at best collections of regularities. Attempts at providing a definitive
set of rules governing article use fail to account for the wide array of
possible uses in written and spoken English. This means that the task
facing L2 learners is indeed challenging.
198 Conclusion

After providing this broad overview of article use in English, this

book presents two studies conducted on the topic of formulaicity and
article use. In the first study, Polish ESL learners were consistently
more successful at using articles when they appear in word combi-
nations that are frequent, which can be interpreted as a sign of the
idiom principle at work. The two main observations that emerged
from Study 1 are that the “phraseological effect” is more apparent in
less advanced learners of English, and that articles remain an area of
difficulty even at advanced levels of English proficiency, as was pre-
viously found in research.
The results obtained in the second study also show that learners
are generally more likely to use the definite article correctly if the
obligatory occasion for its use occurs inside a lexical bundle than
if the context for its use is a combination of words which do not
frequently co-occur. The findings thus suggest that the correct use
of articles by ESL learners may be facilitated by phraseological ties
between words.
The data from both studies allow for the tentative conclusion that
phraseological effects contribute to successful article use even in
those contexts when the use of articles is consistent with the patterns
of regularities normally observed in English. This in turn implies
that the widespread perception of the process of learning to use arti-
cles in ESL correctly as being a process of rule application needs to
be broadened to acknowledge that phraseological effects contribute
to mastering the use of articles. It must be stressed that the studies
presented here are explorative and are only an initial investigation
of the topic. Further research is needed to explore the nature and
effect of collocations on article use, and how those evolve over the
course of L2 development.
Understandably, the extent of the research presented here is not
sufficient to make firm recommendations concerning the teaching of
articles. However, if their findings are supported by future studies,
Conclusion 199

a number of teaching techniques could be recommended for intro-

duction into ESL pedagogy which would encourage the “colloca-
tional” learning of articles. Finally, because the available literature
so far has shown that learners of “articled” and article-less languages
seem to fall into two distinct groups with respect to article use, it is
reasonable to assume that findings on the use of articles by Polish
learners of L2 English may be generalizable to other article-less lan-
guages. Research with such populations is thus warranted.
References

Acton, E. K. (2014). Pragmatics and the social meaning of determiners (doctoral

dissertation). Stanford University.
Amuzie, G. L., & Spinner, P. (2013). Korean EFL learners’ indefinite article use
with four types of abstract nouns. Applied Linguistics, 34(4), 415–434. https://
doi.org/10.1093/applin/ams065
Andersen, R. W. (1984). The one to one principle of interlanguage construction.
Language Learning, 34(4), 77–95.
Andringa, S., & Rebuschat, P. (2015). New directions in the study of implicit and
explicit learning. Studies in Second Language Acquisition, 37(02), 185–196.
https://doi.org/10.1017/S027226311500008X
Arabski, J. (1968). A linguistic analysis of English composition errors made by
Polish students. Studia Anglica Posnaniensia, 1, 71–89.
Arabski, J. (1990). The acquisition of articles and hierarchy of difficulty. In
J. Arabski (Ed.), Foreign language acquisition papers (pp. 11–17). Katowice:
Uniwersytet Śląski.
Arabski, J., & Wojtaszek, A. (2016). Contemporary perspectives on crosslinguis-
tic influence. In R. Alonso Alonso (Ed.), Crosslinguistic influence in second
language acquisition (pp. 215–224). Clevedon: Multilingual Matters.
Ariel, M. (1990). Accessing noun-phrase antecedents. London: Routledge.
Arnon, I., & Snider, N. (2010). More than words: Frequency effects for multi-word
phrases. Journal of Memory and Language, 62, 67–82. https://doi.org/10.1016/j.
jml.2009.09.005
Bahns, J., & Eldaw, M. (1993). Should we teach EFL students collocations? Sys-
tem, 21(1), 101–114.
202 References

Benson, M., Benson, E., & Ilson, R. (1986). The BBI combinatory dictionary of
English. Amsterdam: John Benjamins.
Berezowski, L. (2009). The myth of the zero article. London, New York: Continuum.
Biber, D. (2009). A corpus-driven approach to formulaic language in English:
Multi-word patterns in speech and writing. International Journal of Corpus
Linguistics, 14(3), 275–311. https://doi.org/10.1075/ijcl.14.3.08bib
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman
grammar of spoken and written English. Harlow: Pearson.
Bickerton, D. (1981). Roots of language. Ann Arbor, MI: Karoma.
Bielak, J. (2011). Cognitive linguistics and foreign language pedagogy: An over-
view of recent trends and developments. In M. Pawlak (Ed.), Extending the
boundaries of research on second language learning and teaching (pp. 241–
262). Heidelberg: Springer.
Birner, B., & Ward, G. (1994). Uniqueness, familiarity, and the definite article
in English. In Proceedings of the Twentieth Annual Meeting of the Berkeley
Linguistics Society (pp. 93–102). Berkeley, CA: Berkeley Linguistics Society.
Bitchener, J., & Knoch, U. (2010). Raising the linguistic accuracy level of advanced
L2 writers with written corrective feedback. Journal of Second Language Writ-
ing, 19(4), 207–217. https://doi.org/10.1016/j.jslw.2010.10.002
Bod, R., Hay, J., & Jannedy, S. (Eds.). (2003). Probabilistic linguistics. Cambridge,
London: MIT Press.
Boers, F., & Lindstromberg, S. (2012). Experimental and intervention studies on
formulaic sequences in a second language. Annual Review of Applied Linguis-
tics, 32, 83–110. https://doi.org/10.1017/S0267190512000050
Bolander, M. (1989). Prefabs, patterns and rules in interaction? Formulaic speech
in adult learners’ L2 Swedish. In K. Hyltenstam & L. K. Obler (Eds.), Bilin-
gualism across the lifespan: Aspects of acquisition, maturity and loss (pp.
73–86). Cambridge: Cambridge University Press.
Brinton, L. J., & Brinton, D. M. (2010). The linguistic structure of modern English.
Amsterdam, Philadelphia: John Benjamins.
Brown, R. (1973). A first language: The early stages. Cambridge, MA: Harvard
University Press.
Butler, Y. G. (2002). Second language learners’ theories on the use of English
articles. Studies in Second Language Acquisition, 24(03), 451–480. https://doi.
org/10.1017/S0272263102003042
Bybee, J. (2006). From usage to grammar: The mind’s response to repetition.
Language, 82(4), 711–733.
References 203

Bybee, J. (2007). Frequency of use and the organization of language. Oxford:

Oxford University Press.
Bybee, J., & Hopper, P. (Eds.). (2001). Frequency and the emergence of linguistic
structure. Amsterdam: John Benjamins.
Carroll, J. (1981). Twenty-five years of research on foreign language aptitude. In
K. C. Diller (Ed.), Individual differences and universals in language learning
aptitude (pp. 83–117). Rowley, MA: Newbury House.
Carter, R., & McCarthy, M. (2006). Cambridge grammar of English. Cambridge:
Cambridge University Press.
Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology,
4, 55–81.
Chater, N., & Manning, C. D. (2006). Probabilistic models of language process-
ing and acquisition. Trends in Cognitive Science, 10(7), 335–344. https://doi.
org/10.1016/j.tics.2006.05.006
Chaudron, C., & Parker, K. (1990). Discourse markedness and structural mark-
edness: The acquisition of English noun phrases. Studies in Second Language
Acquisition, 12, 43–64.
Chen, Y.-H., & Baker, P. (2010). Lexical bundles in L1 and L2 academic writ-
ing. Language Learning and Technology, 14(2), 30–49. https://doi.org/10.4304/
tpls.2.3.637-641
Chrabaszcz, A., & Jiang, N. (2014). The role of the native language in the use
of the English nongeneric definite article by L2 learners: A cross-linguis-
tic comparison. Second Language Research, 30(3), 351–379. https://doi.
org/10.1177/0267658313493432
Christophersen, P. (1939). The articles: A study of their theory and use in English.
Copenhagen: Munksgaard.
Conklin, K., & Schmitt, N. (2012). The processing of formulaic language.
Annual Review of Applied Linguistics, 32, 45–61. https://doi.org/10.1017/
S0267190512000074
Conrad, S., & Biber, D. (2004). The frequency and use of lexical bundles in
conversation and academic prose. Lexicographica, 20, 56–71. https://doi.
org/10.1515/9783484604674.56
Council of Europe. (2011). Common European framework of reference for lan-
guage: Learning, teaching, assessment. Council of Europe.
Cruse, D. (1986). Lexical semantics. Cambridge: Cambridge University Press.
Cziko, G. (1986). Testing the language bioprogram hypothesis: A review of chil-
dren’s acquisition of articles. Language, 62, 878–898.
204 References

Davies, M. (2004). BYU-BNC: Based on the British National Corpus from Oxford
University Press. Retrieved from https://www.english-corpora.org/bnc/
Davies, M. (2008). The Corpus of Contemporary American English (COCA).
Retrieved from https://www.english-corpora.org/coca/
Dechert, H. W. (1983). How a story is done in a second language. In C. Faerch
& G. Kasper (Eds.), Strategies in interlanguage communication (pp. 175–195).
London: Longman.
DeKeyser, R. M. (2003). Implicit and explicit learning. In C. J. Doughty & M. H. Long
(Eds.), The handbook of second language acquisition (pp. 313–348). Malden:
Blackwell.
DeKeyser, R. M. (2005). What makes learning second-language grammar diffi-
cult? A review of issues. Language Learning, 55(S1), 1–25.
DeKeyser, R. M. (2016). Of moving targets and chameleons: Why the concept
of difficulty is so hard to pin down. Studies in Second Language Acquisition,
38(2), 353–363. https://doi.org/10.1017/S0272263116000024
Deprez, V., Sleeman, P., & Guella, H. (2011). Specificity effects in L2 determiner
acquisition: UG or pragmatic egocentrism? In M. Pirvulescu (Ed.), Selected
proceedings of the 4th Conference on Generative Approaches to Language
Acquisition North America (GALANA 2010) (pp. 27–36). Somerville, MA:
Cascadilla Proceedings Project.
Deuter, M., Greenan, J., Noble, J., & Phillips, J. (Eds.). (2002). Oxford collocations
dictionary for students of English. Oxford: Oxford University Press.
Dienes, Z., & Perner, J. (2002). The metacognitive implications of the implic-
it-explicit distinction. In P. Chambres, M. Izaute, & P. J. Marescaux
(Eds.), Metacognition (pp. 171–189). Boston, MA: Springer US. https://doi.
org/10.1007/978-1-4615-1099-4_12
Dietz, G. (2002). On rule complexity: A structural approach. EuroSLA Yearbook,
2, 263–286.
Díez-Bedmar, M. B., & Papp, S. (2008). The use of the English article system by
Chinese and Spanish learners. In G. Gilquin, S. Papp, & M. B. Díez-Bed-
mar (Eds.), Linking up contrastive and learner corpus research (pp. 147–175).
Amsterdam: Rodopi.
Donnellan, K. S. (1968). Putting Humpty Dumpty together again. Philosophical
Review, 77, 203–215.
Dörnyei, Z. (2009). The psychology of second language learning. Oxford: Oxford
University Press.
References 205

Dörnyei, Z., & Skehan, P. (2003). Individual differences in second language learn-
ing. In C. J. Doughty & M. H. Long (Eds.), The handbook of second language
acquisition (pp. 589 – 630). Oxford: Blackwell.
Douglas Kozłowska, C., & Dzierżanowska, H. (1988). Selected English collocations.
Warsaw: Wydawnictwo Naukowe PWN.
Downing, A., & Locke, P. (1992). A university course in English grammar. Phoe-
nix: Prentice Hall.
Durrant, P., & Doherty, A. (2010). Are high-frequency collocations psychologi-
cally real? Investigating the thesis of collocational priming. Corpus Linguis-
tics and Linguistic Theory, 6(2), 125–155. https://doi.org/10.1515/cllt.2010.006
Durrant, P., & Schmitt, N. (2009). To what extent do native and non-native writ-
ers make use of collocations? IRAL - International Review of Applied Linguis-
tics in Language Teaching, 47(2), 157–177. https://doi.org/10.1515/iral.2009.007
Durrant, P., & Schmitt, N. (2010). Adult learners’ retention of collocations
from exposure. Second Language Research, 26(2), 163–188. https://doi.
org/10.1177/0267658309349431
Dušková, L. (1969). On sources of errors in foreign language learning. IRAL
- International Review of Applied Linguistics in Language Teaching, 7(1), 11–36.
https://doi.org/10.1515/iral.1969.7.1.11
Ekiert, M. (2004). Acquisition of the English article system by speakers of Pol-
ish in ESL and EFL Settings. Teachers College, Columbia University Working
Papers in TESOL & Applied Linguistic, 4(1), 1–23.
Ekiert, M. (2007). The acquisition of grammatical marking of indefiniteness with
the indefinite article a in L2 English. Columbia University Working Papers in
TESOL & Applied Linguistics, 7(1), 1–43.
Ekiert, M., & Han, Z. (2017). L1-fraught difficulty: The case of L2 acquisition of
English Articles by Slavic Speakers. In R. Alonso (Ed.), Crosslinguistic influ-
ence in second language acquisition (pp. 147–172). Clevedon: Multilingual
Matters. https://doi.org/10.1007/978-3-319-49959-8
Ellis, N. C. (2003). Constructions, chunking, and connectionism: The emer-
gence of second language structure. In C. J. Doughty & M. H. Long (Eds.),
The handbook of second language acquisition (pp. 63–103). Malden, Oxford:
Blackwell.
Ellis, N. C. (2005). At the interface: Dynamic interactions of explicit and implicit
language knowledge. Studies in Second Language Acquisition, 27(2), 305–352.
https://doi.org/10.1017/S027226310505014X
206 References

Ellis, N. C. (2006). Selective attention and transfer phenomena in L2 acquisition:

Contingency, cue competition, salience, interference, overshadowing, block-
ing, and perceptual learning. Applied Linguistics, 27(2), 164–194. https://doi.
org/10.1093/applin/aml015
Ellis, N. C. (2015). Implicit AND explicit language learning: Their dynamic inter-
face and complexity. In P. Rebuschat (Ed.), Implicit and explicit learning of
languages (pp. 3–24). Amsterdam, Philadelphia: John Benjamins.
Ellis, N. C. (2016). Salience, cognition, language complexity, and complex adap-
tive systems. Studies in Second Language Acquisition, 38(2), 341–351. https://
doi.org/10.1017/S027226311600005X
Ellis, R. (2006). Current issues in the teaching of grammar: An SLA perspective.
TESOL Quarterly, 40(1), 83–107.
Ellis, R. (2008). Investigating grammatical difficulty in second language learning:
Implications for second language acquisition research and language testing.
International Journal of Applied Linguistics, 18(1), 4–22. https://doi.org/10.1111/
j.1473-4192.2008.00184.x
Epstein, R. (2002). The definite article, accessibility, and the construction of dis-
course referents. Cognitive Linguistics, 12(4), 333–378.
Erman, B., & Warren, B. (2000). The idiom principle and the open choice prin-
ciple. Text – Interdisciplinary Journal for the Study of Discourse, 20(1), 29–62.
https://doi.org/­10.1515/text.1.2000.20.1.29
Evans, V. (2007). A glossary of cognitive linguistics. Edinburgh: Edinburgh Uni-
versity Press.
Fauconnier, G. (1994). Mental spaces: Aspects of meaning construction in natural
language. Cambridge: Cambridge University Press.
Firbas, J. (1992). Functional sentence perspective in written and spoken communi-
cation. Cambridge: Cambridge University Press.
Gablasova, D., Brezina, V., & McEnery, T. (2017). Collocations in corpus-based
language learning research: Identifying, comparing, and interpreting the
evidence. Language Learning, 67, 155–179. https://doi.org/10.1111/lang.12225
Gass, S. M., & Selinker, L. (2008). Second language acquisition: An introductory
course (3rd edition). New York, London: Routledge.
Gass, S. M., Spinner, P., & Behney, J. (Eds.). (2018). Salience in second language
acquisition. New York: Routledge.
Gillian, E. J. (2017). Gil’s Article Maker. An interactive tool for teaching English
article use to Polish gimnazjum students: Development and evaluation (doc-
toral dissertation). Wydział Anglistyki Uniwersytetu Adama Mickiewicza.
References 207

Głaz, A. (2012). Extended vantage theory in linguistic application: The case of the
English articles. Lublin: Wydawnictwo Uniwersytetu Marii Curie-Skłodowskiej.
Gobet, F., Lane, P. C. R., Croker, S., Cheng, P. C. H., Jones, G., Oliver, I., & Pine,
J. M. (2001). Chunking mechanisms in human learning. Trends in Cognitive
Sciences, 5(6), 236–243. https://doi.org/10.1016/S1364-6613(00)01662-4
Godfroid, A. (2016). The effects of implicit instruction on implicit and explicit
knowledge development. Studies in Second Language Acquisition, 38(2), 177–
215. https://doi.org/10.1017/S0272263115000388
Goldberg, A. E. (2006). Constructions at work: The nature of generalization in
language. Oxford: Oxford University Press.
Goldschneider, J. M., & DeKeyser, R. M. (2005). Explaining the “natural order
of L2 morpheme acquisition” in English: A meta‐analysis of multiple deter-
minants. Language Learning, 55(1), 27–77.
Gozdawa-Gołębiowski, R. (2003). Interlanguage formation: A study of the trig-
gering mechanisms. Warsaw: Instytut Anglistyki UW.
Granfeldt, J. (2000). The acquisition of the Determiner Phrase in bilingual and
second-language French. Bilingualism: Language and Cognition, 3(3), 263–280.
Granger, S. (1998). Learner English on computer. London: Routledge.
Granger, S., & Meunier, F. (2008). The many faces of phraseology. In S. Granger
& F. Meunier (Eds.), Phraseology: An interdisciplinary perspective (pp. xix–
xxviii). Amsterdam, Philadelphia: John Benjamins.
Greenbaum, S. (1991). An introduction to English Grammar. Harlow: Longman.
Grice, P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax
and semantics (Vol. 3, pp. 41–58). New York: Academic Press.
Gries, S. T., & Ellis, N. C. (2015). Statistical measures for usage-based linguistics.
Language Learning, 65(S1), 228–255. https://doi.org/10.1111/lang.12119
Hakuta, K. (1976). A case study of a Japanese child learning English. Language
Learning, 26(2), 321–351. https://doi.org/10.1111/j.1467-1770.1976.tb00280.x
Halliday, M. A. K. (1961). Categories of the theory of grammar. WORD, 17(2),
241–292. https://doi.org/10.1080/00437956.1961.11659756
Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. London: Longman.
Hallin, A. E., & Van Lancker Sidtis, D. (2017). A closer look at formulaic lan-
guage: Prosodic characteristics of Swedish proverbs. Applied Linguistics, 38(1),
68–89. https://doi.org/10.1093/applin/amu078
Hamrick, P. (2014). A role for chunk formation in statistical learning of second
language syntax. Language Learning, 64(2), 247–278. https://doi.org/10.1111/
lang.12049
208 References

Hawkins, J. A. (1978). Definiteness and indefiniteness: A study in reference and

grammaticality prediction. London: Croom Helm.
Hawkins, J. A. (1984). A note on referent identifiability and co-presence. Journal
of Pragmatics, 8(5–6), 649–659.
Hawkins, J. A. (1991). On (in)definite articles: Implicatures and (un)grammati-
cality prediction. Journal of Linguistics, 27(2), 405–442.
Hawkins, R. (2001). Second language syntax: A generative introduction. Oxford:
Blackwell.
Hawkins, R., Al-Eid, S., Almahboob, I., Athanasopoulos, P., Chaengchenkit, R.,
Hu, J., … Velasco-Zarate, K. (2006). Accounting for English article interpre-
tation by L2 speakers. EUROSLA Yearbook, 6, 7–25.
Hewson, J. (1964). Article and noun in English: An essay in psychomechanical
analysis. Ste-Foy, Québec: Université Laval.
Hiki, M. (1991). A study of learners’ judgment on noun countability (doctoral
dissertation). Indiana University.
Hinenoya, K., & Lyster, R. (2015). Identifiability and accessibility in learning
definite article usages: A quasi-experimental study with Japanese learn-
ers of English. Language Teaching Research, 19(4), 397–415. https://doi.
org/10.1177/1362168814541742
Hoey, M. (2005). Lexical priming: A new theory of words and language. London:
Routledge.
Holme, R. (2007). Cognitive linguistics and language teaching. New York: Pal-
grave Macmillan.
Holmes, J. (1988). Doubt and certainty in ESL textbooks. Applied Linguistics,
9(1), 21–44.
Housen, A., & Simoens, H. (2016). Cognitive perspectives on difficulty and com-
plexity in L2 acquisition. Studies in Second Language Acquisition, 38(2), 163–
175. https://doi.org/10.1017/S0272263116000176
Howarth, P. A. (1996). Phraseology in English academic writing: Some implications
for language learning and dictionary making. Berlin: De Gruyter.
Hsu, J. T. (2008). Role of the multiword lexical units in current EFL ESL text-
books. US-China Foreign Language, 6(7), 27–39.
Hua, D., & Lee, T. H.-t. (2005). Chinese ESL learners’ understanding of the
English count-mass distinction. In R. Dekydtspotter, A. Sprouse, & A. Lil-
jestrand (Eds.), Proceedings of the 7th Generative Approaches to Second Lan-
guage Acquisition Conference (pp. 139–149). Somerville, MA: Cascadilla Pro-
ceedings Project.
References 209

Huddleston, R., & Pullum, G. K. (2002). The Cambridge grammar of the English
language. Cambridge, New York, Melbourne: Cambridge University Press.
Hudson, R. (2004). Are determiners heads? Functions of Language, 11(1), 7–42.
Huebner, T. (1983). A longitudinal analysis of the acquisition of English. Ann Arbor,
MI: Karoma.
Huebner, T. (1985). System and variability in interlanguage syntax. Language
Learning, 35(2), 141–163.
Hulstijn, J. H. (2005). Theoretical and empirical issues in the study of implicit
and explicit second-language learning. Studies in Second Language Acqui-
sition, 27, 129–140.
Hyland, K. (2012). Bundles in academic discourse. Annual Review of Applied Lin-
guistics, 32, 150–169. https://doi.org/10.1017/S0267190512000037
Ionin, T., Ko, H., & Wexler, K. (2004). Article semantics in L2-acquisition: the
role of specificity. Language Acquisition, 12, 3–69.
Ionin, T., Zubizarreta, M. L., & Maldonado, S. B. (2008). Sources of linguistic
knowledge in the second language acquisition of English articles. Lingua,
118(4), 554–576. https://doi.org/10.1016/j.lingua.2006.11.012
Jarvis, S. (2002). Topic continuity in L2 English article use. Studies in Second
Language Acquisition, 24(3), 387–418. https://doi.org/10.1017/S02722631020
03029
Jiang, N., & Nekrasova, T. M. (2007). The processing of formulaic sequences by
second language speakers. The Modern Language Journal, 91(3), 433–445.
https://doi.org/10.1111/j.1540-4781.2007.00589.x
Jones, S., & Sinclair, J. M. (1974). English lexical collocations. Cahiers de Lexicol-
ogie, 24, 15–61.
Kaltenbacher, M. (2001). Universal Grammar and parameter resetting in second
language acquisition. Frankfurt: Peter Lang.
Kałuża, H. (1963). Teaching the English articles to speakers of Slavic. Language
Learning, 13(2), 113–124. https://doi.org/10.1111/j.1467-1770.1963.tb01290.x
Kambarov, Z. (2008). The concept of definiteness and its application to automated
reference resolution. New York: Peter Lang.
Kharma, N. (1981). Analysis of the errors committed by Arab university students
in the use of the English definite/indefinite articles. International Review of
Applied Linguistics, 19, 331–345.
Krashen, S. D. (1985). The Input Hypothesis. London and New York: Longman.
Kreidler, C. W. (1998). Introducing English semantics. London: Routledge.
210 References

Król, A. (2005). Applied cognitive linguistics in teaching English articles to Polish

learners. Essen: University of Duisburg.
Król-Markefka, A. (2008). Some theoretical considerations on the use of contras-
tive data in teaching English articles to Polish learners. Studia Linguistica
Universitatis Iagellonicae Cracoviensis, 125, 103–112.
Król-Markefka, A. (2010). Metalinguistic knowledge and the accurate use of
En­glish articles: Effects of applying cognitive grammar in second language
teaching (doctoral dissertation). Jagiellonian University.
Król-Markefka, A. (2012). Pedagogical rules for the use of English articles: An
evaluation and suggestions for improvement. Studia Linguistica Universi-
tatis Iagellonicae Cracoviensis, 129, 98–115. https://doi.org/10.4467/2083462
4SL.12.007.0596
Kuiper, K. (2004). Formulaic performance in conventionalized varieties of
speech. In N. Schmitt (Ed.), Formulaic sequences: acquisition, processing and
use (pp. 37–54). Amsterdam: John Benjamins.
Langacker, R. W. (1987). Foundations of Cognitive Grammar. Stanford: Stanford
University Press.
Langacker, R. W. (2000). A dynamic usage-based model. Stanford: CSLI.
Larsen-Freeman, D. (1975). The acquisition of grammatical morphemes by adult
ESL students. Tesol Quarterly, 9(4), 409–419. https://doi.org/10.2307/3585625
Laufer, B. (1997). What’s in a word that makes it hard or easy? Intralexical factors
affecting the difficulty of vocabulary acquisition. In N. Schmitt & M. McCa-
rthy (Eds.), Vocabulary: Description, acquisition and pedagogy (pp. 140–155).
Cambridge: Cambridge University Press.
Laufer, B., & Waldman, T. (2011). Verb-noun collocations in second language
writing: A corpus analysis of learners’ English. Language Learning, 61(2),
647–672. https://doi.org/10.1111/j.1467-9922.2010.00621.x
Leńko-Szymańska, A. (2012). The role of conventionalized language in the acqui-
sition and use of articles by Polish EFL learners. In Y. Tono, Y. Kawagu-
chi, & M. Minegishi (Eds.), Developmental and crosslinguistic perspectives
in learner corpus research (pp. 83–104). Amsterdam, Philadelphia: John Ben-
jamins.
Leow, R. P. (2015). Explicit learning in the L2 classroom: A student-centered
approach. New York: Routledge.
Leśniewska, J. (2003). The collocational aspects of advanced EFL learners’ lexical
competence (doctoral dissertation). Jagiellonian University.
References 211

Leśniewska, J. (2016). The use of articles in L2 English: A phraseological per-

spective. Studies in Second Language Learning and Teaching, 6(2), 203–224.
https://doi.org/10.14746/ssllt.2016.6.2.2
Leśniewska, J. (2017). New perspectives on the use of articles in L2 English. In
L. Torres-Zúñiga & T. H. Schmidt (Eds.), New methodological approaches
to foreign language teaching (pp. 67–82). Newcastle upon Tyne: Cambridge
Scholars Publishing.
Li, H., & Yang, L. (2010). An investigation of English articles’ acquisition by Chi-
nese learners of English. Chinese Journal of Applied Linguistics, 33(3), 15–31.
Lightbown, P. M., & Spada, N. (2006). How languages are learned (5th edition).
Oxford: Oxford University Press.
Lin, P. M. S. (2012). Sound evidence: The missing piece of the jigsaw in formulaic
language research. Applied Linguistics, 33(3), 342–347. https://doi.org/10.1093/
applin/ams017
Lin, P. M. S., & Adolphs, S. (2009). Sound evidence: Phraseological units in spoken
corpora. In A. Barfield & H. Gyllstad (Eds.), Researching collocations in another
language: Multiple interpretations (pp. 34–48). London: Palgrave Macmillan.
Littlemore, J. (2001). Applying cognitive linguistics to second language learning
and teaching. New York: Palgrave Macmillan.
Liu, D., & Gleason, J. (2002). Acquisition of the article the by nonnative speakers
of English: An analysis of four nongeneric uses. Studies in Second Language
Acquisition, 24, 1–26.
Low, R. M. H. (2005). The phenomenon of the word the in English: Discourse
functions and distribution patterns (doctoral dissertation). State University
of New York.
Lyons, C. (1999). Definiteness. Cambridge: Cambridge University Press.
Lyons, J. (1977). Semantics. Cambridge, London, New York, Melbourne: Cam-
bridge University Press.
Lyons, J. (1995). Linguistic semantics. Cambridge: Cambridge University Press.
MacLaury, R. E. (1995). Vantage theory. In R. E. MacLaury & J. R. Taylor (Eds.),
Language and the cognitive construal of the world (pp. 231–276). Berlin and
New York: Mouton de Gruyter.
MacWhinney, B. (2006). Emergentism – use often and with care. Applied Lin-
guistics, 27(4), 729–740. https://doi.org/10.1093/applin/aml035
Maratsos, M. (1976). The use of definite and indefinite reference in young children:
An experimental study of semantic acquisition. Cambridge: Cambridge Uni-
versity Press.
212 References

Master, P. (1997). The English article system: Acquisition, function, and peda-
gogy. System, 25(2), 215–232. https://doi.org/10.1016/S0346-251X(97)00010-9
Master, P. (2002). Information structure and English article pedagogy. System,
30(3), 331–348. https://doi.org/10.1016/S0346-251X(02)00018-0
Meunier, F. (2012). Formulaic language and language teaching. Annual Review
of Applied Linguistics, 32, 111–129. https://doi.org/10.1017/S0267190512000128
Miller, J. (2005). Most of ESL students have trouble with the articles. International
Education Journal, 5(5), 80–88. https://doi.org/10.1016/S1351-4180(14)70140-0
Mitchell, B., & Robinson, F. C. (1992). A guide to Old English (5th edition). Oxford:
Blackwell.
Mueller, C. M. (2011). English learners’ knowledge of prepositions: Collocational
knowledge or knowledge based on meaning? System, 39(4), 480–490. https://
doi.org/10.1016/j.system.2011.10.012
Muñoz, C., & Singleton, D. (2011). A critical review of age-related research on L2
ultimate attainment. Language Teaching, 44(1), 1–35. https://doi.org/10.1017/
S0261444810000327
Myles, F., & Cordier, C. (2016). Formulaic sequence (FS) cannot be an umbrella
term in SLA. Studies in Second Language Acquisition, 39(1), 1–26. https://doi.
org/10.1017/S027226311600036X
Myles, F., Hooper, J., & Mitchell, R. (1998). Rote or rule? Exploring the role of
formulaic language in classroom foreign language learning. Language Learn-
ing, 48(3), 323–364.
Nation, I. S. P. (2001). Learning vocabulary in another language. Cambridge: Cam-
bridge University Press.
Nattinger, J., & DeCarrico, J. (1992). Lexical phrases and language teaching.
Oxford: Oxford University Press.
Neale, S. (1990). Descriptions. Cambridge, MA: MIT Press.
Nekrasova, T. M. (2009). English L1 and L2 speakers’ knowledge of lexical
bundles. Language Learning, 59(3), 647–686. https://doi.org/10.1111/j.1467-
9922.2009.00520.x
Nesselhauf, N. (2003). The use of collocations by advanced learners of English and
some implications for teaching. Applied Linguistics, 24(2), 223–242.
Ogawa, M. (2008). The acquisition of English articles by advanced EFL Japanese
learners : Analysis based on noun types. Journal of Language and Culture,
Language and Information, 3, 133–151.
Palmer, F. R. (1981). Semantics (2nd edition). Cambridge, New York, Melbourne:
Cambridge University Press.
References 213

Paradowski, M. B. (2006). Uczyć, aby nauczyć – rola języka ojczystego

w gramatyce pedagogicznej i implikacje dla dydaktyki języków obcych. In
J. Krieger-Knieja & U. Paprocka-Piotrowska (Eds.), Komunikacja językowa
w społeczeństwie informacyjnym – nowe wyzwania dla dydaktyki języków
obcych (pp. 125–144). Lublin: Towarzystwo Naukowe KUL.
Paradowski, M. B. (2008). Introducing language interface in pedagogical gram-
mar. In D. Gabryś-Barker (Ed.), Morphosyntactic issues in second language
acquisition (pp. 225–252). Clevedon: Multilingual Matters.
Park, K. (1996). The article acquisition in English as a foreign language. Retrieved
from https://files.eric.ed.gov/fulltext/ED397647.pdf
Parrish, B. (1987). A new look at methodologies in the study of article acquisition
for learners of ESL. Language Learning, 37(3), 361–383.
Parrott, M. (2000). Grammar for English language teachers. Cambridge: Cam-
bridge University Press.
Pawlak, M. (2006). The place of form-focused instruction in the foreign language
classroom. Kalisz, Poznań: Uniwersytet Adama Mickiewicza.
Pawley, A., & Syder, F. H. (1983). Two puzzles for linguistic theory: Nativelike
selection and nativelike fluency. In J. C. Richards & R. W. Schmidt (Eds.),
Language and communication (pp. 191 – 226). London: Longman.
Peters, A. M. (1977). Language learning strategies: Does the whole equal the sum
of the parts? Language, 53, 560–73.
Peters, A. M. (1983). The units of language acquisition. Cambridge: Cambridge
University Press.
Pichette, F., & Leśniewska, J. (2018). Percentage of L1-based errors in ESL: An
update on Ellis (1985). International Journal of Language Studies, 12(2), 1–16.
Quirk, R., & Greenbaum, S. (1973). A concise grammar of contemporary English.
New York: Harcourt Brace Jovanovich.
Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A comprehensive gram-
mar of the English language. London: Longman.
Raupach, M. (1984). Formulae in second language speech production. In
H. Dechert, D. Möhle, & M. Raupach (Eds.), Second language productions
(pp. 114–137). Tübingen: Gunter Narr.
Rebuschat, P., & Williams, J. N. (2012). Statistical learning and language acqui-
sition. Berlin: Mouton de Gruyter.
Ringbom, H. (1985). The influence of Swedish on the English of Finnish learn-
ers. In Foreign language learning and bilingualism (pp. 39–71). Abo: Abo
Akademi.
214 References

Roehr-Brackin, K. (2015). Explicit knowledge about language in L2 learning:

A usage-based perspective. In P. Rebuschat (Ed.), Implicit and explicit learn-
ing of languages (pp. 117–138). Amsterdam, Philadelphia: John Benjamins.
Roehr, K., & Gánem-Gutiérrez, G. A. (2009). Metalinguistic knowledge: A step-
ping stone towards L2 proficiency? In A. Benati (Ed.), Issues in second lan-
guage proficiency (pp. 79–94). London: Continuum.
Romberg, A. R., & Saffran, J. R. (2010). Statistical learning and language acquisi-
tion. Wiley Interdisciplinary Reviews: Cognitive Science, 1(6), 906–914. https://
doi.org/10.1002/wcs.78
Rudzka, B., Channell, J., Putseys, Y., & Ostyn, P. (1981). The words you need. Lon-
don: Macmillan.
Rudzka, B., Channell, J., Putseys, Y., & Ostyn, P. (1985). More words you need.
London: Macmillan.
Rumelhart, D., & McClelland, J. (Eds.). (1986). Parallel distributed processing:
Explorations in the microstructure of cognition. Cambridge: MIT Press.
Russel, B. (1905). On denoting. Mind, 14(56), 479–493. https://doi.org/10.1093/
wbro/lkj004
Saeed, J. I. (2003). Semantics (2nd edition). Malden, MA: Blackwell.
Scheffler, P. (2007). Ćwiczenia tłumaczeniowe w nowoczesnym nauczaniu
języków obcych. Języki Obce w Szkole, (3), 58–62.
Scheffler, P. (2011). Rule difficulty: Teachers’ intuitions and learners’ performance.
Language Awareness, 20(3), 221–237. https://doi.org/10.1080/09658416.2011.570349
Schmitt, N., Grandage, S., & Adolphs, S. (2004). Are corpus-derived recurrent
clusters psycholinguistically valid? In N. Schmitt (Ed.), Formulaic sequences:
Acquisition, processing and use (pp. 127–152). Amsterdam: John Benjamins.
Schwartz, B. D., & Sprouse, R. (1996). L2 cognitive states and the full transfer/
full access model. Second Language Research, 12, 40–72.
Segalowitz, N. (2003). Automaticity and second languages. In C. Doughty
& M. H. Long (Eds.), The handbook of second language acquisition (pp. 382–
408). Malden: Blackwell.
Segalowitz, N. (2005). Automaticity in bilingualism and second language learn-
ing. In F. Kroll & A. M. B. de Groot (Eds.), Handbook of bilingualism: Psy-
cholinguistic approaches (pp. 371–388). New York: Oxford University Press.
Sheen, Y. (2007a). The effect of corrective feedback, language aptitude and learner
attitudes on the acquisition of English articles. In A. Mackey (Ed.), Conver-
sational interaction in second language acquisition (pp. 301–322). Oxford:
Oxford University Press.
References 215

Sheen, Y. (2007b). The effect of focused written corrective feedback and language
aptitude on ESL learners’ acquisition of articles. Tesol Quarterly, 41(2), 255–
283. https://doi.org/10.1002/j.1545-7249.2007.tb00059.x
Shin, Y. K., & Kim, Y. (2017). Using lexical bundles to teach articles to L2 English
learners of different proficiencies. System, 69, 79–91. https://doi.org/10.1016/j.
system.2017.08.002
Shintani, N., & Ellis, R. (2013). The comparative effect of direct written corrective
feedback and metalinguistic explanation on learners’ explicit and implicit
knowledge of the English indefinite article. Journal of Second Language Writ-
ing, 22(3), 286–306. https://doi.org/10.1016/j.jslw.2013.03.011
Shintani, N., Ellis, R., & Suzuki, W. (2014). Effects of written feedback and revi-
sion on learners’ accuracy in using two English grammatical structures. Lan-
guage Learning, 64(1), 103–131. https://doi.org/10.1111/lang.12029
Silva, L. H. R., & Roehr-Brackin, K. (2016). Perceived learning difficulty and
actual performance: Explicit and implicit knowledge of L2 English grammar
points among instructed adult learners. Studies in Second Language Acquisi-
tion, 38(2), 317–340. https://doi.org/10.1017/S0272263115000340
Sinclair, J. (Ed.). (1990). The Collins Cobuild English grammar. London: Harper
Collins.
Sinclair, J. (1991). Corpus, concordance, collocation. Oxford: Oxford University
Press.
Singleton, D. (1999). Exploring the second language mental lexicon. Cambridge:
Cambridge University Press.
Singleton, D. (2000). Language and the lexicon. London, New York: Arnold.
Singleton, D., Leśniewska, J., & Witalisz, E. (2007). Open choice versus the idiom
principle in L2 lexical usage. In M. Pawlak (Ed.), Exploring focus on form in
language teaching (pp. 207–222). Poznań, Kalisz: Adam Mickiewicz Uni-
versity.
Siyanova-Chanturia, A., Conklin, K., & van Heuven, W. J. B. (2011). Seeing
a phrase “time and again” matters: The role of phrasal frequency in the pro-
cessing of multiword sequences. Journal of Experimental Psychology: Learn-
ing Memory and Cognition, 37(3), 776–784. https://doi.org/10.1037/a0022531
Siyanova-Chanturia, A., & Martinez, R. (2015). The idiom principle revisited.
Applied Linguistics, 36(5), 549–569. https://doi.org/10.1093/applin/amt054
Skehan, P. (1998). A cognitive approach to language learning. Oxford, New York:
Oxford University Press.
216 References

Snape, N. (2008). Resetting the Nominal Mapping Parameter in L2 English: Defi-

nite article use and the count–mass distinction. Bilingualism: Language and
Cognition, 11(1), 63–79. https://doi.org/10.1017/S1366728907003215
Sosa, A. V., & MacFarlane, J. (2002). Evidence for frequency-based constituents in
the mental lexicon: Collocations involving the word of. Brain and Language,
83(2), 227–236. https://doi.org/10.1016/S0093-934X(02)00032-9
Spada, N., & Tomita, Y. (2010). Interactions between type of instruction and type
of language feature: A meta-analysis. Language Learning, 60(2), 263–308.
Sperber, D., & Wilson, D. (1995). Relevance: communication and cognition (2nd
edition). Oxford: Blackwell.
Stefanou, C., & Révész, A. (2015). Direct written corrective feedback, learner
differences, and the acquisition of second language article use for generic
and specific plural reference. The Modern Language Journal, 99(2), 263–282.
https://doi.org/10.1111/modl.12212
Stubbs, M. (2009). The search for units of meaning: Sinclair on empirical seman-
tics. Applied Linguistics, 30(1), 115–137.
Sun, G. (2016). The acquisition of English articles by second language learners:
The sequence, differences, and difficulties. SAGE Open, 6(1). https://doi.org/
https://journals.sagepub.com/doi/pdf/10.1177/2158244016635716
Świątek, A. (2013). The acquisition of the English article system by Polish learners
in different proficiency groups juxtaposed with a case study. In E. Piechurs-
ka-Kuciel & E. Szymańska-Czaplak (Eds.), Language in cognition and affect
(pp. 151–170). Berlin, Heidelberg: Springer.
Świątek, A. (2014). The order of the acquisition of the English article system by
Polish learners in different proficiency groups. Newcastle upon Tyne: Cam-
bridge Scholars Publishing.
Szwedek, A. (1973). Some aspects of definiteness and indefiniteness of nouns in
Polish. Papers and Studies in Contrastive Linguistics, 2, 203–212.
Szwedek, A. (1976). Word order, sentence stress and reference in English and Pol-
ish. Edmonton: Linguistic Research.
Tabossi, P., Fanari, R., & Wolf, K. (2009). Why are idioms recognized fast? Mem-
ory and Cognition, 37(4), 529–540. https://doi.org/10.3758/MC.37.4.529
Tarone, E., & Parrish, B. (1988). Task-related variation in interlanguage: The case
of articles. Language Learning, 38(1), 21–44.
Thomas, M. (1989). The acquisition of English articles by first- and second-lan-
guage learners. Applied Psycholinguistics, 10(3), 335–355. https://doi.org/10.1017/
S0142716400008663
References 217

Tomasello, M. (2003). Constructing a language: A usage-based theory of language

acquisition. Cambridge, London: Harvard University Press.
Towell, R., Hawkins, R., & Bazergui, N. (1996). The development of fluency in
advanced learners of French. Applied Linguistics, 17(1), 84–119.
Tremblay, A., & Baayen, R. H. (2010). Holistic processing of regular four-word
sequences: A behavioral and ERP study of the effects of structure, frequency,
and probability on immediate free recall. In D. Wood (Ed.), Perspectives on
formulaic language: Acquisition and communication (pp. 151–173). London:
Continuum International.
Tremblay, A., Derwing, B., Libben, G., & Westbury, C. (2011). Processing advan-
tages of lexical bundles: Evidence from self-paced reading and sentence
recall tasks. Language Learning, 61(2), 569–613. https://doi.org/10.1111/j.1467-
9922.2010.00622.x
Trenkic, D. (2007). Variability in second language article production: Beyond the
representational deficit vs. processing constraints debate. Second Language
Research, 23(3), 289–327. https://doi.org/10.1177/0267658307077643
Trenkic, D. (2008). The representation of English articles in second language
grammars: Determiners or adjectives? Bilingualism: Language and Cogni-
tion, 11(1), 1–18. https://doi.org/10.1017/S1366728907003185
Tyler, A. (2012). Cognitive linguistics and second language learning: Theoretical
basics and experimental evidence. New York, London: Routledge.
Vainikka, A., & Young-Scholten, M. (1996). Gradual development of L2 phrase
structure. Second Language Research, 12, 7–39.
Van Lancker, D., Canter, G., & Terbeek, D. (1981). Disambiguation of ditropic
sentences. Journal of Speech, Language, and Hearing Research, 24(3), 330–335.
https://doi.org/10.1044/jshr.2403.330
Van Lancker, D., & Kempler, D. (1987). Comprehension of familiar phrases by
left- but not by right-hemisphere damaged patients. Brain and Language,
32, 265–277.
Van Lancker-Sidtis, D. (2012). Formulaic language and language disorders.
Annual Review of Applied Linguistics, 32, 62–80. https://doi.org/10.1017/
S0267190512000104
Van Lancker-Sidtis, D., & Postman, W. A. (2006). Formulaic expressions in spon-
taneous speech of left- and right-hemisphere damaged subjects. Aphasiol-
ogy, 20, 411–426.
Van Langendonck, W. (1994). Determiners as heads? Cognitive Linguistics, 5(3),
243–260.
218 References

Verneau, M., van der Kamp, J., Savelsbergh, G. J. P., & de Looze, M. P. (2014).
Age and time effects on implicit and explicit learning. Experimental Aging
Research, 40(4), 477–511.
von Heusinger, K. (n.d.). Definiteness. In Oxford Bibliographies. Oxford Univer-
sity Press. https://doi.org/DOI: 10.1093/OBO/97801997728100063
Warden, D. A. (1976). The influence of context on children’s use of identifying
expressions and references. British Journal of Psychology, 67(1), 101–112.
Weinert, R. (1995). The role of formulaic language in second language acquisi-
tion: A review. Applied Linguistics, 16(2), 180–205. https://doi.org/10.1093/
applin/16.2.180
White, B. (2009). Accounting for L2-English learners’ article choices. MSU Work-
ing Papers in SLS, 1(1), 14–37.
White, B. (2012). A conceptual approach to the instruction of phrasal verbs.
The Modern Language Journal, 96(3), 419–438. https://doi.org/10.1111/j.1540-
4781.2012.01365.x
White, L. (1989). Universal grammar and second language acquisition. Amster-
dam: John Benjamins.
Wierzbicka, A. (1988). The semantics of grammar. Amsterdam, Philadelphia: John
Benjamins.
Wolter, B., & Gyllstad, H. (2013). Frequency of input and L2 collocational pro-
cessing. Studies in Second Language Acquisition, 35(3), 451–482. https://doi.
org/10.1017/S0272263113000107
Wolter, B., & Yamashita, Y. (2018). Word frequency, collocational frequency, L1
congruency, and proficiency in L2 collocational processing: What accounts
for L2 performance? Studies in Second Language Acquisition, 40(2), 395–416.
https://doi.org/10.1017/S0272263117000237
Wong-Fillmore, L. (1976). The second time around: Cognitive and social strategies
in second language acquisition (doctoral dissertation). Stanford University.
Wray, A. (2002a). Formulaic language and the lexicon. Cambridge: Cambridge
University Press.
Wray, A. (2002b). Formulaic language in computer-supported communica-
tion: Theory meets reality. Language Awareness, 11(2), 114–131. https://doi.
org/10.1080/09658410208667050
Wray, A. (2012). What do we (think we) know about formulaic language? An
evaluation of the current state of play. Annual Review of Applied Linguistics,
32, 231–254. https://doi.org/10.1017/S026719051200013X
References 219

Yalçın, Ş., & Spada, N. (2016). Language aptitude and grammatical difficulty.
Studies in Second Language Acquisition, 38(2), 239–263. https://doi.org/10.1017/
S0272263115000509
Yamada, J., & Matsuura, N. (1982). The use of the English article among Japanese stu-
dents. RELC Journal, 13(1), 50–63. https://doi.org/10.1177/003368828201300104
Yamashita, J., & Jiang, N. (2010). L1 influence on the acquisition of L2 colloca-
tions : Japanese ESL users and EFL learners acquiring English collocations.
TESOL Quarterly, 44(4), 647–668.
Yoo, I. W. (2009). The English definite article: What ESL/EFL grammars say and
what corpus findings show. Journal of English for Academic Purposes, 8(4),
267–278. https://doi.org/10.1016/j.jeap.2009.07.004
Yoon, K. K. (1993). Challenging prototype descriptions: Perception of noun
countability and indefinite vs. zero article use. IRAL: International Review
of Applied Linguistics in Language Teaching, 31(4), 269–289.
Yorio, C. (1989). Idiomaticity as an indicator of second language proficiency. In
K. Hyltenstam & L. K. Obler (Eds.), Bilingualism across the lifespan: Aspects
of acquisition, maturity and loss (pp. 55–72). New York: Oxford University
Press. https://doi.org/https://doi.org/10.1017/CBO9780511611780.005
Young, R. (1996). Form-function relations in articles in English interlanguage. In
R. Bayley & D. R. Preston (Eds.), Second language acquisition and linguistic
variation (pp. 135–175). Amsterdam: John Benjamins. https://doi.org/10.1075/
sibil.10.07you
Yule, G. (1998). Explaining English grammar. Oxford: Oxford University Press.
Zabor, L., & Rychlewska, A. (2015). The effectiveness of written corrective feedback
in the acquisition of the English article system by Polish learners in view of the
counterbalance hypothesis. In L. Piasecka, M. Adams-Tukiendorf, & P. Wilk
(Eds.), New media and perennial problems in foreign language learning and
teaching (pp. 131–150). Berlin: Springer.
Zdorenko, T., & Paradis, J. (2008). The acquisition of articles in child second
language English: Fluctuation, transfer or both? Second Language Research,
24(2), 227–250. https://doi.org/10.1177/0267658307086302
Appendix 1 221

Appendix 1

Study 1: Research instrument

The text below does not have articles. Write the articles in the correct
places, as in the example.

Example: He is most wonderful person I’ve ever met.

1. Motorised boats harm ecology of waterways, unless their use is

kept at low level.
2. Glucose, or blood sugar, is produced in our bodies when we eat
carbohydrates.
3. We meet regularly, five times semester, at departmental meeting.
4. Time matters. Please try to send it in as soon as possible – sooner
better.
5. I want to choose foreign language that few people want to study.
Maybe I’ll learn Kurdish.
6. Plants in pots and containers require more water than you actu-
ally might think, smaller pot more critical problem. By midsum-
mer, herbs and vegetables in containers may need water twice day.
7. You should give him spoonful of this syrup every three hours.
222 Appendix 1

8. I’ll remember you until day I die.

9. I see that you haven’t eaten any of food I brought you two days
ago. Can I make you cup of tea?
10. Old leftist political parties are re-emerging to demand that gov-
ernment again expand its role in economy to help poor, even at
price of discouraging foreign investors.
11. I was lucky ball didn’t hit me in face.
12. New version of insurance policy makes number of alternatives
open to insured.
13. Do you speak English?
14. I was recently asked about my hopes for future by friend of mine.
What I know is that I’d like to have kids. And I’d like to live in
luxury apartment one day.
15. Immediately after graduation I need to get job. It doesn’t neces-
sarily have to be in my field, and I’m prepared to move anywhere
where I can find work. Acquaintance of mine was recently offered
position in Berlin and he moved there without moment’s hesita-
tion.
16. What remarkable player he is. His performance today really
impressed me. What shame he didn’t get picked for team.
17. Every member of Royal Family enjoys star status; they are used
to being centre of attention and there is strong unstated rivalry
between them.
18. He was cut in hand in same fight, according to testimony.
Appendix 2 223

Appendix 2

Study 1: Test items and their frequency

in BNC and COCA
 BNC 100,000,000 COCA 450,000,000
Pair Version A/B Article Target phrase Frequency
raw per million raw per million
1. Version A: a a friend of mine High 230 2.30 1,327 2.95
Version B: an acquaintance of mine Low 1 0.01 33 0.07
2. Version A: a what a shame High 120 1.20 173 0.38
Version B: what a remarkable player Low 0 0.00 0 0.00
3. Version A: a twice a day High 142 1.42 754 1.68
Version B: five times a semester Low 0 0.00 0 0.00
4. Version A: the the sooner the better* High 28 0.28 135 0.30
the smaller the pot, the more
Version B: Low 1 0.01 0 0.00
critical the problem
5. Version A: a a cup of tea High 619 6.19 876 1.95
Version B: a spoonful of syrup Low 0 0.00 1 0.00
6. Version A: the the day I die High 11 0.11 81 0.18
Version B: the food I brought Low 0 0.00 1 0.00
7. Version A: the help the poor High 21 0.21 241 0.54
Version B: open to the insured Low 2 0.02 0 0.00
8. Version A: the hit (someone) in the face High 26 0.26 115 0.26
Version B: cut in the hand Low 0 0.00 2 0.00
 BNC 100,000,000 COCA 450,000,000
Pair Version A/B Article Target phrase Frequency
raw per million raw per million
9. Version A: zero speak English High 174 1.74 1,328 2.95
Version B: learn Kurdish Low 0 0.00 0 0.00
10. Version A: a get a job High 299 2.99 1,749 3.89
Version B: live in a luxury apartment Low 0 0.00 0 0.00
11. Version A: zero have kids High 42 0.42 1,158 2.57
Version B: eat carbohydrates Low 0 0.00 11 0.02
12. Version A: the the centre of attention** High 85 0.85 392 0.87
Version B: the ecology of waterways Low 1 0.01 0 0.00

* The frequency count includes both punctuation versions: the sooner the better and the sooner, the better.
** The frequency count includes both the spelling center and centre.
Appendix 3 227

Appendix 3

Study 2: Research instrument

The sentences and texts below do not have articles. Write the articles
in the correct places, as in the example.

Example: He is most wonderful person I’ve ever met.

1. I often wake up in middle of night.

2. Smoking cigarettes results in 30-fold increase in chance of con-
tracting lung cancer.
3. I think I’ll have nightmares about it for rest of my life.
4. People should be judged on merit of their character.
5. There are many interesting places in distant north where you can
experience true arctic winter.
6. He made some final corrections and few slight changes in phras-
ing of his report, and then sent it to his boss.
7. This magazine has strong position in nexus of global multi-na-
tional media.
8. He was waiting for me at top of stairs.
Appendix 3 229

Text 2:
Around 20,000 people live in Kuril Islands, chain of islands connecting
northernmost part of Japan to tip of Russia’s Kamchatka Peninsula. Islands
look like dotted line on one side of which lies vast Pacific Ocean; on other side
is Sea of Okhotsk. Islands have been claimed by both Russia and Japan. Two
countries established diplomatic relations for first time in 1855. Since then,
they have traded possession of islands. Treaty of Saint Petersburg, signed in
1875 at closure of very long negotiations, gave islands to Japan in exchange for
island of Sakhalin. Soviet Union conquered islands at end of World War II. At
period of Cold War, Japan made claim that four southernmost islands should
be returned to Japan. For Russian leaders disputed islands play important
role in context of defence planning, as without them it is impossible to con-
trol access to Sea of Okhotsk. Some attempts at negotiations were made by
two countries at dawning of new millennium.
TECHNICAL EDITOR
Karolina Wąsowska

PROOFREADER
Małgorzata Szul

TYPESETTER
Marta Jaszczuk

Jagiellonian University Press

Editorial Offices: ul. Michałowskiego 9/2, 31-126 Kraków
Phone: +48 12 663 23 80, Fax: +48 12 663 23 83