Summary
#e paper addresses some major misconceptions about article use in English, proceeding from 
purely syntactic issues to those relating directly to pragmatics. It is based on authentic, perfectly 
acceptable examples of article use that many Slovenian students of English would describe as 
‘odd’ or ‘not in accordance with the rules’. #e students’ explanations as to why the examples in 
question should be ruled out confirm the hypothesis that misconceptions about article use are largely 
ascribable to an insufficient understanding of grammatical rules. #e rules governing article use 
are often misunderstood due to inaccurate interpretations of the terms defining/restrictive, definite, 
identifying, specifying, classifying, etc. #e commonest mistake is equating defining with definite, and 
defining/restrictive with identifying, the consequence being the overuse of the definite article. Another 
important point made in the paper is that article use is a matter of pragmatics. #e choice between 
the definite and indefinite articles reflects the speaker’s decision to present a piece of information as 
hearer-old or hearer-new respectively. 
Key words: article use, (in)definiteness, reference, identification, restrictive function, identifying 
function, old and new information
 Povzetek
V članku obravnavam nekaj večjih zmot o rabi člena v angleščini, in sicer od takšnih povsem 
skladenjske narave do takšnih, ki spadajo naravnost v pragmatiko. Članek je zasnovan na 
pristnih angleških, popolnoma sprejemljivih zgledih rabe člena, ki pa so po mnenju mnogih 
slovenskih študentov angleščine ‘čudni’ oziroma ‘niso v skladu s pravili’. Razlage, ki jih ob tem 
navajajo študentje, potrjujejo predpostavko, da je zmote o rabi člena v veliki meri moč pripisati 
pomanjkljivemu razumevanju slovničnih pravil. Napačno razumevanje pravil je pogosto posledica 
nepravilnega tolmačenja izrazov omejevalnost, določnost, identifikacija, specifičnost, vrstnost itd. 
Najpogostejša napaka je enačenje omejevalnosti z določnostjo oziroma identifikacijo, kar ima za 
posledico prekomerno rabo določnega člena. Sploh pa je raba člena stvar pragmatike. Izbira 
med določnim in nedoločnim členom odraža odločitev govorca o tem, ali bo poslušalcu nekaj 
predstavil kot že znano ali kot novo informacijo. 
Ključne besede: raba člena, (ne)določnost, referenca, identifikacija, omejevalna funkcija, 
identifikacijska funkcija, stara in nova informacija
 DOI: 10.4312/elope.3.1-2.99-113

A point came in my teaching career when I realized that the only way to thoroughly understand 
the students’ mistakes was to find out what was going on in their heads. So I asked the students 
in my 3
rd
 and 4
th
 year translation classes to start collecting authentic examples of article use they 
might find ‘odd’ or ‘not in accordance with the rules’. I also asked them to provide explanations 
of why they thought them problematic. Interestingly, most of these ‘striking’ examples turned 
out to be perfectly explicable in terms of the very same rules the students would use as arguments 
to prove the opposite. #is brought me to the root of the problem: the students seem to know the 
rules, but they do not quite understand them. #e paper points out some major misconceptions 
that have arisen regarding article use in English.
 #e first point of attention is the Saxon Genitive and its cooccurrence with the article. A surprising 
number of students find examples like those in (1) ungrammatical because, as they claim, ‘the 
Saxon Genitive cannot be preceded by a determiner’.
(1)♥
1
a. the world’s first national park
b. the Queen’s dazzling crown jewels
c. the unique ecosystem’s death sentence
d. Diamonds are a girl’s best friends.
e. Self-consciousness is a man’s worst enemy.
f. A violin’s authenticity can only be determined through [...] 
It is true that a possessive determiner is mutually exclusive with other (central) determiners 
because the (central) determiner position in an English noun phrase can be filled only once 
(cf. Biber et al. 1999, 294, Quirk et al. 1985, 326). #is explains why a bike and Peter’s bike 
are correct whereas *a Peter’s bike is not. Nevertheless, the misconception that has grown up 
around article-possessor complementarity is that phrases like a boy’s bike are of the same type as 
*a Peter’s bike, with two determiners placed side by side. #e key question is: What counts as 
one determiner? An article, a demonstrative pronoun, a possessive pronoun, a Saxon Genitive 
phrase. It is crucial to make it clear that even in Peter’s bike the determiner is a Saxon Genitive 
phrase, not a Saxon Genitive noun. In other words, the determiner is a noun phrase in the 
genitive case. Bracketing can be of help too: [the/that/a/some boy]’s bike; [the world]’s first 
national park. #ere is something in brackets and that something is in the genitive case, with 
the whole lot (be it one word or more) representing one single determiner. If the article or 
pronoun in question is inside the brackets, its use cannot be precluded by the one-determiner-
only rule. A nice test for that would be shifting the material from the Saxon Genitive phrase 


into the postmodification: the bike owned by [the/that/a/some boy]; the first national park in 
[the world]. 
#e vagueness of the students’ comments on (1) is further proved by cases where a Saxon Genitive 
noun is used as a classifier and does not occupy the determiner position. In the noun phrase a 
lady’s bike, for example, the classifier lady’s restricts the headword bike to a particular type: a [lady’s 
bike]. #e determiner preceding the classifier can also be a Saxon Genitive phrase: [that child]’s 
[lady’s bike] (i.e. the [bike for ladies] owned by [that child]). 
 It’s a common fallacy that an ordinal numeral per se triggers the use of the definite article. 
Examples in (2) are by no means ungrammatical. 
(2)♥
a. #e book he bought at the auction proved to be a first edition.
b. More creative forms of help might be to put them on a credit card as a second card 
holder.
c. #e play was followed by a second sermon, reinforcing the lesson of the representation. 
d. #ey had a third child a year ago.
An ordinal numeral must be preceded by the definite article when a kind of mathematical order 
is  implied. #is, however, is not the case if the numeral has the function of a classifier, forming a 
very close syntactic unit or even a compound with the head noun, or if its meaning corresponds 
to that of other. Compare (3) and (4):
(3) #ey have three children. #e first was born the year they got married, the second a 
couple of years later, and the third a year ago. 
(4)
a. I’d like a second opinion on that matter. 
b. Do you have a second name?
c. I think you should give him a second chance. 
d. She took a second spoon of sugar, and a third one, and a fourth one, and [...] 
 One of the major misconceptions about articles in English is the belief that an of-phrase (or a 
prepositional phrase in general) in postmodifying position triggers the use of the definite article. 
Accordingly, many a student would rule out the sentences in (5) as ungrammatical.
(5)♥ 
a. Immediately, we are in an atmosphere of artificiality.
b. In ( ) areas where a thick blanket of snow lay luxuriantly right down to the valley a hundred 

years ago, only dirty patches can be seen today.
c. In a survey of farmed salmon on sale at major stores, ( ) samples from Morrisons contained 
malachite green.
#e examples in (5) are, of course, perfectly acceptable. #e above-mentioned ‘rule’ applies 
only in cases when the function of the prepositional phrase is identifying, i.e. when it narrows 
down the set of possible referents of the headword to a unique, identifiable referent or subset 
of referents. Very often, however, the prepositional phrase merely restricts the meaning of 
the headword to a particular class. #is restrictive function is not to be confused with the 
identifying one. #e same degree of cautiousness is needed when the head noun is postmodified 
by a restrictive (i.e. defining) relative clause. #e fact that defining is often misunderstood as 
‘making definite’ is likely to lead to confusion. It should be noted that it is not necessary for a 
defining relative clause to identify the referent and ‘make the head noun definite’. Definitions 
of restrictive can be confusing too. Trask (1993, 239), for example, describes a restrictive item 
as ‘a modifier whose presence is essential for identifying the referent of the noun phrase’. 
I find this definition rather misleading, for it blurs the boundary between restrictive and 
identifying. Its vagueness lies in the fact that a restrictive modifier may well be essential for 
identification, but it will not suffice for the hearer to uniquely identify the referent. Or, in 
terms of formal logic, a restrictive modifier is a necessary condition for identification, but 
not a sufficient one.  
Misconceptions about the function of restrictive postmodification, along with the confusing 
terminology, give rise to the overuse of the definite article, especially in the plural. #e key word 
here is reference. #e presence of a restrictive postmodifier often makes it hard to decide whether 
the reference of a plural noun phrase is definite, indefinite or generic. It is therefore advisable to 
put the head noun into the singular and check the items that occur in determiner position. #ese 
should make the reference clear, as in (6) and (7) below. 
(6)
a. #e matter was first brought to our attention by readers of this magazine.
b. #e matter was first brought to our attention by people who read this magazine. 
c. #e matter was first brought to our attention by a/some reader of this magazine.  
d. #e matter was first brought to our attention by a/some person who reads this magazine 
/ by somebody who reads this magazine.
(Reference: specific, indefinite)
(7)
a. Readers of such magazines know what an invaluable source of inspiration they are.
b. People who read such magazines know what an invaluable source of inspiration they are.
c. A/Any reader of such magazines knows what an invaluable source of inspiration they are.
d. A/Any person who / Anybody who reads such magazines knows what an invaluable 
source of inspiration they are.
(Reference: generic)

#e point I want to make here is that notwithstanding the restrictive postmodification of readers 
and people above, their reference is not definite. Both types of postmodifier (i.e. the of-phrase 
and the relative clause) in (6) and (7) are restrictive, but neither of them is identifying. In other 
words, the postmodifier restricts the set of possible referents to a subset, but it does not enable the 
hearer to uniquely identify it. In (6), the speaker may have some specific readers in mind, but he 
presents them as indefinite because he does not expect the hearer to be able to identify them (cf. 
Taylor 1996, 185). In (7), the noun phrase headed by readers refers to a particular type of readers 
in general or, following Taylor (1996, 186), it is ‘used to make a statement about a category of 
entities as a whole’. 
Although the identifying and restrictive functions are to be kept clearly apart, they are not 
unrelated. In fact, the former derives from the latter, provided that the primary set of possible 
referents is already identified. #e situation is illustrated in (8). 
(8)
a. In general, students that are good at maths are good at grammar too. (generic: a–b)
b. In general, a student that is good at maths is good at grammar too. 
c. #e teacher gave the students some extra homework. (specific, definite: c–f)
d. #e teacher gave the students that were good at grammar no additional homework. 
e. #e teacher gave the student that was top in grammar no additional homework. 
f. #e teacher gave the student some extra homework. 
In (8a–b) and (8d–e) the same type of postmodification is used: a restrictive relative 
clause. If we compare (8a) and (8d), or (8b) and (8e), we can see that although the same 
type of clause is used as a postmodifier, the respective headwords do not share the same 
kind of reference. Furthermore, the reference can be definite even if there is no restrictive 
postmodification, as in (8c) and (8f). In (8c), the identification starts with the subject of 
the sentence, which sets in motion the following chain: the teacher > his/her class > his/
her students = the students. #e very same process is repeated in (8d) and (8e). When the 
primary set of referents (i.e. all the students in the class = the students) is identified, the 
postmodifying clause restricts it to a subset. #is final set of referents contains only those 
members of the primary set that belong to the type described by the clause. What might 
come as a real revelation at this point is the observation that the definite article in sentences 
like (8d) and (8e) is not due to the restrictive postmodifier but rather to the specific, 
definite reference of the headword before it becomes postmodified. #e identification of 
the final set of referents is possible only if the primary set of referents is identified first. 
#is goes hand in hand with the view that a restrictive postmodifier is a necessary, but not 
sufficient condition for identification. 
#e identification in (8f) can follow the same path as that in (8c), provided there is only one 
student in the class: the teacher > his/her class > his/her student = the student. If there are more 
students in the class, one of them must be singled out in the discourse before (8f) is uttered. 
(9) and (10) below illustrate the case in point. (8f) is repeated here as (9b) and (10b). 

(9)
a. "e student that had cheated admitted that the use of articles was a mystery to him. 
b. #e teacher gave the student some extra homework. (= 8f) 
(10)
a. A student complained to the teacher that the exercises he had set were too easy. 
b. #e teacher gave the student some extra homework. (= 8f) 
Example (9) illustrates ‘derived identification’ (cf. Anderson 2004). #e identification process in 
(9a) runs parallel to that in (8e), the fully established identity of the student in (9a) being picked 
up in (9b). In (10), on the other hand, the referent in question is introduced into the discourse 
through (10a) as a specific, yet not specifically identified entity. Or, as Anderson (2004) puts it, 
‘the speaker has a specific referent in mind, but does not, or cannot, identify it to the addressee’. 
#us the specific identity of the student in (10b) is not established. #is type of identification is 
referred to as ‘nonspecific identification’ (ibid). 
Example (10) raises another interesting aspect. It may be true that every uniquely identifiable 
entity is referred to by a definite expression, but not every definite expression refers to a uniquely 
identifiable entity. Let us compare (11) and (12).
(11)
a. #ey live on the outskirts of Paris.
b. #e stolen car was found in the vicinity of the station.
c. #ere are two birds nesting in the top of the tree.
(12)
a. #ey live on the outskirts of a big city.
b. #e car was found in the vicinity of an old building.
c. Have you ever seen a bird nesting in the top of a tree?
#e headword of each italicized phrase in (11) and (12) is a relational noun with a one-place 
argument. #e relation between the two can be explained in terms of a part-whole relationship 
where the given ‘part’ (be it an item or a group of items) is the only of its kind within the same 
‘whole’. If, as in (11), the ‘whole’ is presented as a specific, definite, identifiable entity, the ‘part’ 
named by the headword is uniquely identifiable too. #e reference of the outskirts, the vicinity 
and the top in (11) is definite. In (12), on the other hand, the ‘whole’ is presented as an indefinite 
entity, specific in (12a–b), nonspecific in (12c). Since the ‘whole’ cannot be identified, none of its 
‘parts’ can be identified either. #us the reference of the outskirts, the vicinity and the top in (12) is 
indefinite. #e reference can be tested by replacing the given phrase with a one-word equivalent. 
For example:
(13)
a. Have you ever seen a bird nesting in the top of a tree?  (= 12c)
b. Have you ever seen a bird nesting in a treetop?

#e indefinite article in (13b) should be convincing enough. Compare:
(14)
a. #ere are two birds nesting in the top of the tree. (= 11c)
b. #ere are two birds nesting in the treetop. 
#e examples in (12) illustrate the same phenomenon as (10): nonspecific identification. #ere 
are two entities involved, A and B. Entity A could be uniquely identified on the basis of its 
relation with B if B were a specific, definite entity. Since B is indefinite, no specific identity of A 
can be established.
Nonspecific identification can be observed even in cases where entity B (underlined below) is 
definite. Example (15) is taken from Anderson (2004, 444).
(15) 
a. Whoever (s)he may be, the murderer of those people is insane.
b. Whoever (s)he may be, the author of this pamphlet is a liar.
 A strong tendency to overuse the definite article can also be observed when the meaning of the 
headword is restricted by certain types of premodification: 
(16)♥
a. ancient Greece; present-day Germany; 9
th
-century Northumbria 
b. British Egyptologists; Egyptian surgeons; Slovenian archaeologists; Arabic scholars
c. human language; twentieth-century linguistics
d. #e chemical was used as a cheap alternative to licensed medicines for dealing with     
parasites.
#e winners in this respect are items denoting  periods of time (especially when premodifying 
geographical names) and nationalities or countries of origin. #is is likely due to two facts. First, 
the specific, unique, definite reference of these entities makes them highly reliable means of 
identifying other entities. Second, they seem to evoke pictures of contrast in the hearer’s head, 
for example: ‘ancient, not modern’, ‘British, not Slovenian’, etc. Another problem arises when 
the premodifier begins with an ordinal numeral (which I call ‘double trouble’ in class) because 
many a student will see this as an additional argument for using the definite article. T o illustrate 
how the definite article before an ordinal disappears when the ordinal introduces a premodifying 
item, the following example can be used:
(17)
a. works by many sculptors and architects of [the 20
th
 century]
b. works by many [20
th
 century] sculptors and architects

c. works by many of the [20
th
 century] sculptors and architects
d. works by many of the above-mentioned sculptors and architects
It should be noted that the definite article in (17c) has nothing to do with the ordinal numeral. 
#is is confirmed by (17d), where the same article is used, occupying the determiner position in 
the phrase co-headed by sculptors and architects.
Example (17) may help dismiss one argument, but it still remains rather difficult to comprehend 
why no definite article is needed in (16). In the case of countable nouns, the strategy of 
singularization (cf. (6–8)) can be employed to check the reference:
(18)
a. British Egyptologists have made a new, surprising discovery.
b. A/Some British Egyptologist has made a new, surprising discovery.
It should be admitted, though, that the definite article with plural headwords is often merely 
redundant rather than wrong:
(19)
a. #e chemical was used as a cheap alternative to (the) licensed medicines for dealing with     
parasites. (<> 16d)
b. #e chemical was used as a cheap alternative to any licensed medicine (= any of the 
licensed medicines) for dealing with parasites.
One explanation could be the view that despite its generalizing meaning, licensed medicines 
in (16d) involves partitivity, i.e. it assumes the existence of a particular subset of medicines. 
It is obvious that the example does not represent the core generic type, for ‘full genericness 
demands both nonspecificity and nonpartitivity’ (Anderson 2004, 446). My point here is 
that genericness seems to be far closer to definiteness than we like to think. If we think of 
people in general, for example, isn’t it possible to perceive them as ‘definite’ in the sense 
that they represent one big, unique, identifiable ‘group’: mankind? It should nevertheless be 
pointed out that a generic interpretation of a definite plural in English is possible only if it 
has an attributive (Anderson 2004, 448). #is explains why the licensed medicines for dealing 
with parasites can be generic whereas the medicines would yield only a specific, definite 
interpretation.
If, on the other hand, the reference is indefinite, the definite article is unacceptable:
(20)
a. #e chemical was used as a cheap alternative to (some/*the) licensed medicines for 
dealing with parasites.
b. #e chemical was used as a cheap alternative to a/some licensed medicine (= one of the 
licensed medicines) for dealing with parasites.

A similar line of reasonig can be adopted in cases where the reference of a plural noun is restricted 
by the context:
(21)♥
a. #e wheel and its axle are now at the Institute of Archaeology, where they are awaiting 
preservation. [...] ( ) Experts have also started on the reconstruction of the cart. 
b. Britain decided to site a military base at Aldabra, effectively turning it into a giant aircraft 
carrier. ( ) Campaigners rolled back that decision in what was feted as a landmark victory 
for the modern environmental movement.
As regards (16a) and (16c), where no tests like singularization are available, the logic behind the 
omission of the article is much less obvious. Longobardi (2002, 358–9) bases his explanation of 
instances like (16a) on the absence of overt N-to-D raising of proper names in English. Anderson 
(2004, 444–6) points out that proper names are self-identifying and as such definite when used 
as arguments. #us (16a) can partly be explained by drawing a parallel to (8d), where the definite 
article is not due to the restrictive clause postmodying the headword, but to the specific, definite 
reference of the headword before it gets postmodified. #e postmodifier simply restricts this 
already definite set of referents to a (definite) subset. Compare:
 
(22)
a. #e teacher gave the students some extra homework. (= 8c)
b. #e teacher gave the students that were good at grammar no additional homework. (= 8d)
c. I’m reading a book about Egypt.
d. I’m reading a book about ancient Egypt.
 #e article use to be discussed in this section is an intriguing issue. It should be noted that in 
most (if not all) of the examples below, the indefinite article (or its zero counterpart preceding 
a name, an uncountable noun or a noun in the plural) can be replaced by the definite article or 
some other definite determiner. #e question is therefore not that of grammaticality. Example 
(23), for example, is no less grammatical if the indefinite article is omitted: 
(23)
a. ♥ I explained that a Mr. George Cole, with whom I was currently acting, had damaged it.
b. I explained that Mr. George Cole, with whom I was currently acting, had damaged it.
In fact, (23b) seems to be completely in accordance with the general view that a known, old piece 
of information is not to be presented as an indefinite entity. #at the person referred to as Mr. 
George Cole is known to the speaker is obvious from the information given in the postmodifying 
clause. But is he known to the hearer too? According to (23b), the answer is yes. #e use of a 
name in (23b) conveys the speaker’s assumption that the hearer too can identify the referent. As 
pointed out by Anderson (2004, 443), the speaker’s and hearer’s assumed co-identification of 

an entity is reflected in the use of a name, a deictic or some other definite expression. If, on the 
other hand, the speaker assumes that the hearer cannot identify the given entity, an indefinite 
expression should be used. Example (23a) is a case in point. 
It is crucial to understand that an entity that may be perfectly known to the speaker is not 
necessarily identifiable from the hearer’s point of view. #e speaker must tailor his message to the 
hearer’s knowledge and expectations, otherwise the communication is likely to break down. #e 
hearer’s mind is therefore a constant subject of speculation on the part of the speaker. 
Example (24) contains a typical description, with a number of indefinite entities introduced one 
after another. Of interest here are the indefinite article in front of field and the zero article in front 
of bushes. Again, the definite article could be used instead, but that would imply that the hearer 
should already know about the field and the bushes or at least find them identifiable on the basis 
of the given context (cf. the station; the road). #e speaker, however, has decided to present the field 
and the bushes in the same way as the rest of the enumerated entities, that is as hearer-new. 
(24)♥ With a sort of military precision that astonished him, she outlined the route that 
he was to follow. A half-hour railway journey; turn left outside the station; two kilometres 
along the road: a gate with the top bar missing; a path across a field; a grass-grown lane; a 
track between ( ) bushes; a dead tree with ( ) moss on it.
#e message is as follows:
>> ‘You will see a field and a path across it; you will see ( ) bushes and a track between them.’
It often happens in a narrative that something presented as hearer-new is in fact not new for the 
reader but only for a character in the story. Let us suppose that the route described in (24) really 
exists and that somebody who actually lives there happens to be reading the above description. 
Should such a reader find the indefinite article disturbing? No, when he realizes that he is in fact 
only an ‘indirect’ receiver of the message conveyed by the description.
It is possible to imagine a situation where the speaker (person A) describes the route to somebody 
who doesn’t know it (person B), and then reports this to somebody who knows the route very 
well (person C). Let us compare the following two scenarios:
(25)
a. (A to B) #ere is a gate with the top bar missing; a path across a field; a grass-grown lane;      
a track between ( ) bushes; a dead tree with ( ) moss on it.
(A to C) I told B about the gate with the top bar missing; the path across the field; the     
grass-grown lane; the track between the bushes; the dead tree with all that moss on it.
   
b. (A to B) #ere is a gate with the top bar missing; a path across a field; a grass-grown lane;      
a track between ( ) bushes; a dead tree with ( ) moss on it.
(A to C) I told B about a gate with the top bar missing; a path across a field; a grass-grown      
lane; a track between ( ) bushes; a dead  tree with ( ) moss on it. 

#e difference lies in the speaker’s choice of perspective. Although A knows that C can identify 
any part of the route, he can decide to present it from B’s point of view. In other words, the 
speaker can decide to treat a hearer-old entity as hearer-new.
Similarly, the fact that everybody knows about Einstein’s contribution to the development of 
the nuclear bomb does not preclude the possibility of presenting the bomb as a new piece of 
information:
(26)♥ When I failed to become a concert pianist, or even an accompanist for the church 
youth choir, she finally explained that I was late-blooming, like Einstein, who everyone 
thought was retarded until he discovered a bomb.
>> ‘who everyone thought was retarded until something happened: He discovered a bomb.’ 
As is illustrated in (27), information presented as hearer-new typically yields an existential 
interpretation:
2
(27)♥ With ( ) hyperinflation, ( ) wages among the lowest in Europe and an agricultural 
system that sometimes seems a vestige from the last century, the country is facing big 
problems.
>> ‘#ere are ( ) hyperinflation, ( ) wages [...] and an agricultural system [...], and because of 
this the country is facing big problems.’ 
While, due to its existential meaning, (27) is relatively straightforward, more attention should be 
devoted to the following example:
(28)♥ In a life lived largely for the sake of others, it was the one place that was her own.
A 4
th
 year student made the following comment on (28): ‘T o me, In the life lived [...]  would 
sound better since that life is definite: first, it is hers, and second, it is lived for the sake of others.’ 
#e student may have had a point: the life in question is definite. Compare:
(29) In her life / this life of hers, lived largely for the sake of others, it was the one place 
that was her own.
But the second of the two arguments given by the student should be dismissed. #e postmodifier 
in (28) in no way contributes to the definite reference of the headword. It should also be noted 
that the solution proposed by the student (with the definite article and a restrictive postmodifier) 
would imply that the character concerned lived at least two lives. 
#e function of the postmodifier in (28) is to highlight a particular aspect of the character’s 
life, presenting it as new information. #is is reflected in the use of the indefinite article. #e 


indefinite phrase in (28) behaves like a predicative item, a fact which becomes more transparent 
if we consider the following paraphrase:
>> ‘In her life, which was a life lived largely for the sake of others, it was the one place that was her 
own.’
#e rest of the examples follow the same logic:
(30)♥ #e north-eastern region of Karamoja is the only arid region in a country rich in 
lakes and rivers, and suffers from a cyclical problem of drought and famine.
>> ‘the only arid region in this country, which is a country rich in lakes and rivers, [...]’
(31)♥ [...] we were going into the unknown, although South Australia had not been 
foreign to me for quite some time. A dreary day was slowly approaching the evening, and 
our voyage was uncertain with the waves beating mercilessly against the shore.
>> ‘#e day, which was a dreary day, [...]’ 
(32)♥ After a strenuous landing and short consultation it was decided that we would be 
ferried to the island.
>> ‘After the landing, which was a strenuous landing, [...]’
As pointed out elsewhere in the paper, it is a common misconception that a restrictive modifier 
triggers the use of the definite article. Even when its function is defining in the proper sense 
of the word, the definite reference of the headword does not seem to be the consequence of 
this defining function but rather a necessary condition for it (cf. (8)). In examples (28–32), by 
contrast, the article reflects a special function of the modifier. #e reference of the headword 
is definite, and only the highlighting function of the modifier can trigger the indefinite article. 
If the modifier is omitted, a definite determiner is the only choice:
(33)
a. In this life of hers, it was the one place that was her own. 
b. #e north-eastern region of Karamoja is the only arid region in this country, and suffers 
from a cyclical problem of drought and famine.
c. [...] we were going into the unknown, although South Australia had not been foreign to 
me for quite some time. #e day was slowly approaching the evening, [...]
d. After the landing and short consultation it was decided that we would be ferried to the 
island.
It is, of course, possible to add a modifier to any of the headwords in (33) without changing the 
type of determiner. It should be noted, however, that the function of the modifier in such cases is 
solely descriptive, not highlighting. If it comes after the headword, it has to be enclosed in a pair 
of commas as a non-restrictive item:

(34)
a. In this life of hers, lived largely for the sake of others, it was the one place that was her 
own. 
b. #e north-eastern region of Karamoja is the only arid region in this country, rich in lakes 
and rivers, and suffers from a cyclical problem of drought and famine.
c. [...] we were going into the unknown, although South Australia had not been foreign to 
me for quite some time. #e dreary day was slowly approaching the evening, [...]
d. After the strenuous landing and short consultation it was decided that we would be 
ferried to the island.
 Last but not least, there are also some very simple, everyday words and constructions that may 
pose problems when it comes to article use. #ere is, for example, a general misconception that the 
adjectives so-called, unique and most (in the sense ‘very’) trigger the use of the definite article. I asked 
my 3
rd
 year students once how they would justify the article underlined in the example below:
(35) For millenia the atoll known as Aldabra has bloomed in the absence of man. [...] In the 
mid-1960s Britain appeared to deliver the unique ecosystem’s death by [...] 
T o my great disappointment, almost half of the class claimed the reason was the adjective unique. 
It was only omitting the adjective or/and replacing the article with a demonstrative that convinced 
them of the opposite. Unique seems to be misleading for its meaning, most for its form, whereas 
associating so-called with the definite article could be due to the fact that this adjective is often 
found in definite noun phrases, which might create the wrong  impression that so-called and the 
definite article are in a cause-result relationship. #e following sentences should dispel the above 
misconception.
(36)
a. #is is a so-called protractor.
b. #e amphibian known as proteus is a unique creature.
c. It was a most interesting evening.
Another fertile ground for mistakes is exclamatory sentences with what and such. It may be true 
that the pattern with countable, singular heads (e.g. What a clever child!) is a very frequent one, 
but this does not mean that the general pattern is ‘what a / such a + noun’. #e general pattern 
is ‘what/such + [noun phrase]’, the use of the article depending solely on the headword of the 
noun phrase: 
(37) 
a. What [a clever boy] he is!  (countable, singular)
b. What [( )/*a clever children] they are!  (countable, plural)
c. What [( )/*a terrible weather] we’re having!  (non-countable)

A third problematic usage – the last to be mentioned in the paper, yet far from making the 
list of ‘tricky words and constructions’ complete – can be found with measurements, or more 
specifically, with the pattern ‘a length/height/depth/weight/angle etc. of + amount’. For example:
(38) #e flute should be held at an angle of 90° to the axis of the head.
#ere might be a temptation to use the definite article in (38). What appears misleading is the 
fact that the angle size is universally and absolutely defined: be it called 90° or π/2, everybody 
knows exactly how wide the angle is. #e misconception here is that 90° is in apposition to angle 
in the very same way as London is in apposition to city in the city of London. But while London is 
the name of a unique entity, 90° can name an infinite number of entities, which in fact represent 
instances of a type:
 
(39)
a. We visited *a [city of London] / *[cities of London].
b. #e teacher drew an [angle of 90°] using only a ruler and a pair of compasses.
c. #e teacher kept drawing [angles of 90°] using only a ruler and a pair of compasses.
Compare:
(40)
a. #e teacher drew a [right angle] using only a ruler and a pair of compasses. 
b. #e teacher kept drawing [right angles] using only a ruler and a pair of compasses.
c. #e flute should be held at [right angles] to the axis of the head.
It should be admitted, though, that the classifying function of the postmodifier is not always as 
obvious as in the above examples. T o avoid even greater confusion, it is probably best to simply 
learn the pattern because ‘this is how it is’. A clear distinction, however, is to be drawn between 
this pattern and phrases in which the postmodifier designates the ‘possessor’ of the quality named 
by the headword: 
(41)
a. #ese fish can reach a length of over two metres. (<> What (kind of) length?)
b. #e length of these fish can be over two metres. (<> Whose length?)
 #e most frequently held misconceptions about article use seem to be the following: (i) the 
category of article is incompatible with the Saxon Genitive; (ii) an ordinal numeral is always 
preceded by the definite article; (iii) a restrictive modifier per se (especially a postmodifying of-
phrase or a restrictive relative clause) triggers the use of the definite article. As shown in the paper, 
all the three erroneous beliefs can be ascribed to an insufficient understanding of grammatical 
rules. #e terminology used is often interpreted incorrectly. #e commonest mistake is equating 

defining with definite, and defining/restrictive with identifying, which can lead to the false conclusion 
that the presence of a restrictive item yields a specific, definite interpretation. It is therefore 
absolutely vital to thoroughly understand these terms and the concepts they denote. It is also 
worth keeping in mind that article use is primarily a matter of pragmatics: the choice between 
the definite and indefinite articles reflects the speaker’s decision to present a piece of information 
as hearer-old or hearer-new, depending on the hearer’s knowledge and expectations.