FROM A CLASSIFIER LANGUAGE TO A MASS-COUNT LANGUAGE:

WHAT CAN HISTORICAL DATA SHOW US?

Danica MacDonald
University of Calgary

1. Introduction
This paper investigates the historical development and modern-day uses of the Korean
morpheme -tul and proposes a new analysis of this morpheme based on historical data.
The morpheme -tul is generally considered to mark plurality; this is surprising as Korean
is considered to be a classifier language, which does not generally mark plurality
morpho-syntactically.
Korean -tul has been studied extensively; however, there is little consensus as to
the distribution or function of this morpheme. My research takes a new approach to the
analysis of -tul by examining its historical development over the past 100 years. The
goal of this paper is to shed light on the modern-day uses of -tul by illuminating its past.
The general question this paper addresses is whether the development of -tul is consistent
with the properties of a classifier language and whether, based on this development,
Korean still qualifies as a classifier language. I show that this morpheme is currently
undergoing change and that Korean appears to be shifting from a classifier language to a
mass-count language.
This paper is organized as follows. Section 2 provides necessary background
information on mass-count and classifier languages and outlines the specific research
questions investigated in this study. Section 3 presents the historical corpus study which
was conducted in order to investigate the historical development of -tul. The final section,
Section 4, provides a conclusion and directions for future research.

2. Background Information
This section provides background information on the mass-count and classifier language
distinction (Section 2.1) as well as background information concerning the distribution of
the Korean morpheme -tul (Section 2.2) and discusses its historical analysis in previous
literature (Section 2.3). Section 2.4 lays out the specific research questions discussed in
this paper.

2.1 Mass-count and classifier languages

Languages are generally classified into one of two categories with respect to number:
“mass-count” languages and “classifier” languages. In mass-count languages, such as
English, a non-plural count noun generally has a singular interpretation, while plural
reference requires obligatory plural marking. Classifier languages, such as Korean and
Chinese, do not mark plurality morpho-syntactically while in classifier languages a bare
noun can have a singular or plural reading (Cheng & Sybesma, 1999). As an example, the

Actes du congrès annuel de l’Association canadienne de linguistique 2014.

Proceedings of the 2014 annual conference of the Canadian Linguistic Association.
© 2014 Danica MacDonald
 2

bare Korean nouns chayk (‘book’) and haksayng (‘student’) can be interpreted as either
singular or plural. This is illustrated in (1).

(1) (=Nemoto’s (1a))

haksayng-un chayk-ul ilkessta
student-TOP book-ACC read
‘A student read a book.’
‘A student read books.’
‘Students read a book.’
‘Students read books.’
Source: Nemoto (2005: 384)

Another difference between classifier languages and mass-count languages is their use of
classifiers. English can enumerate by using a number plus a noun (for count nouns), as in
(2). However, in languages like Korean, a classifier is necessary for counting both mass
and count nouns, as in (3).

(2) three chairs

(3) (=Shin’s (2b))

chayk twu kwen
book two CL[flat things]
‘two books’
Source: Shin (2009: 132)

To account for differences between mass-count and classifier languages,

Chierchia (2008a,b) developed a formal theory called the Nominal Mapping Parameter.
While the formalities of this theory are outside the scope of the paper the typological
picture which emerges is that no language is predicted to have both a generalized
classifier system and plural marking. This is schematized below in (4).

(4) a. Plural marking ⟹ No generalized classifier system

b. Generalized classifier system ⟹ No plural marking

Korean poses a problem for the typological picture in (4). Like classifier
languages, Korean requires a classifier for counting; however, unlike other classifier
languages, Korean makes use of an optional plural-marker, -tul (e.g. Kang 1994, Baek
2002, Kim 2005), as will be illustrated in detail in the following section. Thus, Korean
appears to have properties of both system types.

2.2 Korean -tul

In Section 2.1, we saw that a bare Korean noun is argued to elicit multiple readings.
Additionally, while Korean lacks obligatory singular/plural morphology, it has an
 3

optional plural-marker -tul (henceforth referred to as “intrinsic -tul”) (e.g. Kang 1994,
Baek 2002, Kwak 2003, Kim 2005). The word sakwa (‘apple’) in (5a) can refer to both a
single ‘apple’ and multiple ‘apples’. In (5b), only a plural reading is possible since the
noun sakwa is marked with the plural morpheme -tul.

(5) (=Kang’s (14a,b))

a. sakwa-ka chayksang wui-ey issta
apple-NOM desk top-at exist
‘There is an apple on the desk’
‘There are apples on the desk.’

b. sakwa-tul-i chayksang wui-ey issta

apple-PL-NOM desk top-at exist
‘There are apples on the desk.’
# ‘There is an apple on the desk.’
Source: Kang (1994: 6)

Kim (2005:49) and Nemoto (2005: 391) claim that -tul can be used with any type
of noun as long as the referent is able to be individuated (or counted), as shown in (6a-c).

(6) (=Nemoto’s (20a-c)

a. haksayng-tul (human noun)
student-PL
‘students’

b. goyangi-tul (animate noun)

cat-PL
‘cats’

c. chayk-tul (inanimate noun)

book-PL
‘books’
Source: Nemoto (2005: 291)

Kim (2005:49) adds that -tul cannot occur with what is generally considered to be a
mass noun (such as mul ‘water’), unless it denotes ‘kinds’ or ‘distributions in containers’.
That is, -tul can only be used with mass nouns if -tul can shift a mass meaning into a
count meaning. The example in (7) illustrates this.
 4

(7) (=Kim’s (40b))

mul-tul (mass)
water-PL
*‘waters’
‘many kinds of water’
Source: Kim (2005: 49)

Kim’s (2005) and Nemoto’s (2005) claims as to what type of noun -tul can attach to
are not, however, uncontroversial. Their analysis supports Song’s (1997:206) findings
that -tul can be attached to both animate and inanimate nouns; however, Song (1997)
states that the plural-marker occurs mainly with human nouns, less frequently with non-
human animate nouns and far less frequently with inanimate nouns (206). Suh (2008:240)
asserts that pluralization is dispreferred on non-human nouns. With the exception of Kim
(2005), the above mentioned authors agree that -tul cannot be suffixed to mass nouns. As
we see, there is little consensus as to the type of nouns that -tul can attach to. Table 1
provides a summary of the distribution of -tul.

Table 1: Summary of the distribution and use of intrinsic -tul

Human Animate Inanimate Mass
Kim (2005)    
Nemoto (2005)    
Song (1997)    
Suh (2008)  ? ? 

With respect to the function of -tul, by far the most widespread claim is that -tul is
a plural-marker (e.g. Kang 1994, Kwak 2003, Kim 2005). This is not, however, the only
claim. As an opposing view, Park (2008) argues that Korean -tul is not a plural marker,
but rather a marker of distributivity or collectivity. Park claims that this shows that -tul
has a distributive function rather than a mere pluralizing effect, and in order to express
distributivity -tul is required, not optional. A third proposal, which discusses a slightly
different use of –tul is that of a focus particle. Song (1997:218) claims that it is not a
random marker, but rather a fully motivated one. Song found that by leaving out different
parts of a sentence marked with -tul, different emphasis or different focus was elicited. I
return to discuss the various debates with respect to the function and distribution of -tul in
Section 4.3.

2.3 Historical proposals on the use of –tul

Park (2010:2) proposes that Modern Korean -tul developed into an inflectional plural-
marking morpheme from Middle Korean. She proposes that subsequently, the function of
plural-marking -tul was further extended to be used as an agreement marker. Park does
not make use of historical data to support her claims, so that her account appears rather
speculative. If it is accurate, we should expect to find -tul used as a plural-marker at
various stages from early Modern Korean to present-day Korean. We should also expect
 5

to find cases of -tul where there is a clear agreement function (and not a focus function,
as was proposed by Song (1997)).
In addition to Park’s (2010) claim, various linguists have put forth claims that
modern day Korean -tul was borrowed from English-language usage. Martin (1992:830)
claims that in modern-day Korean, -tul is spreading and becoming more popular on
analogy with foreign-language plurals, particularly due to language contact between
English speakers and Koreans. Lee (1989:34-5) proposes that Koreans believe that -tul
should function like the English plural formation and so they have recently begun to treat
-tul as an equivalent to English plural-marker. Finally, Baik (1992:25) observes that
between 1965 and 1985, Korea underwent a period of extensive contact with English and
proposes that numerous grammatical borrowings, including the borrowing of the English
number-marking, ensued. While a borrowing hypothesis is an interesting one, it is
evident from historical sources that Korean -tul existed (and was used on nouns with
plural reference) long before contact with and influence by the English language.
Park’s (2010) proposal that -tul developed as a plural-marker from Middle Korean
and the borrowing hypothesis, together with the previous analyses of -tul from Section
2.2, allow us to refine the research question introduced in Section 1.
2.4 Research questions
The specific research questions that the historical investigation was designed to address
are the following:
1. What does the historical data tell us about the use and development of -tul?
How has this morpheme evolved over time?
2. Is -tul really an optional plural-marker (or does it mark some sort of
information structure)?
3. Is there evidence in the data to account for a borrowing hypothesis?
4. What is the status of Korean with respect to the mass-count/classifier
language dichotomy?
The corpus study presented in the following section was designed to help answer
these questions.

3. Corpus study
To answer the questions presented in Section 2.4, I analyzed newspaper articles1 from an
approximate 100-year period. I specifically searched for data on the distribution of -tul,
the number of instances of -tul in each article, the type of nouns which -tul attached to, as
well as cases where -tul could have been used as a plural-marker, but was not. This
section focuses on the methodology I used to collect my data, the data that I collected,
and a discussion of the implications of these findings.

1
This text-type was chosen as it was easily accessible online, provided ample data to consider for this
study, and allowed me to compare texts of the same type produced continuously over the time period
chosen.
 6

3.1 Methodology

The corpus was comprised of 125 newspaper articles, 25 for each of the 5 years that were
selected. The newspaper articles from 1924, 1946, 1970, and, 1995 were retrieved online
from the “Dong-A Ilbo” newspaper archive, while the 2011/13 articles were retrieved
online from the “Chosun Ilbo” newspaper. Within these years, articles of similar length
were chosen for analysis. The average word length for each article was between 380 and
397 words for the five time periods.
Once the articles were selected, the occurrences of -tul were collected and analyzed
for the distribution, the number of occurrences, and the type of noun that this suffix
attached to. Once the instances of -tul were highlighted, two native speakers of Korean2
were asked to provide feedback as to their interpretation of -tul in each case. The data
collected is summarized and discussed in the following sections.

3.2 Data

The number of occurrences of -tul in the 5 different time periods investigated in this
study is summarized in Figure 1 below. The dark gray lines represent cases where -tul is
used in newspaper articles, and the light gray lines represent cases where the text clearly
has a plural reading for a given noun, but where -tul is not used. We can see that in the
earlier newspaper articles, far fewer cases of -tul appeared in the texts and far more cases
where a bare noun conveyed a plural interpretation. In 1924, there are only 9 cases where
-tul is used, compared to 63 cases where there is a clear plural reading but no use of -tul.
Similar results are found in 1946, where there are 14 occurrences of -tul compared to
68 non-occurrences of -tul. The occurrences of -tul in the texts increase much more
rapidly in the later years. In the articles from 1970, there are 49 instances of a noun being
marked with the suffix -tul and only 34 occurrences of a plural reading without -tul. In
1995 there are 89 cases of -tul being used with only 18 instances of a plural reading
without -tul. Finally, in the most recent data from 2011, there are 154 instances of -tul
being used, and only 4 bare nouns which have a plural reading.

2
The native-speakers of Korean were 46 (male) and 34 (female) years of age. Both Koreans were asked to
participate in this study as they both had extensive knowledge of Chinese (Mandarin). This knowledge was
needed as in the earlier texts many nouns and verbs were written in Hanja ‘traditional Chinese characters’
instead of Hangul ‘Korean characters’. As Korean rarely uses Hanja script in Modern Korean, not all
native speakers of Korean are able to read and give judgments on this script.
 7

Instances of -tul in Korean newspapers (by year)

200

150

100

50

0
1924 1946 1970 1995 2011

Occurrences of -tul Non-occurrences of -tul

Figure 1: Instances of -tul in Korean newspapers (by year)

The remainder of this section focuses on the specific data broken down by years.

4.2.1 Data: 1924

As we see in Figure 1 there are only 9 cases of –tul in the data from 1924. Of these 9
cases, there were 8 cases where -tul is used to express plurality. These cases are
interesting, however, in that my Korean consultants claim that -tul only appears on a
noun in cases where emphasis or contrastive focus is placed on the noun and in examples
where the noun marked with -tul is preceded by the demonstrative ku ‘this/these’. In (8), -
tul appears to be used in a case of contrastive focus. Here, the emphasis is placed on the
word ‘materialists’ (as opposed to some other group of people).

(8) yumulonja-tul-un malha-rira…

materialist-PL-TOP say-would…
‘Materialists would say…’
Source: Dong-A Ilbo (1924: Jan 2 (p. 2/4))

In example (9), we see that the first occurrence of the noun salam ‘people’ does not
involve the suffix -tul; however, the second occurrence of salam ‘people’, which is
preceded by the demonstrative ku ‘this/these’, does. Both instances of salam elicit a
plural interpretation; however, only that preceded by the demonstrative is marked with
-tul. We can also see that the nouns, moksori ‘voices’ and si ‘poems’ (nouns which do not
refer to humans), have a plural interpretation but are not marked with the plural suffix
-tul.
 8

(9) celmun-salam-ul manhi-pulletanokho-nun koun-moksori-wa

young-person-ACC many-invited-TOP beautiful-voice-and

si-ro ku saram-tul-ul chwikheyhanun-cungita

poem-with this person-PL-ACC make them intoxicated-are
‘They invited many young people and they are making these people
intoxicated with their beautiful voices and poems.’
Source: Dong-A Ilbo (1924: June 3 (p. 3/4))

If we consider the type of noun that the suffix -tul attaches to, at this time, it only
appears to be used with nouns which refer to humans, such as yumulonja ‘materialist’ or
saram ‘people’. However, it is not used with all human nouns that have a plural reading.
There are many occasions where a noun clearly has a plural interpretation and refers to
humans, but where -tul is not used (one example was the first occurrence of saram
‘people’ in (9)).

4.2.2 Data: 1946

In the 1946 data, there are 14 cases of -tul which show a similar pattern to what we see
for 1924. The nouns with which -tul occurred are also similar to the pattern which we
observe for 1924. One of the differences between the data from 1924 and the data from
1946 is that not all nouns which are marked with -tul are preceded by ku ‘this/these’ or
seem to express some kind of emphasis or focus; in the 1946 data, there are two cases
which have a plural reading (and use -tul) which do not appear with ku ‘this/these’ or
seem to be focused. One of these examples is shown in (10).

(10) gyengchalkwan-tayphyo-tul-i samshipul-ul kyeongKyo-dong

police-representative-PL-NOM 31-on KyeungKyo-district

imshijengpu-lul pangmunhako…
provisional government-ACC visit…
‘Police representatives visited the provisional government in KyeunKyo-
district on the 31st…’
Source: Dong-A Ilbo (1946: Jan 1 (p. 2/2))

However, as we observe with the data from 1924, there are also many occasions in
which nouns referring to humans have a plural reading, but where -tul is not used. A
subset of these nouns is shown in (11).

(11) adong ‘children’

dongji ‘friends’
kisulja ‘engineers’
Source: Dong-A Ilbo (1946: various dates)
 9

The nouns which are not marked with -tul appeared in contexts which require a
plural interpretation, as in (12), where we see an example of a sentence clearly referring
to more than one child and in (13) where the noun shimin ‘citizen’ is modified by the
word baekman ‘many’ (lit. ‘one million’).

(12) man yuk sae-ui adong-un…

complete six year-of child-TOP…
‘…children under six years old…’
(Context: From this year on, children under six years old should, by law, be
accepted into schools.)
Source: Dong-A Ilbo (1946: Jan. 12 (p. 1/4))

(13) sori nophi weychinun baekman-shimin-un

sound (voice) high cry (shout) million-citizen-ACC
‘Many (lit. a million) citizens who cry out at the top of their lungs (lit. voice).’
Source: Dong-A Ilbo (1946: Jan. 1 (p. 2/2))

4.2.3 Data: 1970

In the newspaper articles from 1970, there are significantly more cases of -tul than are
found in the earlier texts. In 1970, there are 49 instances of -tul and 34 cases which have
a plural reading but where -tul was not used. In both 1924 and 1946, the only examples of
nouns marked with -tul are human nouns; however, in 1970, we also observe that while -
tul continues to combine primarily with human nouns, it also combines with a few non-
human, inanimate nouns such as jom-tul ‘points (of view)’ and munjae-tul ‘problems’.
The sentence in (14) shows an example of -tul appearing with an inanimate noun.

(14) kagaekajung-aeseo jaegi-doyeon munjae-tul-i…

all levels of society-LOC raised-have been problem-PL-NOM…
‘Problems have been raised from all levels of society…’
Source: Dong-A Ilbo (1970: Dec. 14 (p. 3/4))

The increase in number of occurrences of -tul is not the only change that occurred
at this time. We can also observe that the range of constructions that -tul can attach to
continues to expand and it is becoming more common to observe examples that do not
include uses with ku ‘this/these’ nor contexts where emphasis or focus is clearly marked.
In the 1970 data, the use of -tul seems to take on the use of functioning as a plural
marker, and not as a way of placing focus on emphasis on a noun.
At this time, it appears that the use of -tul is in a period of transition. While we see
a few cases where -tul seems to extend to include non-human, inanimate nouns, there are
still cases where -tul is not used with human or inanimate nouns, but where these nouns
have a plural interpretation in the text. One of these nouns, gusang ‘idea’, is illustrated in
(15).
 10

(15) sae-chae-gaehyen-ui kibong-pangchim-irago palsu issnun

tax-system-reform-GEN basic-policy-as regarded be

myekgachi gusang-ae
several idea-DAT
‘Several ideas that can be regarded as basic policy for tax system reform…’
Source: Dong-A Ilbo (1970: Dec 14 (p. 3/4))

4.2.4 Data: 1995

The data from 1995 contains 89 cases of -tul being used on both nouns denoting humans
and inanimate nouns. This time period also has a noticeably lower number of non-
occurrences of -tul. In this data, almost all nouns denoting humans which have a plural
interpretation take the suffix -tul and many inanimate nouns also do. In example (16), we
see the noun giep ‘enterprise’ which takes the suffix -tul, but we also see in the same
sentence that the word eryeum ‘difficulty’ has a plural interpretation, but does not take
the suffix -tul.

(16) giep-tul-i yejemhi manhun eryeum-ul gyekko issda

enterprise-PL-NOM still many difficulty-ACC facing are
‘Enterprises are still facing many difficulties.’
Source: Dong-A Ilbo (1995: Dec. 14 (p. 3/24))

At this point, -tul is functioning more like a plural marker and there appears to be
no trace of it marking focus or emphasis on the noun to which it attaches.
In the 1995 data, we also find that most cases which have a plural reading but
which do not use -tul involve inanimate or abstract nouns (such as eryeum ‘difficulties’ in
(16) above). In this data set, there is only one case of -tul not being used with a noun
referencing humans (pohaengja ‘pedestrians’) and this example is used with a classifier.
It is worth noting that amongst the nouns that do not have the plural-marker -tul, many of
these nouns occur with quantifiers such as modun ‘all’ (see (17)) or manhun ‘many’ (see
(16)).

(17) modun iljong

all schedule
‘all schedules’
Source: Dong-A Ilbo (1995: July 1 (page 7/20))

4.2.5 Data: 2011/2013

In the most recent data, there are 154 cases of -tul while there are only 4 instances where
there is a plural reading but no -tul. Of these, all nouns which denote humans use -tul
when expressing a plural reference, while the large majority of nouns which denote
inanimate and abstract nouns do so likewise.
 11

Of the entire set of data from 2011, according to my Korean consultants, only four
nouns had a plural interpretation but did not use the suffix -tul. These examples were
similar to the examples from 1995, in that the nouns are modified with words like
manhun ‘many’ or jeonchae ‘all’. There are also many examples where the noun
modified by quantifiers is marked with -tul, such as the example in (18).

(18) Gaegu-jieok jeonchae haksaeng-tul-i…

GeaGu-region all student-PL-NOM…
‘All the students in DeaGu region…’
Source: Chosun Ilbo (2013: April 19)

The other examples where there are plural-readings but no -tul were found in
examples with classifier constructions. One such construction is shown in (19). (19a) has
no classifier, only a numeral and a noun, while the second construction (19b) has a
numeral and a plural-marked noun. Neither of these classifier constructions appear to fall
under the “grammatical” classification for Korean classifier constructions that we saw in
Section 2.1, with the exception of (19a) being attested under Kim’s (2005) proposal that
highly frequent nouns which refer to humans can appear without a classifier.

(19) a. du saram b. du saram-tul

two people two people-PL
‘two people’ ‘two people’
Source: Chosun Ilbo (2011: various dates)

Another unique construction from the most recent data is the use of -tul with a mass
noun. As we saw in Section 2.1, mass nouns cannot normally take the suffix -tul. In this
case, mul-tul ‘water-PL’ is interpreted as ‘bottles or cups of water’ by my Korean
consultants. This is the only instance of -tul appearing with a mass noun in my texts;
however, it appears that -tul can also be used in some cases to force an individual
interpretation of a mass noun.

4.3 Discussion

In both the 1924 and 1946 data, there were very few cases of -tul. The cases I did find
were limited to use with human nouns. From the data collected, the use of -tul was not
necessary to express plurality (since a plural reading was possible, and in fact most often
occurred, without -tul), and -tul did not seem to be functioning as a plural marker alone in
these early newspaper articles. Instead, -tul seemed to be functioning as a way to place
emphasis or focus on a noun with plural reference to which it attached.
In the later data (from 1970, 1995, and 2011), -tul is used more frequently and its
use is extended to include non-human nouns, and later abstract nouns. After 1970, -tul
does not yet appear on all nouns which have plural referents. However, at later time
periods its use becomes more and more frequent and in the 2011/13 data it appears to be
 12

used consistently in the newspaper data. Table 2 summarizes the uses and development
of this morpheme over an almost 100 year period.

Table 2: Summary of uses of -tul

Time Period Animate (Human) Inanimate Abstract
1924/1946 - few instances - not attested in the data - not attested in the
- nouns appear data
focused

1970 - no longer appears - few inanimate nouns - same as above

focused
1995 - most human nouns - most inanimate nouns - same as above
2011 - all human nouns - most inanimate nouns - includes abstract
nouns

We can now attempt to account for the changes that we have observed in the
newspaper data. If we reconsider some of the claims discussed in Section 2.2, Lee
(1989), Baik (1992), and Martin (1992) have all made claims that modern -tul was
borrowed from English. Baik (1992:25) claims that during the period 1965 – 1985, which
fits in well with the period where we see an increased use of -tul, there was a period of
large contact between English and Korean. This is perhaps due to the influence of the US
military presence in South Korea at this time, and the use of the plural-marker -tul could
be a grammatical borrowing which parallels the English plural-marker -s as Martin
(1992), Baik (1992) and Lee (1989) propose.
However, as Harris & Campbell (1995:120) discuss, syntactic (or grammatical
borrowings) are considered by many to be extremely rare occurrences which only occur
between already similar language systems. Allen (1980 [1977]:380) claims that, “as a
hypothesis, we would expect syntactic influence only when the two languages had a good
deal of syntactic similarity to begin with.” While many consider grammatical borrowings
to be quite rare, many different types of grammatical borrowings (between languages that
belong to the same language families and languages which differ in numerous respects)
have in fact been documented (see Harris & Campbell 1995: 142-149 for discussion).
Harris & Campbell end up concluding, following Thomason & Kaufman (1988: 14), that
“as far as the strictly linguistic possibilities go, any linguistic feature can be transferred
from any language to any other language.” So, the question arises, is it possible that the
new distribution and use of Korean -tul was borrowed from English language usage? If
syntactic borrowing is possible and is more common than we would expect, what are the
chances of the distribution and use of -tul being borrowed?
While a borrowing hypothesis is possible, it is important to remember that Korean
-tul existed (and was used on plural nouns) long before contact with and influence by the
English language. However, it is also important to note that in the newspaper articles
from 1970 and later (the claimed period of extensive language contact between English
and Korean) we see a substantial increase in the use of -tul. While it is highly likely that
language contact with English may have influenced the development of -tul since the
 13

extensive period of language contact, more evidence is needed to be able to say that the
Korean plural-marking was borrowed from English. My study could not reveal any direct
evidence for such a claim and direct evidence may well be hard to come by.
If we return to Park’s proposal that -tul was an inflectional plural-marker that
developed from Middle Korean, as discussed in Section 2.2, it is difficult to make a claim
as to the development of -tul from Middle Korean, as I did not analyze texts from that
period. However, if Park’s (2010) analysis were accurate, then we would expect to find
instances of -tul functioning as a plural-marker in the texts from 1924. This was not the
case; rather, the occurrences of -tul in these texts seemed to mark focus and -tul was only
used with a limited number of nouns in the early texts. It certainly does not appear to be
inflectional as she claimed, at least if by inflectional one means obligatory and
systematic. As there were no cases of extrinsic -tul in the corpus, it is not possible to
assess Park’s (2010) claims that extrinsic -tul represents an extension of intrinsic -tul as
an agreement marker (cf. Section 2.2).
It appears from the data that in the early 1920s, -tul was used to mark focus or
emphasis on nouns with plural reference. Extrapolating from that data, I speculate that at
some point (before -tul lost its focus function), the intrinsic -tul use extended to mark
focus extrinsically (as Song 1997) proposes. At the time when the extension took place,
this would leave us with both uses of -tul having the same primary function: that of
marking focus. While the focus function is maintained on extrinsic -tul nouns in Modern
Korean, the primary function of intrinsic -tul of marking focus will over time have shifted
to one of marking plurality.
Next, if we return to the research questions from Section 2.3, we can see that the
historical data does, in fact, shed light on the modern uses of -tul. Over time, and perhaps
due to language contact with English, the distribution and use of -tul has drastically
expanded. Even though this study focused on an approximately 100-year period, the
increase in the use of -tul is extremely clear even in this short timeframe.
As we saw in Section 2.1, there is significant variability among researchers
working on Korean -tul with respect to its use and distribution. We saw variability with
respect to the noun type that -tul could attach to. Remember for example that Kim (2005)
proposes that -tul can be used with human, animate, inanimate, and mass nouns, while
Nemoto (2005) claims that it cannot be used with mass nouns and Suh (2008) claims that
it is dispreferred on non-human nouns. Given the fact that this morpheme is currently
undergoing development and constantly seems to be expanding in terms of both
distribution and use, it is not surprising that there is no clear consensus as to its use.

5. Conclusion & directions for future research

5.1 Conclusion

To conclude, Park (2010) claim that -tul developed from an autonomous noun which
grammaticalized to an inflectional affix. I propose rather that in Early Modern Korean,
-tul was originally used to mark focus on nouns with plural reference, rather than
functioning as a uniquely plural-marking morpheme in an inflectional Number category.
 14

From this perspective, -tul would not have been optional, but would have been used only
in certain specific contexts, and thus in a sense obligatory. This would establish -tul as
highly motivated and not a random morpheme, as would be the case in the earlier texts
under Park’s analysis.
I also propose, following Martin (1992), Baik (1992), and Lee (1989) that if -tul
functions like an optional plural-marker today, this is not entirely due to the development
from Middle Korean; rather, it is due, at least in part, to language contact.

5.2 Directions for future research

This study provides an initial examination of the historical development of -tul, and sheds
light on the changes of use of this morpheme over the past 100 years. However, it still
leaves many questions unanswered and this section provides various directions for future
research.
An extension of this study will also include texts from earlier time periods. Park
(2010) claim that -tul was used as a plural-marker in Middle Korean; however, it would
be interesting to re-examine this claim. Perhaps -tul was used in Middle Korean to mark
focus on nouns with plural reference instead. This is one of the predictions that my
proposal would make, since I suggest, following Martin (1992), that -tul was only being
used on a regular basis as a plural-marker following intensive contact with English.
I would also like to investigate data from North Korean newspapers. The North
Korean language has not had the contact with English that South Korean dialects of
Korean have had and it would be interesting to investigate what, if any, changes have
occurred in the North Korean data. If North Korean changes with the use of -tul were
observed similar to the changes in the South Korean data then we could speculate that
this is a language-internal change. If, on the other hand, -tul follows a different pattern in
North Korean dialects, it is possible that there were language-external changes taking
place.
A final plan for future research would involve investigating different text-types.
Newspapers are a formal source of data where colloquial language is rarely used. In order
to gain a more complete understanding of the evolution of -tul it would be interesting to
investigate text-types such as letters, or spoken language.
The loss of classifiers from classifier constructions in Korean, the obligatoriness
of the plural-marking -tul in certain constructions, the interpretation of a bare noun as
‘singular’, and the use of -tul on mass nouns as a form of individuating raises questions
as to the changing status of Korean with respect to Number-marking typology and helps
explain the difficulties that linguists encounter when working on Korean plural-marking
data. It appears that Korean started out as a classifier language, not unlike other typical
classifier languages, but the changes that the language is undergoing are causing it to
shift more towards becoming a mass-count language. Addressing the issue from a
historical perspective is something that has not been done before, and has led me to make
new insights. It would appear, then, that change of this kind is gradual, rather than
discrete, contrary to what Chierchia (2008a, b) would predict. Only time will tell what
 15

will eventually happen with the Korean classifier system and with the interpretation of
Korean bare nouns.

References

Allen, C. L. 1980 [1977]. Topics in diachronic English syntax. PhD dissertation, University of Massachusetts
(Published 1980. New York: Garland.)
An, Y.-R. 2007. Korean tul and English all. LSO Working Papers in Linguistics 7: Proceedings of WIGL
2007: 1-16.
Baek, M.-H. 2002. Handkwuke pokswu uymi yenkwa. [A study on Korean plural senses.] Discourse and
Recognition 9(2): 59-78.
Baik, M. J. 1992. Language shift and identity in Korea. Journal of Asian Communication 3(1): 15-31.
Cheng, L. & R. Sybesma. 1999. Bare and not-so-bare nouns and the structure of NP. Linguistic Inquiry
30(4), 509-542.
Chierchia, G. 1998a. Reference to kinds across languages. Natural Language Semantics6: 339-405.
Chosun Ilbo Newspaper (Online version): http://www.chosun.com
Cole, J. F. 1995. The prosodic domain hierarchy in reduplication. PhD Dissertation, Stanford University.
Corbett, G. 2000. Number. Cambridge: Cambridge University Press.
Dong-A Ilbo Newspaper (Online archived version): http://dna.naver.com/search/searchByDate.nhn#
Fluid Surveys (Online Surveys): www.fluidsurveys.com
Hwang, S.-H. & D. Lardiere. 2013. Plural-marking in L2 Korean: A feature-based approach. Second
Language Research 29(1): 57-86.
Kang, B.-M. 1994. Plurality and other semantic aspects of common nouns in Korean. Journal of East
Asian Linguistics 3: 1-24.
Kim, C.-H. 2005. The Korean plural marker tul and its implications. PhD Dissertation, University of
Delaware.
Kuh, H. 1987. Plural copying in Korean. In Harvard Studies in Korean Linguistics ,,eds. S. Kuno et al.,
239-250.
Kwak, E.-J. 2003. Interpretation of plural noun phrases in Korean. Linguistics 35:3-38.
Harris, A. & L. Campbell. 1995. Historical syntax in cross-linguistic perspective. Cambridge: Cambridge
University Press.
Lee, K.-M. & S. Ramsey. 2011. A history of the Korean language. Cambridge: Cambridge University
Press.
Lee, S. 1989. The subversion of Korean. English Today 20: 34-37.
Madigan, S., M. Yamada, & A. E. Peng. 2007. Inclusive reference in Korean: Implications for a theory of
plurality. Proceedings of Harvard ISOKL 2007: Online version.
Martin, S. 1992. A reference grammar of Korean. Tokyo: Tuttle Publishing.
Nemoto, N. 2005. On mass denotations of bare nouns in Japanese and Korean. Linguistics 43(2): 383-
413.
Park, S. 2010. Grammaticalization of plurality marker tul. Seoul International Conference on Linguistics.
Park, S.-Y. 2008. Plural marking in classifier languages: a case study of the so-called plural marking –tul
in Korean. Toronto Working Papers in Linguistics 28: 281-295.
Rullmann, H. & A. You. 2003. General number and the semantics and pragmatics of indefinite bare nouns
in Mandarin Chinese. Ms., University of Calgary.
Shin, K.-Y. 2009. Numeral quantifiers: NP modifiers and relational quantity nominal. Language
Research 45(1): 131-156.
Sohn, H.-M. 1999. The Korean language. Cambridge: Cambridge University Press.
Song, J.-J. 1997. The so-called plural copy in Korean as a marker of distribution and focus. Journal of
Pragmatics 27: 203-224.
Suh, E. 2008. The usage and interpretation of Korean –tul ‘plural’ by Heritage Language speakers. In
Selected Proceedings of the 2007 Second Language Research Forum, eds. M Bowles, et al., 239-251.
Somerville, MA: Cascasilla Proceedings Project.