Using an Ontology to Determine English Countability
Francis Bond and Caitlin Vatikiotis-Bateson  
* bond@cslab.kecl.ntt.co.jp ** caitlinvb@yahoo.com
2-4 Hikari-dai Seika-cho, Kyoto, Japan 619-0237
NTT Communication Science Laboratories,
Nippon Telegraph and Telephone Corporation
Abstract
In this paper we show to what degree the count-
ability of English nouns is predictable from their
semantics. We found that at 78% of nouns’
countability could be predicted using an ontol-
ogy of 2,710 nodes. We also show how this
predictability can be used to aid non-native
speakers to determine the countability of En-
glish nouns when building a bilingual machine
translation lexicon.
1 Introduction
In English, nouns heading noun phrases are typ-
ically either countable or uncountable (also
called count and mass). Countable nouns can
be modi ed by denumerators, prototypically
numbers, and have a morphologically marked
plural form: one dog, two dogs. Uncountable
nouns cannot be modi ed by denumerators, but
can be modi ed by unspeci c quanti ers such
as much, and do not show any number dis-
tinction (prototypically being singular): * one
equipment, some equipment, *two equipments.
Knowledge of countability is important when
translating from a source language without
obligatory number and countability distinctions
to a target language that does make num-
ber distinctions. Some examples are Japanese-
to-English (Ehara and Tanaka, 1993; Bond,
2001), Japanese-to-German (Siegel, 1996), and
Chinese-to-English.
For a system generating English, it is impor-
tant to know the countability of the head noun,
as this determines whether it can become plural,
and the range of possible determiners. Knowl-
edge of countability is particularly important in
machine translation, because the closest trans-
  This research was done while the second author was
visiting the NTT Communication Science Laboratories
lation equivalent may have di erent countabil-
ity from the source noun. Many languages, such
as Chinese and Japanese, do not mark count-
ability, which means that the choice of count-
ability will be largely the responsibility of the
generation component.
In this paper, we measure how well seman-
tic classes predict countability. Obviously, the
answer depends both on how many countability
distinctions are made, and how many semantic
classes are used. If every sense of every word
belongs to its own semantic class, then seman-
tic classes will uniquely, although not usefully,
predict countability. This is e ectively the po-
sition taken by Wierzbicka (1988), where the
semantics of a noun, given in the Natural Se-
mantic Metalanguage, always provides enough
information to predict the countability. On the
other hand, if there are only a handful of se-
mantic classes, then they will have little pre-
dictive power. We  rst de ne countability, and
discuss its semantic motivation (x 2). Then we
present the lexical resources used in our exper-
iment (x 3), including the ontology of 2,710 se-
mantic classes. Next, we describe the experi-
ment, which uses the semantic classes of words
in a Japanese-to-English transfer dictionary to
predict their countability (x 4). Finally, we
present the results and discuss the theoretical
and practical implications in (x 5).
2 Linguistic Background
Grammatical countability is motivated by
the semantic distinction between object
and substance reference (also known as
bounded/non-bounded or individuated/
non-individuated). Imai and Gentner (1997)
show that the presence of countability in En-
glish and its absence in Japanese in uences how
native speakers conceptualize unknown nouns
as objects or substances. There is de nitely
some link between countability and conceptual-
ization, but it is a subject of contention among
linguists as to how far grammatical countability
is motivated and how much it is arbitrary. Jack-
endo (1991) assumes countability and number
to be fully motivated, and shows various rules
for conversion between countable and uncount-
able meanings, but does not discuss any of the
problematic exceptions.
The prevailing position in the natural lan-
guage processing community is to e ectively
treat countability as though it were arbitrary
and encode it as a lexical property of nouns.
Copestake (1992) has gone some way toward
representing countability at the semantic level
using a type form with subtypes countable
and uncountable with further subtypes below
these. Words that undergo conversion between
di erent values of form can be linked with lexi-
cal rules, such as the grinding rule that links a
countable animal with its uncountable inter-
pretation as meat. These are not, however di-
rectly linked to a full ontology. Therefore there
is no direct connection between being an animal
and being countable.
Bond et al. (1994) suggested a division of
countability into  ve major types, based on
Allan (1980)’s noun countability preferences
(NCPs). Nouns which rarely undergo conver-
sion are marked as either fully countable,
uncountable or plural only. Nouns that are
non-speci ed are marked as either strongly
countable (for count nouns that can be con-
verted to mass, such as cake) or weakly
countable (for mass nouns that are readily con-
vertible to count, such as beer). Conversion is
triggered by surrounding context. Noun phrases
headed by uncountable nouns can be converted
to countable noun phrases by generating clas-
si ers: one piece of equipment, as described in
Bond and Ikehara (1996).
Full knowledge of the referent of a noun
phrase is not enough to predict countability.
There is also language-speci c knowledge re-
quired. There are at least three sources of ev-
idence for this: the  rst is that di erent lan-
guages encode the countability of the same ref-
erent in di erent ways. To use Allan (1980)’s
example, there is nothing about the concept de-
noted by lightning that rules out *a lightning
being interpreted as a  ash of lightning. In
both German and French (which distinguish be-
tween countable and uncountable uses of words)
the translation equivalents of lightning are fully
countable (ein Blitz and un  eclair respectively).
Even within the same language, the same ref-
erent can be encoded countably or uncount-
ably: clothes/clothing, things/stu , jobs/work.
The second evidence comes from the psycho-
linguistic studies of Imai and Gentner (1997)
who show that speakers of Japanese and En-
glish characterize the same referent in di erent
ways depending on whether they consider it to
be countable (more common for English speak-
ers) or uncountable (more common for Japanese
speakers). Further evidence comes from the En-
glish of non-native speakers, particularly those
whose native grammar does not mark countabil-
ity. Presumably, their knowledge of the world
is just as complete as English native speakers,
but they tend to have di culty with the English
speci c conceptual encoding of countability.
In the next section (x 3) we describe the re-
sources we use to measure the predictability of
countability by meaning, and then describe our
experiment (x 4).
3 Resources
We use the  ve noun countability classes of
Bond et al. (1994), and the 2,710 seman-
tic classes used in the Japanese-to-English ma-
chine translation system ALT-J/E (Ikehara et
al., 1991). These are combined in the machine
translation lexicons, allowing us to quantify how
well semantic classes predict countability.
3.1 Semantic Transfer Dictionary
We use the common noun part of ALT-J/E’s
Japanese-to-English semantic transfer dictio-
nary. It contains 71,833 linked Japanese-
English pairs. A simpli ed example of the entry
for usagi \rabbit" is given in Figure 1. Each
record of the dictionary has a Japanese index
form, a sense number, an English index form,
English syntactic information, English seman-
tic information, domain information and so on.
English syntactic information includes the part
of speech, noun countability preference, default
number, default article and whether the noun
is inherently possessed. The semantic informa-
tion includes common and proper noun seman-
tic classes. In this example, there are two se-
mantic classes: animal subsumed by living
thing, and meat subsumed by foodstu .
Because the dictionary was developed for
a Japanese-to-English machine translation sys-
tem, many of the English translations are longer
than the Japanese source terms: many concepts
encoded in a single lexical item in Japanese may
need multiple words in English. Of the 71,833
entries, 41,285 are multi-word expressions in
English (57.4%).
3.2 Semantic Ontology
ALT-J/E’s ontology classi es concepts to use
in expressing relationships between words. The
meanings of common nouns are given in terms of
a semantic hierarchy of 2,710 nodes. Each node
in the hierarchy represents a semantic class.
Edges in the hierarchy represent is-a or has-
a relationships, so that the child of a semantic
class related by an is-a relation is subsumed
by it. For example, organ is-a body-part.
The semantic hierarchy and the Japanese dic-
tionary marked with it have been published as
Goi-Taikei: A Japanese Lexicon (Ikehara et al.,
1997).
The semantic classes are primarily used to
distinguish between word-senses using the se-
lectional restrictions which predicates place on
their arguments. Countability has not been
used as a criterion in deciding which word
should go into which class. In fact, because
the dictionary has been built mainly by native
Japanese speakers, who do not have reliable in-
tuitions on countability, it was not possible to
use countability to help decide into which class
to put a given word.
Although the dictionary has been extensively
used in a machine translation system, errors still
exist. A detailed examination of user dictionar-
ies with the same information content, made by
the same lexicographers who built the lexicon,
found errors in 11{21% of the entries (Ikehara
et al., 1995). A particularly common source of
errors was words being placed one level too high
or low in the hierarchy. The same study found
that 90% of words entered into a user dictio-
nary could be automatically assigned to lexical
classes with 13{25% errors, although words were
assigned to too many semantic classes 32{56%
of the time (the range in errors is due to di er-
ent results from di erent domains: newspapers
and software manuals).
3.3 Noun Countability Preferences
Nouns in the dictionary are marked with one
of  ve major countability preference classes:
fully countable, strongly countable,
weakly countable, uncountable and plural
only, described at length in Bond (2001).
In addition to countability, default values
for number and classi er (cl) are also part
of the lexicon. The classes and additional
features are summarized in Table 1, along with
their distribution in ALT-J/E’s common noun
dictionary.1 The most common NCP is fully
countable, followed by uncountable.
The two most basic types are fully
countable and uncountable. Fully countable
nouns such as knife have both singular and plu-
ral forms, and cannot be used with determiners
such as much, little, a little, less and overmuch.
Uncountable nouns, such as furniture, have no
plural form, and can be used with much.
Between these two extremes there are a vast
number of nouns, such as cake, that can be
used in both countable and uncountable noun
phrases. They have both singular and plu-
ral forms, and can also be used with much.
Whether such nouns will be used countably or
uncountably depends on whether their refer-
ent is being thought of as made up of discrete
units or not. As it is not always possible to
determine this explicitly when translating from
Japanese to English, we divide these nouns into
two groups: strongly countable, those that
refer to discrete entities by default, such as cake,
and weakly countable, those that refer to non-
bounded referents by default, such as beer. At
present, these distinctions were made by the
lexicographers’ intuition, as there are no large
sense-tagged corpora to train from.
In fact, almost all English nouns can be used
in uncountable environments, for example, if
they are given the ground interpretation. The
only exception is classi ers such as piece or bit,
which refer to quanta, and thus have no un-
countable interpretation.
Language users are sensitive to relative fre-
quencies of variant forms and senses of lexi-
cal items (Briscoe and Copestake, 1999, p511).
The division into fully, strongly, weakly
1We ignore the two subclasses in this paper:
collective nouns are treated as fully countable and
semi-countable as uncountable.
2
66
66
66
66
66
4
Index usagi
sense 1
2
66
66
66
64
English Translation rabbit
Part of Speech noun
Noun Countability Pref. strongly countable
Default Number singular
Semantic Classes
h
common noun animal, meat
i
3
77
77
77
75
3
77
77
77
77
77
5
Figure 1: Japanese-English Noun Lexical Entry (usagi , rabbit)
Table 1: Noun Countability Preferences
Noun Countability Code Example Default Default # %
Preference Number Classi er
fully countable CO knife sg | 47,255 65.8
strongly countable BC cake sg | 3,110 4.3
weakly countable BU beer sg | 3,377 4.7
uncountable UC furniture sg piece 15,435 21.5
plural only PT scissors pl pair 2,107 2.9
and uncountable is, in e ect, as a coarse way
of re ecting this variation for noun countability.
The last major type of countability prefer-
ence is plural only: nouns that only have a
plural form, such as scissors. They can neither
be denumerated nor modi ed by much. plural
only are further divided depending on what
classi er they take. For example, pair plural
only nouns use pair as a classi er when they
are denumerated: a pair of scissors. This is
motivated by the shape of the referent: pair
plural only nouns are things that have a bi-
partite structure. Such words only use a sin-
gular form when used as modi ers (a scissor
movement). Other plural only such as clothes
use the plural form even as modi ers (a clothes
horse). In this case, the base (unin ected) form
is clothes, and the plural form is zero-derived
from it. The word clothes cannot be denumer-
ated at all. If clothes must be counted, then
a countable word of similar meaning is substi-
tuted, or clothing is used with a classi er: a
garment, a suit, a piece of clothing.
Information this detailed about noun count-
ability preferences is not found in standard
dictionaries. To enter this information into
the transfer lexicon, a single (Australian) En-
glish native speaker with some knowledge of
Japanese examined all of the entries in Goi-
Taikei’s common-noun dictionary and deter-
mined appropriate values for their countability
preferences.
4 Experiment and Results
To test how well the semantic classes predict the
countability preferences, we carried out a series
of experiments.
We ran the experiments under several condi-
tions, to test the e ect of combinations of se-
mantic classes and single-word or multi-word
entries. In all cases the baseline was to
give the most frequently occurring noun count-
ability preference (which was always fully
countable).
In the experiments, we use  ve NCPs (fully,
strongly, weakly countable, uncountable
and plural only), we do not consider default
number in any of the experiments.
For each combination of semantic classes
in the lexicon, we calculated the most com-
mon NCP. Ties are resolved as follows: fully
countable beats strongly countable beats
weakly countable beats uncountable beats
plural only. For example, consider the se-
mantic class 910:tableware with four mem-
bers: shokki , tableware (UC), youshokki ,
dinner set (CO), youshokki , Western-style
tableware (UC) and toukirui , crockery (UC).
Conditions Entries % Range Baseline
Training=Test all 77.9 76.8{78.6 65.8
Tenfold Cross Validation all 71.2 69.8{72.1 65.8
Tenfold Cross Validation single-word 66.6 65.6{67.7 58.6
Tenfold Cross Validation multi-word 74.8 73.9{75.8 71.1
Table 2: Results
The most common NCP is UC, so the NCP as-
sociated with this class is uncountable.
In our  rst experiment, we calculated the per-
centage of entries whose NCP was the same
as the most common one. For example,
the NCP associated with the semantic class
910:tableware is uncountable. This is correct
for three out of the four words in this semantic
class. This is equivalent to testing on the train-
ing data, and gives a measure of how well se-
mantic classes actually predict noun countabil-
ity in ALT-J/E’s lexicon: 77.9% of the time.
This is better than the base-line of all fully
countable which would give 65.8%. All the re-
sults are presented in Table 2.
In order to test how useful countability would
be in predicting the countability of unknown
words, we tested the system using strati ed
ten-fold cross validation. That is, we divided
the common noun dictionary into ten sets, then
tested on each set in turn, with the other nine-
tenths of the data used as the training set. In
order to ensure an even distribution, the data
was strati ed by sorting according to semantic
class with every 10th item included in the same
set. If the combination of semantic classes was
not found in the test set, we took the count-
ability to be the overall most common NCP:
fully countable. This occurred 11.6% of the
time. Using only nine tenths of the data,
the accuracy went down to 71.2%, 5.4% above
the baseline. In this case the training set for
910:tableware will still always contain a ma-
jority of uncountable nouns, so it will be asso-
ciated with UC. This will be correct for all the
words in the class except youshokki , dinner
set (CO).
Finally, we divided the dictionary into single
and multiple word entries (looked at from the
English side) and re-tested. It was much harder
to predict countability for single words (66.6%)
than it was for multi-word expressions (74.8%).
We will discuss the reason for this in the next
section.
5 Discussion
The upper bound of 78% was lower than we
expected. There were some problems with
the granularity of the hierarchy. In English,
the class names of heterogeneous collections
of objects tend to be uncountable, while the
names of the actual objects are countable.
For example, the following terms are all hy-
ponyms of tableware in Wordnet (Fellbaum,
1998): cutlery, chopsticks, crockery, dishware,
dinnerware, glassware, glasswork, gold plate,
service, tea set, .... Most of the entries
are either uncountable, or multi-word expres-
sions headed by group classi ers, such as ser-
vice and set. The words below these classes
are almost all countable, with a sprinkling of
plural only (like tongs). Thus in the three
levels of the hierarchy, two are mainly un-
countable, and below that mainly countable.
However, ALT-J/E’s ontology only has two
levels here: 910:tableware has four daugh-
ters, all leaf nodes in the semantic hierarchy:
911:crockery, 912:cookware, 913:cutlery
and 914:tableware (other). The majority
NCPs for all four of these classes are fully
countable. The question arises as to whether
words such as cutlery should be in the upper or
lower level. Using countability as an additional
criterion for deciding which class to add a word
to makes the task more constrained, and there-
fore more consistent. In this case, we would add
cutlery to the parent node 910:tableware, on
the basis of its countability (or add a new layer
to the ontology).
Adding countability as a criterion would also
help to solve the problem of words being entered
in a class one level too high or too low, as noted
in Section 3.2.
We were resigned to getting almost all of the
pair plural only wrong, and we did, but they
amount to less than 3% of the total. Although
there are some functional similarities, such as
a large percentage of 820:clothes for the
lower body, it was more common to get one or
two in an otherwise large group, such as tongs in
the 913:cutlery class, which is overwhelmingly
fully countable. Because the major di eren-
tiator is physical shape, which is not included in
our semantic hierarchy, these words cannot be
learned by our method. This is another argu-
ment for the importance of representing phys-
ical shape so that it is accessible for linguistic
processing.
We had expected single word entries to be
easier to predict than multiple word entries, be-
cause of the lack of in uence of modi ers. How-
ever, the experiment showed the opposite. In-
vestigating the reason found that single word
entries tended to have more semantic classes per
word (1.38 vs 1.34) and more varied combina-
tions of semantic classes. This meant that there
were 5.1 entries per combination to train on for
the multi-word entries, but only 3.7 for the sin-
gle word entries. Therefore, it was harder to
train for the single word entries.
As can be seen in the case of tableware given
above, there were classes where the single-word
and multi-word expressions in the same seman-
tic class had di erent countabilities. Therefore,
even though there were fewer training exam-
ples, learning the NCPs di erently for single
and multi-word expressions and then combing
the results gave an improved score: 72.0%.
Finally, there were also substantial numbers
of genuine errors, such as a0a2a1a4a3a6a5a8a7a10a9 sofuto
kar a which has two translations soft colour and
soft collar. Their semantic classes should have
been hue and clothing respectively, but the
semantic labels were reversed. In this case the
countability preferences were correct, but the
semantic classes incorrect.
An initial analysis of the erroneous predic-
tions suggested that the upper bound with all
genuine errors in the lexicon removed would be
closer to 85% than 78%. We speculate that
this would be true for languages other than En-
glish because is not speci cally tuned to En-
glish, it was developed for Japanese analysis.
Unfortunately we do not have a large lexicon of
French, German or some other countable lan-
guage marked with the same ontology to test
on.
5.1 Further Work
First, we would like to look more at multi-
word expressions. There is a general trend
for the head of a multiword expression to de-
termine the overall countability, which we did
not exploit. Modi ers can also be informative,
particularly for quanti ed expressions such as
zasshoku , various colors whose English part
must be countable as it is explicitly denumer-
ated.
Second, we would like to investigate further
the relation between under-speci ed semantics
and countability. Words such as usagi , rab-
bit are marked with the semantic classes for
animal and meat, and the single NCP strongly
countable. It may be better to explicitly iden-
tify countability with the animal sense, and un-
countability with the meat sense. In this way,
we could learn NCPs for each semantic class
individually (ignoring plural only) and look
at ways of combining them, or of dynamically
assigning countability during sense disambigua-
tion. Learning NCPs for each class individually
could also help to predict NCPs for entries with
idiosyncratic combinations, for which training
data may not be found.
Finally, from a psycho-linguistic point of
view, it would be interesting to test whether un-
predictable countabilities (that is those words
whose countability is not motivated by their se-
mantic class) are in fact harder for non-native
speakers to use, and more likely to be translated
incorrectly by humans.
5.2 Applications
In general, many errors in countability that had
been overlooked by the lexicographers in the
original compilation of the lexicon and its subse-
quent revisions became obvious when looking at
the words grouped by semantic class and noun
countability preference. Most entries were made
by Japanese native speakers, who do not make
countability distinctions. They were checked
by a native speaker of English, who in turn
did not always understand the Japanese source
word, and thus was unable to identify the cor-
rect sense.
Adding a checker to the dictionary tools,
which warns if the semantic class does not pre-
dict the assigned countability, would help to
avoid such errors. Such a tool could also be
used for  ne tuning the position of words in the
hierarchy, and spotting  at-out errors.
Another application of these results is in au-
tomatically predicting the countability of un-
known words. It is possible to automatically
predict semantic classes up to 80% of the time
(Ikehara et al., 1995). These semantic classes
could then be used to predict the countability
at a level substantially above the baseline.
6 Conclusions
Even with a limited ontology and noisy lexicon,
semantics does predict countability around 78%
of the time. Therefore countability is shown to
correlate with semantics. This semantic motiva-
tion can be used to build tools to (a) automat-
ically predict countability for unknown words,
and (b) serve as a check on consistency when
building a dictionary.
Acknowledgments
The authors would like to thank the other mem-
bers of the NTT Machine Translation Research
Group, as well as the NTT Linguistic Media
Group, Timothy Baldwin and Ann Copestake.
This research was supported by the research
collaboration between the NTT Communica-
tion Science Laboratories, Nippon Telegraph
and Telephone Corporation and CSLI, Stanford
University.

References

Keith Allan. 1980. Nouns and countability. Lan-
guage, 56(3):541{67.

Francis Bond and Satoru Ikehara. 1996. When and
how to disambiguate? | countability in machine
translation |. In International Seminar on Mul-
timodal Interactive Disambiguation: MIDDIM-
96, pages 29{40, Grenoble. (Reprint of MIDDIM-
1996).

Francis Bond and Kentaro Ogura. 1998. Reference
in Japanese-to-English machine translation. Ma-
chine Translation, 13(2{3):107{134.

Francis Bond and Kyonghee Paik. 1997. Classifying
correspondence in Japanese and Korean. In 3rd
Paci c Association for Computational Linguistics
Conference: PACLING-97, pages 58{67. Meisei
University, Tokyo, Japan.

Francis Bond, Kentaro Ogura, and Satoru Ikehara.
1994. Countability and number in Japanese-to-
English machine translation. In 15th Interna-
tional Conference on Computational Linguistics:
COLING-94, pages 32{38, Kyoto. (http://xxx.
lanl.gov/abs/cmp-lg/9511001).

Francis Bond. 2001. Determiners and Number in
English contrasted with Japanese | as exempli ed
in Machine Translation. Ph.D. thesis, University
of Queensland, Brisbane, Australia.

Ted Briscoe and Ann Copestake. 1999. Lexical
rules in constraint-based grammars. Computa-
tional Linguistics, 25(4):487{526.

Ann Copestake. 1992. The Representation of Lexi-
cal Semantic Information. Ph.D. thesis, Univer-
sity of Sussex, Brighton.

Terumasa Ehara and Hozumi Tanaka. 1993.
Kikaihonyaku-ni-okeru shizengengo shori (natu-
ral language processing in machine translation).
Journal of Information Processing Society of
Japan, 34(10):1266{1273. (in Japanese).

Christine Fellbaum, editor. 1998. WordNet: An
Electronic Lexical Database. MIT Press.

Satoru Ikehara, Satoshi Shirai, Akio Yokoo, and Hi-
romi Nakaiwa. 1991. Toward an MT system with-
out pre-editing { e ects of new methods in ALT-
J/E{. In Third Machine Translation Summit:
MT Summit III, pages 101{106, Washington DC.
(http://xxx.lanl.gov/abs/cmp-lg/9510008).

Satoru Ikehara, Satoshi Shirai, Akio Yokoo, Fran-
cis Bond, and Yoshie Omi. 1995. Automatic de-
termination of semantic attributes for user de-
 ned words in Japanese-to-English machine trans-
lation. Journal of Natural Language Processing,
2(1):3{17. (in Japanese).

Satoru Ikehara, Masahiro Miyazaki, Satoshi Shirai,
Akio Yokoo, Hiromi Nakaiwa, Kentaro Ogura,
Yoshifumi Ooyama, and Yoshihiko Hayashi. 1997.
Goi-Taikei | A Japanese Lexicon. Iwanami
Shoten, Tokyo. 5 volumes/CDROM.

Mutsumi Imai and Dedre Gentner. 1997. A crosslin-
guistic study of early word meaning: Univer-
sal ontology and linguistic in uence. Cognition,
62:169{200.

Ray Jackendo . 1991. Parts and boundaries. In
Beth Levin and Steven Pinker, editors, Lexical
and Conceptual Semantics, pages 1{45. Blackwell
Publishers, Cambridge, MA & Oxford, UK.

Kazumi Kawamura, Yasuhiro Katagiri, and
Masahiro Miyazaki. 1995. Multi-dimensional
thesaurus wth various facets,. In IEICE Technical
Report NLC94-48, pages 33{40. (in Japanese).

Melanie Siegel. 1996. De niteness and number
in Japanese to German machine translation. In
D. Gibbon, editor, Natural Language Processing
and Speech Technology, pages 137{142. Mouton
de Gruyter, Berlin.

Anna Wierzbicka. 1988. The Semantics of Gram-
mar. John Benjamins, Amsterdam.
