HG2002 Semantics and Pragmatics

Word Meaning

Francis Bond
Division of Linguistics and Multilingual Studies
http://www3.ntu.edu.sg/home/fcbond/
bond@ieee.org

Lecture 3
https://bond-lab.github.io/Semantics-and-Pragmatics/
Creative Commons Attribution License: you are free to share and adapt
as long as you give appropriate credit and add no additional restrictions:
https://creativecommons.org/licenses/by/4.0/.

HG2002 (2021)
 Overview

ã Revision: Meaning, Thought and Reality

â Reference as a Theory of Meaning
â Deixis
â Mental Representations
â Words, Concepts and Thinking
ã Defining word
ã Problems with defining word meaning
ã Lexical Relations
â Wordnet
ã Derivational Relations
ã Lexical Universals
ã Next week: Chapter 4: Sentence Relations and Truth

HG2002 (2021) 1
 Revision:
Meaning, Thought and
Reality

HG2002 (2021) 2
 Referential View

Speaker Refer
Say
Expression Referent
Denote

Referential view is focused on direct relationships between

expressions (words, sentences) and things in the world (realist
view). (More in Chapter 10)

HG2002 (2021) 3
 Representational View

Refer Concept
Speaker
Denote Represent
Say
Expression Referent

Representational view is focused on how relationships between

expressions (words, sentences) and things in the world are medi-
ated by the mind (cognitive linguistics). (More in Chapters 9 and
11)

HG2002 (2021) 4
 Two types of naming

ã The description theory: Names are like short hands for de-
scriptions:
William Shakespeare = “the playwright who wrote Hamlet”

ã The causal theory: Names begin with some event of naming

(e.g. a christening) before becoming commonly accepted.
William Shakespeare = “the guy other people call William
Shakespeare”

HG2002 (2021) 5
 Mental Representations

ã Divide meaning into

â reference: the relation to the world
â sense: the rest of the meaning

ã Introduce concepts
â Represented by Necessary and Sufficient Conditions
â Prototypes
∗ Concepts are organized in groups around a prototype
∗ These have typical members (remembered as exemplars)
∗ prototypes have characteristic features
∗ Some categories (concepts) seem to be more psycholog-
ically basic than others: basic level categories

HG2002 (2021) 6
 What is Deixis

ã any linguistic element whose interpretation necessarily makes

reference to properties of the extra-linguistic context in which
they occur is deictic
Person relative to the speaker and addressee
Spatial Location demonstratives; …
Temporal Location tense; yesterday, today, tomorrow
Social relative to the social status: professor, you, uncle, boy

ã Discourse deixis: referring to a linguistic expression or chunk

of discourse

More than 90% of the declarative sentences people utter are

indexical in that they involve implicit references to the speaker,
addressee, time and/or place of utterance in expressions like first
and second person pronouns, demonstratives, tenses, and ad-
verbs like here, now, yesterday (Bar-Hillel 1954: 366).

HG2002 (2021) 7
 Spatial Deixis

ã Two (three) way systems (English, …)

proximal this here close to the speaker

distal that there far to the speaker
interrogative what where

ã Three (four) way systems (Japanese, …)

proximal kore “this” koko “here” close to speaker

medial sore “that” soko “there” close to addressee
distal are “that” asoko “over there” far from both
interrogative dore “which” doko “where”

ã Can decompose: this “this thing”, here “this place”

Interrogative often not included 8

 Person Deixis

ã Commonly a three way division

First Person Speaker I

Second Person Addressee you
Third Person Other he/she/it

ã Often combined with

â gender: he/she/it
â number: I/we, ’anta “you:m”, ’antumaa “you:dual”, ’antum
“you:m:pl”
(Arabic)
â inclusion: núy “we including you”, níi “we excluding you”
(Zayse)
â honorification: kimi “you:inferior”, anata “you:equal”,
don’t use pronouns for superiors: sensei “teacher”,
…(Japanese)

HG2002 (2021) 9
 Social Deixis

In European languages, a two-way choice in 2nd person

pronominal reference is known as the T/V distinction, based on
the French forms for “you”.

ã T/V distinctions in European languages

Familiar 2sg Polite 2sg

French tu vous
German du Sie
Spanish tú usted

ã Shift from asymmetric use showing power (superior uses

du; inferior uses vous) to symmetric use showing solidarity
(strangers use vous; intimates use du): typically the socially
superior person must invite the socially inferior person to use
the familiar form

HG2002 (2021) 10
 Linguistic Relativity

ã The language we think in makes some concepts easy to ex-

press, and some concepts hard

ã The idea behind linguistic relativity is that this will effect how
you think

ã Do we really think in language?

â We can think of things we don’t have words for
â Language under-specifies meaning

ã Maybe we store a more abstract representation

the language of thought or Mentalese

HG2002 (2021) 11
 Word Meaning

HG2002 (2021) 12
 Defining word

ã How many words are there in the following?

(1) He who laughs last laughs longest.

(2) If he is right and I am wrong, are we both in trouble?
(3) I’m gonna go.
(4) 他们结婚了 ta1men jie2hun1 le “they got married” (他
们结了婚)

ã Tokens: Individual instances of a class

ã Types: The class as a whole

HG2002 (2021) 13
ã Why do we need a definition for word?
Psychological reality People can divide language into words
Phonological contours People pronounce words as unit
Orthographic practice Many languages put spaces between
words (although this practice only began around 600 CE for
Latin, and did not spread to all European languages until as
late as the 1600s)
â Some put them between phrases (Korean)
â Some words include spaces New York, ad hoc

HG2002 (2021) 14
 Bloomfield’s grammatical definition

A word, then, is a free form, which does not consist entirely

of (two or more) lesser free forms; in brief, a word is a mini-
mum free form.

(Bloomfield 1984: p178)

In practice, the definition is somewhat task specific: it may

make more sense to talk of orthographic words, semantic
words or predicates, ….

HG2002 (2021) 15
 Problems with defining
word meaning

HG2002 (2021) 16
 Definitional Semantics

ã Standard lexicographic approach to lexical semantics:

semantics = the study of language meaning
tailor = a person whose occupation is making and altering
garments

ã Definitions are conventionally made up of;

â genus: what class the lexical item belongs to
â differentiae: what attributes distinguish it from other mem-
bers of that class

ã Often hard to understand if you don’t already know the mean-

ing!

HG2002 (2021) 17
 Definitional Semantics: pros and cons

ã Pros:
â familiarity (we are taught to use dictionaries)

ã Cons:
â subjectivity in sense granularity (splitters vs. lumpers) and def-
inition specificity
â circularity in definitions
â consistency, reproducibility, …
â often focus on diachronic (historical) rather than synchronic
(current) semantics

HG2002 (2021) 18
 Starting at the Beginning ...

ã Lexical semantics is concerned with the identification and rep-

resentation of the semantics of lexical items

ã If we are to identify the semantics of lexical items, we have to

be prepared for the eventuality of a given word having multiple
interpretations
â Polysemy: having multiple meanings
â Monosemy: having only one meaning

ã Homonyms are words with two unrelated meanings:

â homographs: same spelling
bow vs bow; keep vs keep
â homophones: same pronunciation
right vs write; keep vs keep

19
 Distinguishing Polysemes

ã The polysemy of a word can be tested by a variety of means,

including:
â Antagonism: can the word be used in a sentence with mul-
tiple competing interpretations that are incompatible?
Kim can’t bear children
∗ Cannot have children
∗ Doesn’t like children
â Zeugma: can the word be used in a context where multiple
competing interpretations are simultaneously evoked?
Kim and her visa expired
∗ died
∗ ran out
â Paraphrase/Translation: Is there more than one (clearly
different) way to paraphrase/translate the word.

HG2002 (2021) 20
 Lexical Relations

HG2002 (2021) 21
 Words/Concepts are related in many ways

ã Hyponymy/Hypernemy

ã Synonymy

ã Antonymy (Opposites)

ã Meronymy
â Member-Collection
â Portion-Mass
â Element-Substance

ã Domain (lexical field)

HG2002 (2021) 22
 Hypernymy and Hyponymy

ã Hyponymy: X is a hyponym of Y iff f (X) entails f (Y ) but f (Y )

does not entail f (X) (for all or most f ):
Kim has a pet dog → Kim has a pet animal
Kim has a pet animal ̸→ Kim has a pet dog
N.B. complications with universal quantifiers and negation:
Kim likes all animals → Kim likes all dogs
Kim likes all dogs ̸→ Kim likes all animals

ã Hypernymy: Y is a hypernym of X iff X is a hyponym of Y

ã Can a word have multiple hypernyms?

(5) tank1 ⊂ military vehicle1; ⊂ tracked vehicle1; ⊂ ar-

mored vehicle1; ? ⊂ weapon1

HG2002 (2021) 23
 Properties of hypernymy/hyponymy
ã Asymmetric
ã applies only to lexical items of the same word class
ã applies at the sense level
ã Transitive: dog1 ⊂ mammal1 ⊂ animal1
ã Not all nodes are lexicalized
neutral (Hyper) male −balls female child
sheep ram wether ewe lamb
cow bull steer cow calf
goose gander goose gosling
snake
horse stallion gelding mare foal: colt/filly
ã Can you do this for pig, cat or chicken? ?
ã Can you give an example of this in another language? ?

HG2002 (2021) 24
 Synonymy

ã Propositional synonymy: X is a propositional synonym of Y

if
â (i) X and Y are syntactically identical,
â (ii) substitution of Y for X in a declarative sentence doesn’t
change its truth conditions
e.g., violin and fiddle

ã Why propositional synonymy is over-restrictive:

â syntactic identity (cf. eat and devour)
â collocations (cf. cemetery and graveyard)
â gradability (cf. sofa/settee vs. boundary/frontier)

HG2002 (2021) 25
 Near Synonymy

ã Synonyms are substitutable in some/most rather than all con-

texts

ã Synonymy via semantics: synonyms share “common traits” or

attributional overlap, walking the fine line between “necessary
resemblances” and “permissible differences”:
grain vs. granule; green vs. purple; alsation vs.
spaniel

HG2002 (2021) 26
ã Permissible differentiation via clarification:
Here is a grain, or granule, of the substance.
* The cover is green, {or,that is to say} purple.
and contrast:
Here is a grain or, more exactly, granule
* He likes alsations, or more exactly, spaniels

HG2002 (2021) 27
 Properties of synonymy

ã Symmetric

ã traditionally applies only to lexical items of the same word class

but what about
â can vs be able to
â immediately vs at once

ã applied at the sense or lexical item-level?

ã ≈ converse of polysemy

HG2002 (2021) 28
 Antonymy (opposites)

ã Give me some new examples of each ?

ã Simple antonyms: the negative of one implies the positive of

the other.

(6) dead/alive
(7) pass/fail

ã Gradable Antonyms: points along a scale

(8) boiling/hot/warm/tepid/cool/cold/freezing
(9) like HG2002/fascinating/interesting/dull/boring/

ã Reverses: reverse the direction of a motion

(10) ascend/descend

HG2002 (2021) 29
 (11) up/down; right/left

ã Converses: the same act from different points of view

(12) above/below; right/left

(13) employer/employee

(Slightly non-standard usage)

ã Taxonomic Sisters: children of the same (grand)parent

(14) Monday/Tuesday/…/Sunday
in WordNet: day of the week ⊃ weekday, weekend
(15) LMS/English/Chinese/…
Context dependent

HG2002 (2021) 30
 Meronymy

ã Meronomy refers to the part-whole relation

â meronym is the part
â holonym is the whole
car
wheel engine door steering wheel
tire rim piston valve

ã It is not always transitive shirt

button
button hole

But we don’t normally say that a button hole is part of a shirt.

HG2002 (2021) 31
 Member-Collection

ã The relation between a collection and one of the units that

makes it up

(16) tree–forest
(17) sheep–flock
(18) fish–school
(19) book–library
(20) member–band
(21) musician–orchestra
(22) student–class

HG2002 (2021) 32
 Portion-Mass

ã The relation between a mass noun and a typical unit of mea-

surement

(23) drop–liquid
(24) grain–sand/salt/truth
(25) sheet/ream–paper
(26) lump–coal (or just about anything)
(27) strand–hair
(28) rasher–bacon

ã Similar to classifiers in many ways, e.g. in Malay

(29) ekor “tail”–animal

(30) orang “human”–person

HG2002 (2021) 33
 Domain (lexical field)

The domain in which a word is typically used with this meaning.

(31) driver1 — the operator of a motor vehicle

(32) driver2 — someone who drives animals that pull a vehicle
(33) driver3 — a golfer who hits the golf ball with a driver [GOLF]
(34) driver4 — (≃ device driver) a program that determines
how a computer will communicate with a peripheral de-
vice [COMPUTER SCIENCE]
(35) driver5 — (≃ number one wood) a golf club (a wood) with
a near vertical face that is used for hitting long shots from
the tee [GOLF]

Some GOLF terms: approach9, approach shot1, golf course1,

links course1, wedge5, tee1, scratch9, putt1, slice1, hook1

Examples from WordNet 3.0 34

 And More

ã There are many, many more lexical relations advocated by var-

ious theories including:
â Troponymy/hypernymy (cf. walk vs. lollop) “way of doing
something”
â Entailment (cf. snore vs. sleep) “if you do one thing, you must
be doing the other”
â Operator (cf. question vs. ask) “the thing you do by doing
something”
â Magnifier (cf. wound vs. badly) “intensifier, diminisher”
â Usage (cf. strong-willed vs. pig-headed “stubborn”)
strong-willed is pejorative

35
 Wordnet

HG2002 (2021) 36
 WordNet

ã WordNet is an open-source electronic lexical database of En-

glish, developed at Princeton University
http://wordnet.princeton.edu/

ã Made up of four separate semantic nets, for each of nouns,

verbs, adjectives and adverbs

ã WordNets exist for many languages, at LMS we work on:

â Japanese
â Bahasa Malay/Indonesian
â Chinese
â Myanmar
â Kristang
â The shared open multi-lingual wordnet (34+ languages)
http://compling.hss.ntu.edu.sg/omw/

Miller (1998); Fellbaum (1998); Bond and Foster (2013) 37

 Wordnet Structure

ã Lexical items are categorised into ∼115K (and counting)

glossed synsets (= synonym sets)
1. enrichment -- (act of making fuller or more
meaningful or rewarding)
2. enrichment -- (a gift that significantly increases
the recipient's wealth)

ã Lexical relations at either the synset level or sense (= combi-

nation of lexical item and synset) level

ã Strongly lexicalist (orginally):

â synsets only where words exist
â but many multiword expressions (≈ 50%)

HG2002 (2021) 38
 Psycholinguistic Foundations of WordNet

ã Strong foundation on hypo/hypernymy (lexical inheritance)

based on
â response times to sentences such as:
a canary {can sing/fly,has skin}
a bird {can sing/fly,has skin}
an animal {can sing/fly,has skin}
â analysis of anaphora:
I gave Kim a novel but the {book,?product,...} bored her
Kim got a new car. It has shiny {wheels,?wheel nuts,...}
â selectional restrictions

ã Is now often used to calculate semantic similarity

â The shorter the path between two synsets the more similar
they are
â Or the shorter the path to the nearest shared hypernym, …

39
 Word Meaning as a Graph

entity

object process substance abstract

artifact agent move manipulate communication

equipment person propel control writing

golf-club worker swing operate code

driver driver drive drive driver

instrument agent

It ends up being a very big graph 40

 Wordnet in this course

ã We use wordnet to test our skills in determining word meaning

â tag a short text from this year’s story or stories
â discuss differences with other annotators

ã LMS students have used wordnets for:

â Japanese derivational relations (Bond and Wei, 2019)
â pronoun representation for Japanese, Mandarin and English
(Seah and Bond, 2014)
â exclamatives and classifiers (Mok et al., 2012; Morgado da
Costa and Bond, 2016)
â sentiment analysis (Le et al., 2016; Bond et al., 2019)
â cross-lingual sense annotation (Bonansinga and Bond,
2016)
â multilingual crosswords (Tan, 2012)
…

HG2002 (2021) 41
 Synonyms for a dead Parrot

be dead, be demised, be deceased, pass on,

be no more, cease to be, expire, go to meet
one’s maker, be a stiff, be bereft of life, rest
in peace, push up the daisies, one’s metabolic
processes are now history, be off the twig,
kicked the bucket, shuffle off this mortal coil,
ring down the curtain, join the choir invisible,
be an ex-parrot

From the “Dead Parrot Sketch”, also known as the “Pet Shop
Sketch” or “Parrot Sketch”, originally in Monty Python’s Flying Cir-
cus, first performed in the eighth episode of the show’s first series,
”Full Frontal Nudity” (7 December 1969).

HG2002 (2021) 42
 Derivational Relations

HG2002 (2021) 43
 Diathesis Alternations

ã Causative/inchoative alternation:
Kim broke the window ↔ The window broke
also the window is broken (state)

ã Middle construction alternation:

Kim cut the bread ↔ The bread cut easily

ã Conative alternation:
Kim hit the door ↔ Kim hit at the door

ã Body-part possessor ascension alternation:

Kim cut Sandy’s arm ↔ Kim cut Sandy on the arm

Levin (1993) 44
 Diathesis Alternations and Verb Classes

ã A verb’s (in)compatibility with different alternations is a strong

predictor of its lexical semantics:
break cut hit touch
Causative YES NO NO NO
Middle YES YES NO NO
Conative NO YES YES NO
Body-part NO YES YES YES

break = {break, chip, crack, crash, crush, ...}

cut = {chip, clip, cut, hack, hew, saw, ...}
hit = {bang, bash, batter, beat, bump, ...}
touch = {caress, graze, kiss, lick, nudge, ...}

Levin (1993) 45
ã Corollary: we can predict the syntax of novel words we are
given the semantic class for

ã The principal weakness of syntax-based verb classification is

that there are often subtle divergences in semantics between
synonyms (cf. eat vs. devour vs. gobble)

Levin (1993) 46
 Agentive Nouns

ã An agentive noun is a word that is typically derived from an-

other word denoting an action, and that identifies an entity that
does that action.
verb + -er, -or, -ant

(36) murderer, commentator, whaler, director, computer

(37) ?? undertaker, cooker, footballer (Saeed also in-
cludes these)

ã Should murderer be listed separately from murder in the dic-

tionary? Why or why not?

ã Also recipient nouns that show the undergoer: verb + -ee:

employee, trustee

47
 Agentive Nouns in Other Languages

ã Japanese (suffix distinguishes person/machine)

â 運転する → 運転者 unten-sha “driver”
â 計算する → 計算者 計算機 keisan-sha/ki “computer”
â 研究する → 研究者 研究員 kenkyuu-sha/in “researcher”
â 読む → 読み手 読者 yomite/dokusha “reader”

ã Malay (prefix can convert any part of speech)

â bantu (v) “help” → pembantu “assistant/helper”
â potong (v) “cut” → pemotong “cutter (human/machine)”
â terbang (v) “fly” → penerbang “pilot (not passenger)”
â gunting (n) “scissors” → penyunting “(editor – human)”

Thanks to Yeo Jia Qi (Malay) 48

 Agentive Nouns in Other Languages

ã Tamil, can convert verb or noun

â ேவைல vēlai “work” → ேவைலக்காரர் vēlaikkārar
“worker”
â சைமயல் samaiyal “cook” → சைமயல்காரர் samaiyalkārar
“chef”
â பாடல் pāl “song” → பாடகர் pālkārar “singer”
ã Endings can mark gender, similar to pronouns
â Singer
∗ பாடகன் pāṭagan (male)
∗ பாடக pāṭaki (female ≈ male + இ i )
∗ பாடகர் pāṭaka (formal)
â Pronouns
∗ அவன் avaṉ “he”
∗ அவள் avaḷ “she”
∗ அவர் Avar “they” (Formal/Gender-neutral)

Thanks to Shalini (Tamil) 49

 Lexical Universals

HG2002 (2021) 50
 Color Terms

ã Basic Color Terms

â Monolexemic
â Not a hyponym of any other color
â Can be widely applied
â Not derived from a noun

ã Focal Colors are related to the neurophysiology of our visual

system

ã Seem to come in an order  

{ } { } 
 PURPLE 

 
WHITE/DARK GREEN PINK
< RED < < BLUE < BROWN <
BLACK/LIGHT YELLOW 
 ORANGE 

 
GREY

See also: http://wals.info/chapter/133: Colour Terms by Paul Kay and Luisa Maffi 51
 Core Vocabulary

ã Some universal terms can be used to compare languages

â lexicostatistics (quantitative language relatedness assess-
ment)
â glottochronology (language divergence dating)

ã The Swadesh list, developed by Morris Swadesh from 1940

onward

ã Chosen for their universal, culturally independent availability in

as many languages as possible

en.wiktionary.org/wiki/Appendix:Swadesh_list 52
 I, You, we, this, that, who, what, not, all, many, one, two, big,
long, small, woman, man, person, fish, bird, dog, louse, tree,
seed, leaf, root, bark, skin, flesh, blood, bone, grease, egg, horn,
tail, feather, hair, head, ear, eye, nose, mouth, tooth, tongue,
claw, foot, knee, hand, belly, neck, breasts, heart, liver, drink, eat,
bite, see, hear, know, sleep, die, kill, swim, fly, walk, come, lie, sit,
stand, give, say, sun, moon, star, water, rain, stone, sand, earth,
cloud, smoke, fire, ash(es), burn, path, mountain, red, green, yel-
low, white, black, night, hot, cold, full, new, good, round, dry,
name

ã Available in many languages (hundreds);

ã Now linked to wordnet (Morgado da Costa et al., 2016)

en.wiktionary.org/wiki/Appendix:Swadesh_list 53
 Natural Semantic Meta Language

ã Try to define everything in terms of semantic primitives and re-

ductive paraphrase
â simple, indefinable, and universally lexicalized concepts
â breaking complex concepts down into simpler concepts

X feels unhappy=
sometimes a person thinks something like this:
something bad happened to me
I don’t want this
if I could, I would do something
because of this, this person feels something bad
X feels like this

ã Very hard to do consistently and reproducibly

Wierzbicka (1996) 54
 The Semantic Primitives

ã substantives: I, YOU, SOME- ã location, existence, posses-

ONE, PEOPLE, SOMETHING/THING, sion, specification: BE (SOME-
BODY WHERE), THERE IS, HAVE, BE
ã relational substantive: KIND, (SOMEONE/THING)
PART ã life and death: LIVE, DIE
ã determiners: THIS, THE SAME, ã time: WHEN/TIME, NOW, BE-
OTHER/ELSE FORE, AFTER, A LONG TIME, A
ã quantifiers: ONE, TWO, MUCH/MANY, SHORT TIME, FOR SOME TIME,
SOME, ALL MOMENT
ã evaluators: GOOD, BAD
ã space: WHERE/PLACE, HERE,
ã descriptors: BIG, SMALL
ABOVE, BELOW, FAR, NEAR,
ã mental predicates: THINK, SIDE, INSIDE
KNOW, WANT, FEEL, SEE,
ã logical concepts: NOT, MAYBE,
HEAR
CAN, BECAUSE, IF
ã speech: SAY, WORDS, TRUE
ã actions, events, movement, ã intensifier, augmentor: VERY,
contact: DO, HAPPEN, MOVE, MORE
TOUCH ã similarity: LIKE/WAY

HG2002 (2021) 55
 Acknowledgments and References

ã Definitions from WordNet: http://wordnet.princeton.edu/

ã Images from
â the Open Clip Art Library: http://openclipart.org/
â Steven Bird, Ewan Klein, and Edward Loper (2009) Natural
Language Processing with Python, O’Reilly Media
www.nltk.org/book

ã Video: Dead parrot sketch Monty Python

HG2002 (2021) 56
 *
References

Giulia Bonansinga and Francis Bond. 2016. Exploring cross-lingual sense

mapping in a multilingual parallel corpus. In Proceedings of the 8th Global
Wordnet Conference (GWC 2016), pages 45–49.

Francis Bond and Ryan Foster. 2013. Linking and extending an open mul-
tilingual wordnet. In 51st Annual Meeting of the Association for Com-
putational Linguistics: ACL-2013, pages 1352–1362. Sofia. URL http:
//aclweb.org/anthology/P13-1133.

Francis Bond, Arkadiusz Janz, and Maciej Piasecki. 2019. A comparison of

sense-level sentiment scores. In Proceedings of the 11th Global Wordnet
Conference (GWC 2019).

HG2002 (2021) 57
Francis Bond and Ryan Lim Dao Wei. 2019. Generating derivational relations
for the Japanese wordnet: The case of agentive nouns. In 2019 Pacific
Neighborhood Consortium Annual Conference and Joint Meetings (PNC),
pages 1–7.

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

MIT Press.

Tuan Anh Le, David Moeljadi, Yasuhide Miura, and Tomoko Ohkuma. 2016.
Sentiment analysis for low resource languages: A study on informal Indone-
sian tweets. In Proceedings of The 12th Workshop on Asian Language Re-
sources, page 123–131. Osaka.

Beth Levin. 1993. English Verb Classes and Alternations. University of Chicago
Press, Chicago, London.

George Miller. 1998. Foreword. In Fellbaum (1998), pages xv–xxii.

HG2002 (2021) 58
Hazel Shuwen Mok, Eshley Huini Gao, and Francis Bond. 2012. Generating
numeral classifiers in Chinese and Japanese. In Proceedings of the 6th
Global WordNet Conference (GWC 2012). Matsue. 211-218.

Luís Morgado da Costa and Francis Bond. 2016. Wow! what a useful extension
to wordnet! In 10th International Conference on Language Resources and
Evaluation (LREC 2016). Portorož.

Luís Morgado da Costa, Francis Bond, and František Kratochvíl. 2016. Linking
and disambiguating Swadesh lists: Expanding the open multilingual word-
net using open language resources. In GLOBALEX 2016: Lexicographic
Resources for Human Language Technology (Workshop at LREC 2016).
Portorož.

Yu Jie Seah and Francis Bond. 2014. Annotation of pronouns in a multilingual

corpus of Mandarin Chinese, English and Japanese. In 10th Joint ACL -
ISO Workshop on Interoperable Semantic Annotation. Reykjavik.

Jeanette Yi Wen Tan. 2012. Automatic Generation of Multilingual Crossword

HG2002 (2021) 59
 Puzzles with WordNet. Final year project, Linguistics and Multilingual Stud-
ies, Nanyang Technological University.

Anna Wierzbicka. 1996. Semantics: Primes and Universals. Oxford University

Press, Oxford.

HG2002 (2021) 60