LIN 1180 Semantics Lecture7 Ambiguity and vagueness Continuation from last week Ambiguity vs Vagueness I In context a word can seem to have several distinct senses Some may appear more related than others ID: 418438
Download Presentation The PPT/PDF document "Albert Gatt" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Albert Gatt
LIN 1180 – Semantics
Lecture7Slide2
Ambiguity and vagueness
Continuation from last weekSlide3
Ambiguity vs. Vagueness (I)
In context, a word can seem to have several distinct senses. Some may appear more related than others.
In
our example:
run
1
= physical act of running
run
2
= place where fowl are kept
So
run is
2-ways ambiguous
(2 senses)
But
run
1
exhibits
vagueness
between a general sense of running, and the more specialised sense used in cricket.Slide4
Ambiguity vs. vagueness (II)
Similarly:
daħla
1
= entrance or inlet
daħla
2
= introduction to a text
2-ways
ambiguous
daħla
1
is
vague
between the sense of “entrance” and that of “inlet”Slide5
Ambiguity vs. vagueness (IV
)
Ambiguity:
In this case, the context will select one of the meanings/senses
We often don’t even notice ambiguity, because context clarifies the intended meaning.
Vagueness
:
Context adds information to the sense.
Therefore the sense of the word itself doesn’t contain all the information.
It is
underspecified
.Slide6
Tests for ambiguity and vagueness
There
are some tests to decide whether meaning distinctions involve ambiguity or vagueness.
The
do-so
test of meaning identity
The synonymy or sense-relations testSlide7
The
do-so
test: preliminary example
I
ate a sandwich
and Mary
The
do-so
construction is interpreted as
identical to the preceding verb phrase
Similar constructions in Maltese:
Kilt biċċa ħobż u
anka
MarijaKilt biċċa ħobż u Marija għamlet hekk ukoll.
did so toodid tooSlide8
The
do-so
test and meaning identity
Main principle:
if a particular sense is selected for a word in a verb phrase, it will also be the same sense in the
do-so
phrase
Therefore, very useful to test if two meanings are two distinct senses.Slide9
Do-so examples
Lili
għoġbitni
d-daħla
u lil Jimmy
wkoll
(I liked the entrance/introduction and so did Jimmy)
Suppose
daħla
here = “introduction”
Is it possible that
I liked the introduction and Jimmy liked the entrance
?
If not, then these are two distinct senses or daħlaI made
a run and so did PriscillaIf “I made a run” = “I ran”, then Priscilla cannot have made a run for her chickens...So, again, these are two distinct senses of run.Slide10
The sense relations test
Basic principle:
Words exhibit synonymy or similarity of meaning to other words.
Therefore, if a word is ambiguous, we can substitute it for a similar word in the same context, and see if the meaning
stays
roughly
the same.Slide11
Sense relations examples
Recall:
run
1
= physical act of running (
similar word:
jog
)
run
2
= a closed space for animals (
similar word:
enclosure)Pete went for .We can’t substitute one set of words for another and still keep the same meaning.
√ a run
√
a jog
*an enclosureSlide12
Lexical relations: basic concepts
LIN 1180 -- Semantics
We have established that:
words in the lexicon
can
have multiple senses (
ambiguity
)
they can also be
vague
, so that the actual meaning is underspecified and becomes clearer in context
In addition:
Words are
not merely listedthey are often related to one anotherSlide13
Homonymy,
polysemy
, synonymy
Part 1Slide14
How is the lexicon structured?
LIN 1180 -- Semantics
Lexical items belong to semantic fields
words that belong to the same “topic” ,“subject” or “usage”
lexical relations are often strongest within a semantic field
different senses of a word often fall into different fields
Examples
:
computing
:
gigabyte, CPU, memory, disk,
monitor
administration/diplomacy/politics:
green, monitor, parliament, electionNotice that monitor here has two senses, each falling in a different field.Slide15
Homonymy -- I
LIN 1180 -- Semantics
Homonyms are
unrelated senses of a the same phonological or orthographic word
.
sometimes we use
homographs
for unrelated senses of a written word
could be considered different words
lexicographers often treat derivationally related forms as homonyms
Examples
:
bank (river) / bank (financial)
ring / wring
house (N) / house (V)right / writeSlide16
Two subtypes
of homonymy
LIN 1180 -- Semantics
homphony
ring / wring
same phonology
different
orthography
homography
articulate (ADJ) / articulate (V)
Maltese:
domna
(V) (stay-late.3PL) /
domna (N) (religious icon)
different phonologysame orthographySlide17
Polysemy
LIN 1180 -- Semantics
One phonological word, multiple senses
(ambiguity)
senses are related, though distinguishable
cf.
daħla
(entrance) vs.
daħla
(in
let
)
in traditional dictionaries, multiple senses are listed under the same head word.Slide18
Homonymy vs. polysemy
LIN 1180 -- Semantics
Relatedness:
homonymy: senses are unrelated;
polysemy: senses are related
either historically or
based on speaker intuition
NB
: Not always a clear-cut distinction. Speakers’ intuitions vary considerably.
Do you consider
sole
(“bottom of foot”) and
sole
(“flat, riverbed fish”) related?Slide19
Synonymy
LIN 1180 -- Semantics
Different phonological words with
highly related meanings
:
sofa / couch
boy / lad
żgħir (small) / ċkejken (little)
moxt
(comb) /
petne
(comb)
Very
very difficult to find examples of perfect synonyms.Slide20
Imperfect synonymy
LIN 1180 -- Semantics
Synonyms often exhibit slight differences, espcially in connotations
petne
(“
comb
”)
has Romance origins; probably used by most speakers today
moxt
(“
comb
”)
has Semitic origins (cf.
xuxa “hair”)Usage differs depending on dialect, context…Slide21
The importance of register
LIN 1180 -- Semantics
With near-synonyms, there are often register-governed conditions of use.
Register = a style of language specific to a situation (e.g. formal, colloquial etc)
E.g
.
naive
vs
gullible
vs
ingenuous
gullible
/
naive seem critical, or even offensiveingenuous more likely in a formal contextSlide22
Synonymy vs. Similarity
LIN 1180 -- Semantics
Native speakers often have strong intuitions about words which are “related”, though not necessarily identical, in meaning.
E.g
. boat/ship; car/truck; man/woman
But
also near-synonyms such as: snake/serpent
Similarity
is broader than synonymy, since even words with “opposite” or “antonymous” meanings can be judged as similar; e.g. large/smallSlide23
When are two words similar?
LIN 1180 -- Semantics
Contextual view of meaning (Wittgenstein, 1953…):
the meaning of linguistic expressions can be characterised by looking at how they are
used
two words are similar to the extent that they’re used in similar waysSlide24
Example: master/pupil
LIN 1180 -- Semantics
These words have very different meanings, but share a core set of uses.
Both refer to human roles which tend to be practised in the same real world contexts (school etc).
Is
this reflected in the way we use the words?
master of X school, pupil of X school
past master, past pupil
…
Rather
than in contextual terms, we could view similarity as simply arising from links in a network of concepts.Slide25
Opposites and antonymy
Part 2Slide26
Semantic opposition
LIN 1180 -- Semantics
Traditionally,
antonyms
are words which are opposite in meaning.
dead – alive
We can find other kinds of opposition:
hot – cold
explode – implode
writer – reader, employer – employee
black – white, red – orange
(?)Slide27
Simple vs Gradable antonyms
LIN 1180 -- Semantics
Simple antonyms:
dead – alive
,
hit – miss
truth of one implies falsity of the other
?
X is dead but he’s alive
.
Gradable antonyms:
hot – cold
,
big – small both may be “false”: neither tall nor shorttypically, many terms to express gradations:
hot >> warm >> tepid >> cool >> coldoften modifiable with intensifiers:very hot, somewhat cold
exhibit global dependencies: If we say X is big, we mean “big for an object of type X”
big elephant
is much bigger than a
big mouseSlide28
Reverses and converses
LIN 1180 -- Semantics
Reverses:
explode – implode
a kind of opposition where one terms “reverses” the other.
often found with terms related to movement (
go/come
, etc)
Converses:
employer – employee, own – belong to
describe a relation between two entities from different viewpoints
“complement eachother”
if X is Y’s employer, then Y is X’s employeeif X owns Y, then Y belongs to XSlide29
Taxonomies
LIN 1180 -- Semantics
Colour
red
orange
yellow
green
blue
Taxonomies are classification systems, often in the form of a tree.
Sisters are elements at the same level.Slide30
Taxonomic sisters
LIN 1180 -- Semantics
Usually taken to be complementary or “opposed” or “incompatible”
or “mutually exclusive”
NB: Taxonomies are often our way of imposing a discrete categorisation on a continuum (e.g. colour).Slide31
Opposites and similarity
LIN 1180 -- Semantics
To many native speakers, the most highly related word to an adjective is its antonym or opposite.
also typical of taxonomic sisters
does this mean that opposites are synonymous?
No! It just means that “similarity” under the contextual view is much broader than synonymy.Slide32
Hyponymy and other relations
Part 3Slide33
Definition of hyponymy
LIN 1180 -- Semantics
Hyponymy is a
relation of inclusion
.
Arrows can be interpreted as “IS-A” relations.
Unlike taxonomic sisterhood, which is horizontal, hyponymy is vertical.
ANIMAL
MAMMAL
BIRD
CANARY
SPARROWSlide34
Elements of hyponymy
LIN 1180 -- Semantics
If
Y IS-A X
then:
X is the
superordinate
or
hypernym
of Y
Y is a
subordinate
or
hyponym of Xe.g. HUMAN is the hypernym of MAN, TOOL is the hypernym of CHAINSAWInclusion:if Y is a hyponym of X then Y contains the meaning of X (plus something extra)e.g. MAN includes all the features of HUMAN, plus the specification of ADULT and MALE.
Transitivity:if X IS-A Y and Y IS-A Z, then X IS-A ZSlide35
Transitivity -- illustration
LIN 1180 -- Semantics
A CANARY IS-A BIRD
A BIRD IS-A ANIMAL
Therefore, a CANARY IS-A ANIMAL
ANIMAL
BIRD
SPARROW
CANARY
MAMMALSlide36
Special cases of taxonomic relations
LIN 1180 -- Semantics
Sometimes, language exhibits special cases of relations that are:
well-established and lexicalised
seem to depend on an underlying taxonomy or hierarchy
ADULT-YOUNG
dog – puppy, duck – duckling, etc
MALE-FEMALE
woman – man, dog – bitch, drake – duck, etc
NB: These pairs are often
asymmetric
. The unmarked case in the MALE-FEMALE is the MALE.
We tend to use it for the name of the species.Slide37
Meronymy or part-whole
LIN 1180 -- Semantics
A different kind of taxonomic relationship. Arrows are interpreted as “
HAS-A
”
ANIMAL
BIRD
WING
LEG
HAS-A
IS-A
HAS-ASlide38
Meronymy vs. Hyponymy
LIN 1180 -- Semantics
Meronymy tends to be less regular than hyponymy:
NOSE is perceived as a necessary part of a FACE
CELLAR may be part of HOUSE, but not necessarily
Meronymy need not be transitive:
If X HAS-A Y and Y HAS-A Z, it does not follow that Y HAS-A Z
window HAS-A pane
room HAS-A window
??room HAS-A pane
Common-sense knowledge plays a very important role in acceptability of these relations.Slide39
Member-collection relations
LIN 1180 -- Semantics
We often lexicalise names of collections of specific things:
flotta (fleet) : a collection of ships
merħla (flock): a collection of sheep
Native
speakers know there is a member-collection relation:
flotta (fleet) – vapur (ship)
armata (army) – suldat (soldier)
merħla (flock) – nagħġa (sheep)
Can
be viewed as a special, lexicalised case of meronymy.Slide40
Are collections singular or plural?
LIN 1180 -- Semantics
In many languages, there is the possibility of switching from:
a view of a collection as a single entity vs. the “contents” of the collection as a group or set
English:
The band played well tonight.
It drove the crowd nuts
[SG]
They drove the crowd nuts
[PL]
Maltese:
L-armata rtirat
(The army retreated.SG)?L-armata rtiraw. (The army retreated.PL)Perhaps not as acceptable? Only with some nouns?Slide41
Portion-mass
LIN 1180 -- Semantics
Mass nouns:
nouns denoting things which have no units
noun is also true of portions of the substance
liquid, coal, hair
Langu
a
ges often have lexicalised concepts denoting portions of specific substances:
qatra
(drop) for liquids
strand
of hairSlide42
Summary
LIN 1180 -- Semantics
This lecture gave an overview of some standard ways to classify relations between lexical items.
homonymy vs. polysemy
synonymy (and contextual similarity)
taxonomic relations: part-whole and hyponymy