2. Comprehension Questions
1.
Write the sense relationships which reflect the similarities between lexical
p
units.
2.
State the kinds of antonyms and explain each of them.
3.
Explain ‘frame semantics’ theory.
4.
What are the roles of the frames, frame elements and context?
5.
Which steps are involved during the analysis process carried out by using frame
semantics theory?
6.
What is the importance of using the frame semantics approach?
7.
Write the information types which are relevant to make a lexicographic record
of a word.
8.
What are the inherent properties of the headword?
9.
Why learning the properties of the source texts is sometimes needed?
3. 5.1
5 1 Preliminaries
A grounding in linguistic theory is not a prerequisite for being a proficient
lexicographer – still less a guarantee of success i th fi ld
l i
h
till l
t
f
in the field.
But there are certain basic linguistic concepts which are invaluable in preparing
people to analyse data and to produce concise, accurate dictionary entries.
An awareness of linguistic theory can help lexicographers to do their jobs more
effectively and with greater confidence.
This chapter reviews those linguistic theories which we have found to have direct
application to our work as dictionary planners and dictionary makers.
5.2 Sense relationships: similarities
This section summarizes different types of ‘similarity’ between lexical units:
o
Th
th t h
ti
t
ti (hyponym and synonymy)
d
)
Those that share some semantic property or properties (h
o
Those that denote a part-whole relationship between objects in the real world (meronymy)
o
Those that allow similar metaphorical sense extensions (regular polysemy)
5.2.1
5 2 1 Hyponymy
The nodes of this hierarchy are the ‘superordinate’ and the ‘hyponym’.
This relationship can be summarized as ‘if a hyponym then a superordinate’.
Its major significance for lexicographers is that the ‘genus expression’ (the central
genus expression
word or words) in a definition should ideally be the superordinate of the headword.
Hyponymy rule of thumb: X is a Y but Y is not only an X (a terrier is a dog).
4. The hyponymy hierarchy is rarely found in adjectives, and consequently there is a real
lack of superordinates.
Cohyponyms rule of th b X and Y are both Zs ( rose and a tulip are both flowers).
C h
l f thumb:
d
b th Z (a
d t li
b th fl
)
5.2.2 Synonymy
Synonyms are words which have the same meaning like ‘pavement’ and ‘sidewalk’.
True synonyms are extremely rare; the nearest you get is usually a pseudo-synonymy and
l
h
ll
d
d
synonyms in dictionaries often turn out to be cohyponyms or superordinates.
Pure synonymy is rare across languages, except for the names of concrete objects which
the two cultures share.
Synonymy rule of thumb: X is Y and Y is X (shut is close and close is shut).
5.2.3 Meronymy
Meronymy reflects the relationship of the part to the whole and vice versa.
It’s difficult to define the part without mentioning the whole but the part is only
occasionally referred to in the definition of the whole.
Meronymy rule of thumb: X and other parts of Y.
Quasi-meronymy reflects the relationship of the member to the group or class of
people, or collection of objects.
This is a rather loose relationship: it’s difficult to word a formula appropriate to all the
cover.
varied Lus it should cover
Quasi-meronymy rule of thumb: X belongs to / in a Y.
5.2.4 Regular polysemy
Some polysemous words have a particular relationship with others i th i ‘l i l s t’ in
S
l s
s
ds h
ti l
l ti shi
ith th s in their ‘lexical set’ i
that several of their meanings seem to parallel each other.
5. Certain specific semantic components result in sets of words behaving lexicographically
in a very similar way. This is known as ‘regular polysemy’.
When you re planning the editorial work in a dictionary project it’s obciously a help to
you’re
project, it s
the team if you can list the major instances of regular polysemy, either by producing
template entries or simply by issuing lists of headwords related in this way.
English morphology encourages a wider range of regular polysemt than is found in
languages that have specific forms for verbs and nouns.
l
h h
ifi f
f
b
d
5.3 Sense relationships: differences
This section summarizes relationships between LUs that are in some way opposite in
h
l
h
b
L
h
meaning.
The kinds are complementary, polar and directional antonymy.
5.3.1
5 3 1 Complementary antonymy
This relationship is sometimes called ‘contradiction’.
Complementary antonymy rule of thumb: If it isn’t X then it must be Y and vice versa.
5 3 2 Polar antonymy
5.3.2
This relationship is similar to, but more complex than, complementarity.
There is a gradient between X and Y in polar antonymy. X and Y are at the poles of
gradient,
area,
this gradient but in between there is an indeterminate area where more X and less Y
are found.
Polar antonymy rule of thumb: If it’s X then it can’t be Y and vice versa, but it can be
somewhere in between.
5.3.3 Directional antonymy
Directional antonyms include various subtypes: some denote contrary movement or
position, for instance, pairs of words representing opposing poles along a shared axis.
6. 5.3.4 Converseness
Converseness holds between pairs of words which have a certain semantic symmetry, so
that although not antonyms one of the pair is felt in some way to be linked by
g
y
p
y
y
oppositeness to the other.
There is little direct application of converse pairs in dictionaries, but if a word is
difficult to define a look at its converse’s definition can be helpful.
5.4 Frame semantics
The application of this theory to practical lexicography results in the approach to
lexicographic relevance which helps lexicographers to identify useful facts in corpus
texts.
This theory which was produced by Fillmore describes words, their various meanings,
and how these are combined with others to form the utterances and sentences of a
language.
Its aim is to analyse and record for each sense of a word or phrase, the full range of
its semantic and syntactic relations.
To do this they have devised a suite of codes denoting semantic roles (frame elements)
this,
and grammatical relationships, which allow them to document in detail the corpus
contexts in which a word is found.
The work is computer-assisted.
The frame semantics approach to word behaviour is the most helpful and appropriate
approach to corpus data, ensuring that the analysis is correctly carried out, and no vital
fact is overlooked.
5.4.1
5 4 1 What are frames and frame elements
Frame semantics describes the meanings of words and phrases in terms of the frame to
which they belong and the contexts in which these LUs are found.
7. A semantic frame is a schematic representation of a situation type together with a list of the typical
participants, props, and concepts that are to be found in such a situation; these are the semantic
roles, or frame elements (Fes).
The context is normally the phrase or clause and maximally the sentence in which the target word
clause,
appears in corpus data.
5.4.2 How is the analysis done?
There are several distinct steps in the analysis process:
o First, the frame is defined, and its core elements named and described.
o Next, a list is made of as many words as can be found which in one of their senses evoke that
frame.
o Then for each sense or LU a set of corpus sentences is extracted, in which the word is used in
Then,
sense,
LU,
extracted
the particular sense.
o Each sentence is annotated by marking off any section which instantiates an FE and by recording
for each FE thus identified:
o Its phrase type
o Its grammatical function
5.4.3 Why is this useful for lexicographers?
The frame semantics approach, grounded in a coherent theory, offers the possibility of a more
systematic less subjective way of analysing corpus data and gives us confidence that all relevant
systematic,
data,
features are being captured.
5.5 Lexicographic relevance
When looking at the concordances the corpus offers for a word we have to determine ‘what is
lexicographically relevant’?
We consider lexicographic relevance from the standpoint of Fillmore’s frame semantics.
8. Three types of information are relevant to making a lexicographic record of a word:
o
What we know, as native speakers, about the headword (its inherent properties)
o
What we learn from its use in corpora and elsewhere (its contextual features)
o
What we know about where the citations came from (the properties of the source texts)
It’s important to remember that ‘lexicographic relevance’ relates to what is relevant to an
LU, and not to a lemma, the focus is the headword in one of its senses, not the whole word.
5.5.1 Inherent properties of the headword
This is the knowledge of our language that we all bring to analysing the corpus data and
writing the dictionary entry.
The properties of the headword that principally concern us can be summarized very briefly:
o
Its wordclass
o
Its wordforms
o
Its grammatical behaviour
o
Its semantics
In corpus lexicography we use our inherent knowledge of the headword rather to help us
discover the really useful facts in the corpus, and make sure the entry is comprehensive and
h
l
l
i
i
k
the examples are pleasing to our native speakers’’ ears.
Our knowledge of the headword’s inherent properties serves as quality control during our
work on corpus data, as we discover and record its contextual features in each of its LUs.
5.5.2 Contextual features of the headword
An understanding of lexicographic relevance helps you identify in a corpus sentence all the
essential components of the headword’s context, all the facts that you need to take into
account when writing any entry for that word.
9. 5.5.2.1 Case study: argue
The headword argue contains four LUs; each represents a distinct sense of the
headword.
headword
LU 1: the sense of ‘quarrel, dispute’, the communication frame.
LU 2: the sense of ‘maintain, make a case for’, the reasoning frame.
LU 3: the sense of ‘indicate’, the evidence frame
indicate
frame.
LU 4: the sense of ‘persuade’, the persuasion frame.
They will differ for each LU, since the frame elements depend on the frame the LU
belongs to.
ongs
participant 1
participant 2
topic
The sentence is ‘Sam / was arguing / with his brother / about the money’.
We have to examine its;
o Frame elements or semantic roles
o Phrase types
o Grammatical function
Frame elements:
Participant 1
Participant 2
Topic
Sam
with his brother
about the money
Phrase types + grammatical functions
NP: Subject / PP-with: Complement / PP-about: Complement
The phrase type information allows you to mark off in the sentence the actual sections
relevant to your description.
The information about grammatical function lets you assess the importance of the
component for your database.
10. The set of threefold descriptions of each component contains most of the
information you need to extract from this sentance for your database.
Whan we are analysing corpus data in an attempt to collect the facts about a
word for our dictionary entry, it’s important to be able to discover from the
concordances tha actual source of each citation.
This information is stored in the ‘document headers’ of each text in the corpus.
Using this information the computer can tell you whether a particular citation
comes mainly from spoken or written language, or political documents, or feminist
publications and so on.
If you h
have any d b about the way a word i used, i ’ useful to b able to
doubts b
h
d is
d it’s
f l
be bl
check up on where the citation came from.
To summarize lexicographic relevance: the wordclass of the word is central to
what is relevant to record, and there are lists of the principal co constituents of
record
co-constituents
a clause that are relevant for each of the four major wordclasses.
11. Turkish Summary
Bu bölümde sözlük oluşturan kişilerin çalışmalarına katkı sağlayabilecek bazı
ş
ş
ç ş
ğ y
dilbilimsel kuramlardan ve bu kuramların uygulanış biçimlerinden bahsedilmektedir.
Sözcükler arasında birtakım anlamsal ilişkiler vardır ve bu ilişkiler benzerlik ve
farklılıklara göre şekillenir. Sözcükler arasındaki benzerlikten doğan ilişkiler
sözcüklerin bazı anlambilimsel özellikleri paylaşmalarına nesneler arasındaki
paylaşmalarına,
parça-bütün ilişkilerine ve sözcüklerin benzer mecazi alt anlamları
çağrıştırmalarına göre farklılık gösterir. Sözcükler arasındaki farklılıklardan
doğan ilişkiler ise sözcüklerin birbirlerini zıtlıklarla tamamlamalarına, tamamen
farklı kavramları çağrıştırmalarına ve birbirlerine zıt olmalarına göre farklılık
gösterir. Dilbilimci Fillmore sözlükbilime de uygulanabilen ve ‘yapısal anlambilim’
olarak adlandırılan bir kuram geliştirmiştir. Bu kuram sözcükleri ve onların farklı
anlamlarını tasvir eder ve bir dildeki sözce ve tümcelerin oluşturulması için
ükl i diğ
ükl l
l birleştirilmesi gerektiğini açıklar. Bunu yapmak
i il
i
k iği i
kl
B
k
sözcüklerin diğer sözcüklerle nasıl bi l
için de anlambilimsel rolleri ve dilbilgisel ilişkileri gösteren bazı kodlar tasarlanır.
Bu kuramın sözlükbilime uygulanması sonucunda sözlükbilimcilere bütünce
metnindeki faydalı bilgileri tanımlama konusunda yardımcı olan sözlükbilimsel ilişki
y
g
y
ş
yaklaşımı ortaya çıkmıştır. Bir sözcüğün sözlükbilimsel ilişkilerini belirlemek için
sözcüğün genel özelliklerinin, bağlamsal özelliklerinin ve kaynak metinlerin
özelliklerinin bilinmesi gereklidir.
12. Comprehension Questions / Answers
Write the sense relationships which reflect the similarities between lexical
units.
Different types of ‘similarity’ between lexical units:
o
Those that share some semantic property or properties (hyponym and synonymy)
o
Those that denote a part-whole relationship between objects in the real world (meronymy)
o
Those that allow similar metaphorical sense extensions (regular polysemy)
State the kinds of antonyms and explain each of them.
The kinds are complementary, polar and directional antonymy.
Complementary antonymy
This relationship is sometimes called ‘contradiction’.
Complementary antonymy rule of thumb: If it isn’t X then it must be Y and vice versa.
P l antonymy
t
Polar
This relationship is similar to, but more complex than, complementarity.
There is a gradient between X and Y in polar antonymy. X and Y are at the poles of this
gra nt, ut n tw n th r s
gradient, but in between there is an indeterminate ar a, where more X and less Y ar foun .
n t rm nat area, wh r mor
an
ss are found.
Polar antonymy rule of thumb: If it’s X then it can’t be Y and vice versa, but it can be
somewhere in between.
Directional antonymy
Directional antonyms include various subtypes: some denote contrary movement or position,
for instance, pairs of words representing opposing poles along a shared axis.
Explain ‘frame semantics’ theory.
13. This theory which was produced by Fillmore describes words, their various meanings, and how
these are combined with others to form the utterances and sentences of a language. Its aim is
to analyse and record for each sense of a word or phrase, the full range of its semantic and
syntactic relations.
What are the roles of the frames, frame elements and context?
A semantic frame is a schematic representation of a situation type together with a list of the
typical participants, props, and concepts that are to be found in such a situation; these are
the semantic roles, or frame elements (Fes). The context is normally the phrase or clause and
roles
(Fes)
clause,
maximally the sentence in which the target word appears in corpus data.
Which steps are involved during the analysis process carried out by using frame
semantics theory?
There are several distinct steps in the analysis process:
o First, the frame is defined, and its core elements named and described.
o Next, a list is made of as many words as can be found which in one of their senses evoke
that f
th t frame.
o Then, for each sense, or LU, a set of corpus sentences is extracted, in which the word is
used in the particular sense.
o Each sentence is annotated by marking off any section which instantiates an FE and by
y
g
y
y
recording for each FE thus identified:
o Its phrase type
o Its grammatical function
What is the importance of using the frame semantics approach?
The frame semantics approach, grounded in a coherent theory, offers the possibility of a
more systematic, less subjective way of analysing corpus data, and gives us confidence that
all relevant features are being captured.
Write the information types which are relevant to make a lexicographic record
of a word.
Three types of information are relevant to making a lexicographic record of a word:
14. o What we know, as native speakers, about the headword (its inherent properties)
o What we learn from its use in corpora and elsewhere (its contextual features)
o Wh we know about where the citations came from (the properties of the source
What
k
b
h
h i i
f
( h
i
f h
texts)
What are the inherent properties of the headword?
The properties of the headword that principally concern us can be summarized very
briefly:
o Its wordclass
o Its wordforms
ts wor forms
o Its grammatical behaviour
o Its semantics
Why learning the properties of the source texts is sometimes needed?
Using this information the computer can tell you whether a particular citation comes
mainly from spoken or written language, or political documents, or feminist publications
and so on. If you have any doubts about the way a word is used, it’s useful to be able to
check up on where the citation came from
from.