Online Appendix 1

This is an appendix to Bednarek, Monika (2008). Emotion Talk across Corpora. Palgrave.

Appendix to Chapter 1 of Emotion Talk Across Corpora (Appendix 1)
Contents:

A 1.1

Design of BRC

2

A 1.1.1 Conversation

3

A 1.1.2 News reportage

5

A 1.1.3 Fiction

6

A 1.1.4 Academic discourse

8

A 1.1.5 Summary

9

A 1.1.6 Key words in the BRC sub-corpora

11

A 1.1.7 List of files included in BRC

13

A 1.2

Analyzing emotion terms

17

A 1.3

List of emotion terms

24

Notes

35

A 1.1

Design of BRC

The British Register Corpus (BRC) is a register-sensitive corpus with four distinct subcorpora, which was compiled through a principled sampling from the BNC (British National
Corpus). The following sections describe its design in detail in order to follow one of the most
important principles in corpus linguistic research, namely to outline ‘[t]he design and
composition of a corpus […] with information about the contents and arguments in
justification of the decisions taken’ (Sinclair 2004a: 8).
As noted, the compilation of the BRC was exclusively based on the BNC. This is a general corpus of British English, consisting of over 100 million words from different varieties of
modern English (about ten million words of spoken and about 90 million words of written
English). Not all of the various parts of the BNC were compiled to be representative of the
register they represent, since the BNC was designed as a general corpus of English (representative of the British English language as a whole rather than representative of the different
registers that make up the corpus). As pointed out by Aston & Burnard (1998), ‘[a]lthough the
BNC distinguishes several different geographical, sociological, and generic varieties, it does
not necessarily provide a reliable sample for any particular set of such criteria’ (Aston &
Burnard 1998: 28). For more detailed information on the BNC see e.g. the BNC website at
http://www.natcorp.ox.ac.uk/what/index.html, Burnard (1995), Crowdy (1995), Rayson et al
(1997) or Leech et al (2001: 2-4). Since the BNC offers only a very broad categorization system for the texts included in the corpus (domain, context, socio-economic class), Lee’s (2001,
2002) much more detailed categorization system was used to select texts from the BNC to
compile the BRC. For details of this system and some of the difficulties and complexities involved see Lee (2001, 2002). The design of the BRC thus resembles the design of the fourmillion word BNC baby (Burnard 2003, http://www.natcorp.ox.ac.uk) in that 1) both are fourpart corpora, 2) both use the text classifications provided by David Lee. However, the BRC is
about five times bigger than the BNC baby, and there is no complete overlap of files. Other
differences include the sampling procedures adopted in selecting texts for each register. Readers particularly interested in the issue of corpus design may wis h to compare the design of the
BRC (as described below) with the design of the BNC baby (Burnard 2003) and the LSWE
corpus used by Biber et al (1999) (the BRC registers are at times less broad and more homogenous than the LSWE registers and contain no American English).

2

A 1.1.1 Conversation

Design

The BRC conversation sub-corpus is equivalent to the BNC spoken demographic section,
which is made up of 153 files (4,206,058 words). This sub-corpus of the BNC comprises casual conversational British English, collected by 124 ‘respondents’ (aged 15 and older), who
taped their conversations with other speakers during a certain time period. The respondents
were selected with the help of demographic sampling, and a relatively high degree of representativeness was achieved by including ‘as far as possible, […] equal numbers of men and
women, equal numbers from each of the six age groups, and equal numbers from each of four
social classes’ (Burnard 1995: 20).
To achieve a high degree of balance in terms of social group and gender is extremely important for the study of affect/emotion since it cannot be assumed a priori that both social
group and gender do not have an effect on the usage of emotion vocabulary: ‘[i]n stratified
societies, social groups are frequently perceived as having different affective styles, and class
identification rests in part on the individual’s affective demeanor’ (Besnier 1990: 435).1 Presumably, age difference is also potentially important in this respect (Schrauf & Sanchez 2004
note a difference in the emotion words that have psychological salience for young and old
speakers), although it may be difficult to pinpoint exactly which social factor is responsible
for a given variation, and such an analysis of social variables and affect/emotion was well
beyond the scope of Bednarek (2008).
In total, the BRC conversation sub-corpus includes female and male respondents from
four social groupings: AB (top or middle management, administrative or professional), C1
(junior management, supervisory or clerical), C2 (skilled manual), and DE (semi-skilled or
unskilled). It also comprises respondents from six different age groups. Table A.1 below
presents the breakdown of respondents by social class (according to Lee 2002), age
(according to Burnard 1995) and gender (according to Burnard 1995), in number of words as
well as in percentages (from Leech et al 2001: 3):

3

Table A.1 BRC conversation sub-corpus
Respondent social class
Number of files
AB
59
C1
36
C2
32
DE
19
Unknown
7
Respondent Age
Number of files
0-14
26
15-24
36
25-34
29
35-44
22
45-49
20
60+
20
Respondent Gender
Number of files
Male
73
Female
75
Unclassified
5
Total
153

Number of words
1,363,571
1,097,023
1,192,120
515,981
37,363
Number of words (W-units)
265,716
668,947
847,236
839,026
956,474
633,817
Number of words (W-units)
1,732,731
2,462,339
16,146
4,206,058

Percentage
32.54
26.02
25.64
14.88
0.89
Percentage
6.3
15.88
20.11
19.92
22.71
15.05
41.14
58.47
0.38
100

Issues

The BRC conversation sub-corpus is probably the most representative of all the BRC subcorpora, since it was originally designed for maximum representativeness. It has been used in
much linguistic research, and is argued to ‘[provide] an unparalleled resource for investigating, on a large scale, the conversational behaviour of the British population in the 1990s’
(Rayson et al 1997: 134). However, the sampling only relates to the respondents, rather than
their interlocutors, and concerning cross-tabulation it is difficult to say whether there is indeed
a balance for all combinations of factors (e.g. are there the same number of male speakers of
age group 1 and social class 1 as there are female speakers of the same age and social class,
and as there are male speakers of other age groups and social classes?). Furthermore, Table
A.1 does not tell us much about the number of words that is actually produced by the different
speakers. In fact, as pointed out by Rayson et al (1997) female speech is slightly overrepresented and male speech is underrepresented in the conversational sub-corpus of the BNC
(compare Table A.2):
Table A.2 Female vs. male speakers in conversation sub-corpus
Number of speakers
Number of turns
Number of words spoken
Number of turns per speaker
Number of words per turn
(from Rayson et al 1997: 136)

Female speakers
561
250,955
2,593,452
447.33
10.33

Male speakers
536
179,844
1,714,443
335.53
9.53

4

However, no emotion terms are among the most over or underrepresented lexical items listed
by Rayson et al (1997: 136-139), so this appears relatively unproblematic at least for the study
of affect/emotion.

A 1.1.2 News reportage

Design

The news section of the BNC (that is, all files from newspapers) consists of 486 files
(9,345,878 words). For the BRC news reportage sub-corpus, texts were sampled from across
the various topic and section categories listed by Lee (2002). These are: arts/cultural material,
commerce/finance, home/foreign news, science, lifestyle/leisure/belief & thought, and sports
(coded by Lee as arts, commerce, report, science, social, sports). Texts were also sampled
from both ‘quality’ broadsheet newspapers and ‘popular’ tabloid newspapers. The following
British national daily newspapers are included in the corpus: The Independent, The Guardian,
The Daily Mirror, and The Daily Telegraph, representing different readerships and political
attitudes. Local and regional newspapers (as well as a miscellaneous (unclear) category and
spoken news) were excluded, because researchers have found that ‘while the conditions for
the formulation of media language are similar, since practitioners are bound by the strictures
of their discourse community of media- makers, the results are realized differently in different
local contexts’ (Cotter 2001: 429). Since the focus was on informative rather than persuasive
discourse, newspaper editorials were also excluded, making up a distinct register with a different global purpose.
On account of the complex process of news writing (for a summary see Bednarek 2006)
multiple authors are involved, which means that the characteristics of the speaker/writer (age,
sex, regional origin) that are important for conversation are somewhat irrelevant for news reportage. Table A.3 below shows the breakdown of news texts according to
broadsheets/tabloids, topic/section (Lee 2002) and newspaper in terms of number of files and
words:

5

Table A.3 BRC news reportage sub-corpus

Type
Broadsheet

Topic/section
Arts/cultural material

No of files
51

Commerce/finance

44

Home/foreign news
49
Science
29
Lifestyle/leisure/belief & thought 36

Tabloid
Total

Sports
no further sub-categories

24
6
239

Newspapers
Independent, Guardian, Daily
Telegraph
Independent, Guardian, Daily
Telegraph
Independent, Guardian
Guardian, Daily Telegraph
Independent, Guardian, Daily
Telegraph
Independent, Guardian
Daily Mirror
Daily Mirror, Daily Telegraph,
Guardian, Independent

No of words
351,811
424,895
663,355
65,293
81,895
297,737
728,413
2,613,399

Issues

One of the disadvantages of the news reportage sub-corpus is that it includes only samples
from four of the ten national British broadsheet newspapers: three broadsheets and just one
tabloid newspaper (Daily Mirror). However, the newspapers at least represent different political stances, and the number of words sampled from the tabloid is relatively high. A more
problematic issue relates to the fact that the tabloids are not sub-categorized in terms of different topics/sections/subject matter, which makes a more detailed analysis of intra-register
variation difficult. In terms of tabloid newspaper language better large-scale corpora are
clearly needed.

A 1.1.3 Fiction

Design

In the BNC the genre W_fict_prose (novels and short stories) consists of 432 files
(15,926,677 words). For the BRC fiction sub-corpus it was decided to focus on a less broadly
defined and more homogenous register of relatively contemporary adult book fiction, comprising only fiction from 1985-1994 in book form. For this reason, files that indicated a collection or compilation of stories were excluded to filter out short stories. Texts were sampled
with the aim of achieving balance of author sex (i.e. a similar number of male/female authors)
6

and author age (i.e. a similar number of younger/older authors). Audience sex was not taken
into consideration, though the majority of files is addressed to a mixed audience. The breakdown for the included fiction texts by gender and age (according to Lee 2002) is given in
Table A.4:
Table A.4 BRC fiction sub-corpus
Gender
Number of files
male
86
female
84
Unknown
2
Age
Number of files
25-34
14
35-44
28
45-59
26
60+
30
Unknown
74
Total
172

Number of words
3,326,231
3,297,300
64,928
Number of words
532,585
1,091,130
1,011,365
1,137,025
2,916,354
6,688,459

Issues

Although the number of words for male/female and older/younger authors is relatively similar, there are still some imbalances when cross-tabulation is involved (the number of
male/female speakers of a certain age). As Table A.5 shows, there are, for instance, more
texts of male authors between 25-44 than female authors of that age, and more texts of female
authors that are between 45-60+ than of male authors.
Table A.5 Cross-tabulation of age and sex
Age
Number of files
Female author
Male author
25-34
5
9
35-44
7
21
45-59
16
10
60+
17
13
unknown
39
33

Number of words
Female author Male author
209,311
323,274
251,777
839,353
635,887
375,478
644,108
492,917
1,556,217
1,295,209

However, achieving complete balance would have meant a considerable reduction of the size
of the corpus. Another disadvantage of the fiction sub-corpus is that no systematic sampling
could be undertaken in terms of the different types or kinds of fiction (such as ‘mystery’, ‘romance’, ‘historical’, ‘adventure’, ‘science’, ‘general’) – which may well represent different
registers – because this information is not explicitly given by Lee (2002) and would have had
to be guessed from the contents or titles.

7

A 1.1.4 Academic discourse

Design

For the academic discourse sub-corpus of the BRC, samples of academic writing were taken
from a wide range of (groups of) disciplines contained in the BNC (according to Lee 2002):
the humanities, medicine, natural sciences, politics/law/education, social/behavioural
sciences, and technology/computing/engineering (an approach similar to that adopted by
Oakey 2002).2 In order to achieve balance, about one million words from each of these ‘discipline’ groups were sampled, with the aim of including material that was as diverse as possible
(taking into account the criteria of medium, keywords, author sex and type of authorship), e.g.
research articles and book extracts (however, this was sometimes difficult since the material
contained in the BNC is limited in variety). As a result, the academic discourse sub-corpus of
the BRC represents a rather ‘broad’ register of academic writing, rather than a more specialized corpus of e.g. journal articles, book chapters or doctoral theses (for such studies see e.g.
Hyland 1999, Thompson & Tribble 2001). The breakdown of texts according to subject area
is listed in Table A.6.
Table A.6 BRC academic discourse sub-corpus
Subject area
Humanities
Medicine
Natural sciences
Politics, la w, education
Social and behavioural sciences
Technology, computing, engineering
Total written academic prose

Number of files
26
14
38
41
38
23
180

Number of words
1,001,982
1,001,196
999,855
1,130,111
1,141,785
686,004
5,960,933

As becomes apparent, there is a relative balance in terms of the number of words sampled
from different discipline groups, with the exception of technology/computing/engineering,
which derives from the fact that the BNC contains only 686,004 words of this variety.

Issues

It is slightly problematic that different academic subjects such as politics, law and education
or technology, computing and engineering are classified as one category rather than as several
distinct categories. This means that there are fewer words concerning politics than e.g. medicine. In this, the academic discourse sub-corpus is not completely balanced. However, as with
8

fiction, the alternative (a more balanced corpus in these terms) would have resulted in a considerable reduction of corpus size (a sampling of far fewer than one million words from each
academic subject).

A 1.1.5 Summary

To sum up, the British Register Corpus (BRC) is made up of four sub-corpora: conversation,
news reportage, fiction, and academic discourse with different sub-categories for academic
discourse, conversation and news reportage according to topic, discipline and social class:3

BRC (British Register Corpus):

Acad

hu

me

ns

sos

as

ple

Fiction
BRC
tb
News

bs

art

fin

nw

Convers

AB C1

C2 DE uk

sc

lst

sp

The BRC contains samples rather than whole texts, which is not completely in line with a
functional approach (Butler 2004: 152). However, many of the samples are rather substantial
(full chapters, paragraphs, whole articles, conversations); thus, the standard size in the BNC
of a sample from a book is 40,000 words (Kilgarriff 1997a: 138). It is also worth pointing out
that the corpus only includes English up to the early 1990s, and perhaps no longer represents
contemporary English as spoken at the beginning of the 21st century. Finally, it is important to
emphasize that the BNC was not originally designed for studies of individual registers or register variation (see above, Aston 2001, and Oakey 2002: 115). Designing the BRC on the basis of the BNC hence violates one of the principles established by Sinclair: ‘Only those components of corpora which have been designed to be independently contrastive should be contrasted’ (Sinclair 2004a: 3). Sinclair continues,

9

the existence of components differentiated according to the criteria discussed below,
or identified by archival information, does not confer representative status on them,
and so it is unsafe to use them in contrast with other components. Now that with
many corpus management systems it is possible to ‘dial-a-corpus’ to your own requirements, it is important to note that the burden of demonstrating representativeness lies with the user of such selections and not with the original corpus builder. It
is perfectly possible, and indeed very likely, that a corpus component can be adequate for representing its variety within a large normative corpus, but inadequate to
represent its variety when freestanding. This point cannot be overstated; a lot of research claims authenticity by using selections from corpora of recognised standing
(Sinclair 2004a: 3).
However, I tried to be as careful and balanced in the design of the BRC as possible, and to
provide maximum transparency regarding its design, so that the degree of representativeness
that the BRC achieves can be judged by fellow researchers.
Furthermore, the only alternative is to design one’s own corpus, which results in a number of problems because of time and other (e.g. financial) constraints where large-scale corpus
research is concerned. And what the design of the BRC loses in representativeness, it gains in
replicability. In any case, a corpus that fulfils all the design criteria mentioned by Sinclair
remains yet to be compiled, and ‘[w]hile these [representativeness and balance] are not precisely definable and attainable goals, they must be used to guide the design of a corpus and
the selection of its components’ (Sinclair 2004a: 9). This was clearly the case in the design of
the BRC: Each of its four sub-corpora was designed to be as representative as possible of the
register it represents, and as balanced as possible, while at the same time being as large as
possible. The important phrase here is as possible, which means that some compromise had to
be made, as has become apparent in the discussion above. Much more could be said on representativeness and corpus design, which is a thorny issue in corpus linguistics (for discussions
compare e.g. Hunston 2002, Mahlberg 2004, Wynne 2005), but this is not the place to do so.
Instead, it must be noted that the results from the BRC are to be considered as indicative, but
not definitive (compare Sinclair 2004b: 81), and that no claims of guaranteed representativeness are made. Clearly, the corpus is not ‘representative’ of the chosen registers as such,
rather it is as representative as possible given the material contained in the BNC. As Sinclair
concludes:

It is important to avoid perfectionism in corpus building. It is an inexact science, and
no-one knows what an ideal corpus would be like. […] [C]ompilers make the best
corpus they can in the circumstances, and their proper stance is to be detailed and
honest about the contents. From their description of the corpus, the research community can judge how far to trust their results, and future users of the same corpus
can estimate its reliability for their purposes. We should avoid claims of scientific
10

coverage of a population, of arithmetically reliable sampling, of methods that guarantee a representative corpus. The art or science of corpus building is just not at that
stage yet (Sinclair 2004b: 81).

1.6 Key words in the BRC sub-corpora
100 Wordsmith key words in the BRC sub-corpora (reference corpus BRC, log- likelihood
test, max p value 0.1, minimum frequency 3):

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41

Academic
#
OF
THE
IS
IN
WHICH
BY
ARE
MAY
PATIENTS
OR
BE
THESE
SUCH
FORMULA
BETWEEN
NDASH
ALSO
HOWEVER
EXAMPLE
DATA
SOCIAL
THIS
AS
STUDY
LSQB
RSQB
SYSTEM
THUS
HAS
THEREFORE
FORM
AN
FIG
ANALYSIS
LANGUAGE
NON
CELLS
DISEASE
SHOWN
WITHIN

News
#
POUND
MR
YESTERDAY
YEAR
HAS
BY
MDASH
LAST
WILL
BRITISH
NEW
GOVERNMENT
CENT
MILLION
BRITAIN
WORLD
PER
FOR
MARKET
SHARES
THE
COMPANY
LONDON
ENGLAND
EUROPEAN
SEASON
AFTER
CITY
CHAIRMAN
TEAM
FORMER
UNION
DAVID
MINISTER
PRESIDENT
LEAGUE
WHO
DOLLA R
ITS
PLAYERS

Fiction
EQUO
BQUO
HER
SHE
HE
HIS
HAD
WAS
HIM
HELLIP
LOOKED
SAID
EYES
MY
ME
FACE
BACK
KNEW
MAN
HEAD
ROOM
HERSELF
INTO
DOOR
TURNED
COULD
FELT
VOICE
SEEMED
ASKED
WOULD
MDASH
FATHER
HAND
STOOD
AT
AWAY
SMILED
HIMSELF
HANDS
CAME

Conversation
I
YOU
YEAH
T
S
IT
OH
WELL
GOT
ER
MM
KNOW
WHAT
DON
NO
ERM
VE
THAT
DO
THEY
YES
RE
GET
LL
COS
WE
GO
JUST
THINK
THERE
GONNA
RIGHT
LIKE
MEAN
SO
CAN
M
ALRIGHT
YOUR
REALLY
ONE

11

42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97

C
FIGURE
INFORMATION
B
STUDIES
P
ET
RESULTS
SECTION
G
EFFECT
DIFFERENT
USING
NUMBER
IMPORTANT
TREATMENT
DEVELOPMENT
PARTICULAR
ACT
CASES
AREAS
LEVEL
STATE
GROUPS
AL
FUNCTION
BOTH
EVIDENCE
EACH
PROCESS
SPECIFIC
SYSTEMS
EDUCATION
DNA
STRUCTURE
ORDER
CONTROL
USE
ACTIVITY
CASE
POSSIBLE
ACID
VALUES
DESCRIBED
ASSOCIATED
E
SPECIES
SIGNIFICANT
H
GASTRIC
INDIVIDUAL
POPULATION
WOMEN
SIMILAR
CHAPTER
PER

NATIONAL
WEST
WIN
JOHN
SPOKESMAN
PROFITS
INTERNATIONAL
PARTY
COMPANIES
HONG
MONTH
KONG
LABOUR
DIRECTOR
SECRETARY
EAST
CLUB
GAME
AGAINST
MATCH
EUROPE
TAX
BID
UNITED
STAR
SPORT
TV
BANK
YEARS
CONFERENCE
PAGE
WEEK
PRICE
GERMANY
MANAGER
CORRESPONDENT
EDITION
SALES
CUP
THEIR
CORPUS
CASH
LEADER
INDUSTRY
UNIVERSITY
GERMAN
SMITH
FIRST
POLICE
SHARE
AMERICAN
SOVIET
INVESTORS
GROUP
FINAL
BILLION

SAT
D
DOWN
MOTHER
DARK
OUT
THOUGHT
AGAIN
TOOK
NODDED
QUOT
SAW
MOMENT
NEVER
NOTHING
WOMAN
SUDDENLY
SMILE
SHOOK
TOO
HAIR
BEGAN
BEHIND
ACROSS
STARED
UP
WALKED
HEARD
TOLD
FEET
LOOKING
SIR
MOUTH
LAUGHED
BED
AM
CORBETT
HOUSE
AROUND
BESIDE
LIPS
ARMS
SOMEONE
REPLIED
ROSE
WINDOW
TELL
OPENED
PULLED
SILENCE
TOWARDS
WONDERED
FLOOR
WANTED
GIRL
GLANCED

WANT
AH
GOING
THEN
PUT
BIT
ME
SEE
DIDN
OOH
HAVEN
COME
NICE
THEM
ISN
MUM
LOOK
SAY
HAVE
DID
OKAY
GOTTA
UP
DOING
SORT
THING
BLOODY
HERE
GOOD
TWENTY
GOES
D
HA
IF
FIVE
ALL
LOT
HUNDRED
ANYWAY
WHY
DOESN
NOW
AYE
ACTUALLY
WON
WANNA
INNIT
FUCKING
DONE
MY
HE
SAYS
SAID
DAD
MHM
SOMETHING

12

98
99
100

GENERAL
X
REQUIRED

TORY
CHAMPION
FINANCIAL

HADN
MRS
ATHELSTAN

AIN
WOULDN
EH

1.7 List of files included in BRC

Conversation sub-corpus
Spoken_AB
KB0
KB3
KB8
KBC
KBD
KBK
KBM
KBR
KBS
KBU
KBW
KC0
KC3
KC4
KC8
KC9
KCB
KCD
KCH
KCR
KCV
KCW
KD4
KDL
KDR
KDU
KE1
KNR
KP2
KP3
KP4
KP6
KP8
KP9
KPA
KPB
KPC
KPE
KPF
KPG
KPH
KPK
KPL
KPN
KPP
KPT

Spoken_C1
KB9
KBG
KBH
KBJ
KBL
KBV
KBY
KC6
KCC
KCK
KCN
KCP
KCS
KD0
KD2
KD5
KD7
KDB
KDG
KDJ
KDK
KDM
KDP
KDV
KDW
KDY
KE0
KE3
KE4
KP1
KP5
KPJ
KPM
KPR
KPU
KR0

Spoken_C2
KB5
KB7
KBA
KBB
KBF
KBN
KBT
KBX
KC1
KCE
KCF
KCG
KCL
KCM
KCT
KCX
KCY
KD1
KD3
KD8
KD9
KDA
KDC
KDD
KDE
KDH
KDT
KDX
KE2
KE6
KP7
KPD

Spoken_DE
KB1
KB2
KB4
KB6
KBE
KBP
KC2
KC5
KC7
KCA
KCJ
KCU
KD6
KDF
KDN
KDS
KE5
KPS
KSS

Spoken_Unclass
KNS
KNT
KNU
KNV
KNW
KNY
KP0

13

KPV
KPW
KPX
KPY
KR1
KR2
KSN
KSP
KSR
KST
KSU
KSV
KSW

News reportage sub-corpus:
News_
Arts
A1D
A1K
A1R
A20
A24
A25
A2B
A2D
A2G
A2R
A2U
A32
A35
A36
A3H
A3R
A42
A4A
A4E
A4L
A4S
A51
A54
A5B
A5E
A5F
A7S
A83
A8F
A8S
A93
A9C
A9K
A9T
AA2
AA9
AAH

News_
News_
Commerce Report
A1E
A1S
A21
A26
A2H
A2V
A37
A3J
A3K
A3S
A43
A4F
A55
A5G
A5S
A5T
A7T
A7U
A85
A86
A8G
A8H
A8U
A8V
A94
A9D
A9L
A9U
AA3
AAA
AAJ
AAS
AHB
AHJ
AHT
AJ2
AJ9

A1G
A1J
A1V
A1Y
A28
A2A
A2M
A2P
A2X
A30
A3D
A3G
A3U
A3W
A46
A49
A4H
A4K
A4N
A4X
A50
A57
A59
A5M
A5R
A7V
A7W
A87
A88
A8J
A8K
A8W
A8X
A95
A96
A9E
A9F

News_
Science

News_
Social

News_
Sports

A1M
A3Y
A82
A8A
A8E
A8R
A92
A9B
AAG
AH9
AHD
AHL
AHP
AHV
AHY
AJ4
AJ7
AJB
AJE
AJK
AJS
AK0
AK3
AK7
AKA
AKF
AKN
AKW
AL4

A1L
A1X
A2C
A3M
A3X
A48
A4M
A5V
A7Y
A8B
A8M
A8Y
A98
A9Y
AHE
AHH
AHM
AHS
AHW
AJ1
AJ5
AJC
AJG
AJL
AJT
AJW
AK1
AK8
AKC
AKG
AKK
AKP
AKT
AKX
AL1
AL5

A1N
A22
A2E
A2S
A33
A3L
A40
A4B
A4P
A52
A5C
A5U
A80
A8C
A8N
A90
A99
A9H
A9R
AA0
AA7
AAE
AAN
AAW

News_
Tabloids
CH1
CH2
CH3
CH5
CH6
CH7

14

AAR
AHA
AHG
AHR
AJ0
AJ8
AJF
AJN
AJV
AK4
AKB
AKJ
AKS
AL0

AJH
AJP
AJX
AKD
AKL
AKU
AL2

A9M
A9N
A9V
A9W
AA4
AA5
AAB
AAC
AAK
AAL
AAT
AAU

Fiction sub-corpus:
A0L
A6J
A6N
A74
AB9
ABW
AC2
AC3
ACE
ACK
AD9
ADS
ADY
AE0
AEA
ALJ
AN7
ANL
ANY
APM
APR
APU
AR2
AR3
ARK
AS7
ASD
ASE
ASN
ASS
AT7
ATE
B1X
B20
B3J
BMN
BMR

BPA
C86
C8D
C8E
C8S
C8T
C98
CA0
CAM
CB5
CCD
CCM
CD2
CDB
CDE
CEB
CEC
CEH
CFY
CHG
CJA
CJF
CK9
CKB
CKC
CKD
CKE
CKF
CLD
CMJ
CN3
CR6
CRE
EA5
ECK
EDN
EEW

F9C
F9X
FAB
FAJ
FB0
FB9
FNT
FNU
FP0
FP3
FP6
FP7
FPF
FPK
FPM
FPX
FR3
FRF
FRS
FS8
FSC
FSE
FSF
FYV
FYY
G04
G0A
G0N
G0S
G0X
G10
G15
G16
G17
G1L
G1S
GUE

GVP
GW2
GWF
GWG
H7F
H7H
H7P
H7W
H84
H85
H8A
H8B
H8X
H90
H98
H9C
H9G
H9N
HA2
HGL
HGN
HGU
HH0
HH5
HHC
HJH
HNJ
HR8
HR9
HTH
HTM
HTR
HTT
HTU
HTW
HU0
HW8

15

BMW
BNC
BP0
BP1
BP7
BP9

EF1
EV1
EVC
EVG
EWH

GUG
GUK
GV2
GV6
GV8
GVL

HWE
HWL
HWN
HWP
JY3
K8R
K95

Academic discourse sub-corpus:
Academic_
humanities
A6G
A6U
ARD
F98
GWM
J0V
J1A
A04
A05
A07
CG0
EA7
EE2
EEE
G1N
G1R
GUW
HY6
A6B
B2C
CFK
CK1
CM2
CM8
H7Y
H8V

Academic_
medicine
B0X
B30
B33
EWX
CAN
HU2
HWU
EA0
HU3
HWV
EA1
HU4
HWW
HWS

Academic_
natural
science
EV6
ARY
FTB
FTC
FTD
FTE
K5N
K5P
K5R
K5S
K5T
K5U
K5V
K5W
K5X
B2J
B2K
CMA
CMH
E9X
EV9
EVW
FEF
AML
AMM
FU0
FU9
G1E
GU5
GU8
GV0
GW6
H79
H8K
H9R
H9S
HRG
J12

Academic_
Politicslaw
education
B12
CJG
EF3
GW1
HPX
HXT
HXW
HYB
J6R
J76
FBS
FBT
FBU
FBV
FBW
FBX
FBY
FC0
FC1
FC2
FC3
FC4
FC5
FC6
FC7
A5Y
A62
A64
A6F
A6M
ABP
ACJ
ADD
AM6
AN5
APE
APN
ASB
HTF
HXD
HXE

Academic_ Academic_
socialscience techengin
AN3
B16
CGF
CGT
CMN
CMR
CRF
ECB
ECE
FRL
J14
J7H
J7L
AS6
EB1
EWA
HP2
J7F
J7G
J7J
J7K
J7P
K93
ALM
ALN
ALP
B2X
FBH
GWJ
B1G
CJ1
CLH
FST
G08
G1J
GVA
H0J
H8D

B2M
BP2
CA4
CG7
CG8
CG9
CGA
CHF
EES
EUS
EWW
FE6
FNR
FPG
G3N
H0U
H7R
HGR
HGX
HR3
HRK
HX9
K90

16

Further information on these files is provided in David Lee’s excel spread sheet, available to
download at http://clix.to/davidlee00.

A 1.2

Analyzing emotion terms

One of the central problems in the analysis of affect/emotion is how to establish a list of emotion terms. Affect is used in Bednarek (2008) as a cover term for feelings, emotions, moods,
affect, etc and thus includes more than just emotion, but how much more? Where is the line
between an emotional state of mind and a state of mind that does not involve emotion? Like
other semantic-conceptual categories, affect is a fuzzy category, with no clear boundaries,
presumably organized with the help of prototypes as cognitive reference points and family resemblances (e.g. Shaver et al 1987, Russell 1991, Oatley et al 2006: 184). For example, does
confusion denote an emotion or not? How can we delineate the semantic domain of affect?4
What Ortony et al (1987) call the affective lexicon comprises words that implicate emotions in
various ways, as has also been noted by other emotion researchers. Words can denote

expressions of emotions (e.g. laughter, smiling, crying, tears, frown), bodily states
associated with emotions (e.g. strong, tiredness), properties of emotion (e.g. deep,
positive, negative, expressive, mixed, disturbed, uncontrollable, turbulent), characteristics of behaviour motivated by emotion (e.g. sincerity, giving, helping, sharing,
violence), personality traits related to emotion (e.g. outgoingness, gentleness, sensitive, stubbornness, hardness, vulnerability, hyperactive), states of mind associated
with emotions (e.g. confusion, uncertainty, arousal, control, conflict, thinking,
meditating, alert), and cognates and superordinates of emotion (e.g. reactions, responsive, state, communication, expression). (Johnson-Laird & Oatley 1989: 87-88)
As Fiehler notes: ‘Es gibt sicherlich einen Kernbereich prototypischer Emotionen, bei denen
auch intersubjektiv große Übereinstimmung herrschen wird, daß es sich um Emotionen handelt […] . Es gibt aber einen breiten Übergangsbereich, in dem die Urteile differieren.’
(Fiehler 1990: 56).5 For instance, there is no clearly defined border between affective and
other mental processes (Fiehler 1990: 56), and many existing taxonomies involve differences
of degree (see, for example, Ortony et al 1987: 353). Consequently, while the list of 1060
emotion terms used in Bednarek (2008) contains many lexical items that would unanimously
be judged as prototypical emotion terms by many researchers (e.g. hate, love), it also includes
terms that may be excluded by some researchers, and excludes terms that may be included by

17

other researchers. As noted in Chapter 1 (Note 6) different lists of emotion vocabulary have
very different lengths.
In Bednarek (2008), the Encarta Thesaurus (ET) was used as the basis for establishing a
list of emotion terms, but it was not adopted on a one-to-one basis. While some words had to
be excluded for purely methodological reasons (Bednarek 2008: Section 1.6.3), other
decisions related to the ET itself. For instance, its categorization is sometimes not consistent
(e.g. laugh and weep are not inc luded but close to tears, crying, sigh and crack up are), and it
is not clear on which methodological principles its categorization is based. In some cases it
looks as if one word if included as adjective was automatically included as noun, verb etc
even if this made no sense. Other words that are problematic include canonise, oracle or
martyrdom where the connection to affect seems controversial. A possible solution to this
problem would be to use the ET list as basis for a prototype rating of emotion terms, working
with a representative sample of native speakers. However, to establish the status of almost
2000 words was not feasible (and existing lists of prototype ratings for emotion terms were
considered too small). Instead, it was decided to adopt a relatively broad approach to affect, in
effect following the compilers of the ET, but excluding terms that seemed to refer to states of
mind or character rather than emotions, and focusing on emotion talk rather than emotional
talk (Bednarek 2008: Section 1.3). For instance, the following categories of the ET were
excluded, as expressing rather than denoting emotions/states of mind, as referring to states of
mind rather than to emotions, or as referring to psychological and related phenomena:6

(1)

Expressions of surprise (e.g. gee, oh my etc), regret (e.g. alas), uncertainty (e.g.
apparently, seemingly)

(2)

Calmness, confidence, and composure (e.g. accustomed, calm), Pensiveness and
interest (attentive, pensive), Prejudice (e.g. bias, chauvinism), Neutrality and indifference (e.g. absently, apathy), Bore and fail to interest (e.g. bore, weary), Soothe
and calm (e.g. appease, sweeten), Appeal to and arouse interest (e.g. captivate, inspire), Encourage (e.g. fortify, hearten), Change of mood and composure (e.g. adapt,
take heart), Experience and encounter (e.g. suffer, undergo), Ignorance (e.g. untrained, ignorant), Knowledge and wisdom (e.g. academic, bookish), Uncertainty
(e.g. agnostic, doubt), Certainty (e.g. certainly, belief)

(3)

Eccentricity and irrationality (e.g. barmy, raving), Fears and phobias (e.g. acrophobia), Fads, fetishes, and idolatry (e.g. craze, cult), Devotees and addicted people
(e.g. addict, patriot)
18

(4)

The will and willingness (e.g. cooperative, submissive), Unwillingness and
stubbornness (e.g. averse, difficult), Rebelliousness and disobedience (e.g. radical,
turbulent), Uncooperative or rebellious person (e.g. activist, warmonger)

In Bednarek (2008), establishing a list of emotion terms was a methodological issue, rather
than an issue that I was interested in from a more theoretical view point. The final list of more
than 1000 emotion terms certainly covers a large variety of different emotion terms in British
English.
After having established a list of emotion terms to investigate in the corpus, decisions
had to be made as to which of their usages to include as affect. As mentioned, the frequency
analyses reported in Chapter 2 are the result of a semi-automated count of the meaning of
emotion terms, involving a manual deletion of some concordance lines. Only the ‘affective’
meanings and usages of emotion terms were counted, with ‘non-affective’ meanings
excluded. This means that the analysis is much more ‘subjective’ than a purely automatic frequency analysis that is not sense-sensitive (à la Leech et al 2001). Ideally, such a meaningsensitive analysis would be as detailed as the analyses undertaken by West (1953), based on a
corpus of 2½-5 million words. Compare Table A.7 on page 20:

19

Table A.7: Analysis of meanings of emotion terms in West (1953)
love (2059 n, v):
affection:
Love of/for a friend (19%)
The Goddess of Love (25%)
Phrases:
in love with, fall in love, my love (4%)
Love-affair, love-letter, etc (2%)
hope (1283 n, v):
Have some/no/a hope of success, lose hope; raise the hopes of, build/fix one’s hopes on (48%)
The hope of his party, our last hope (3%)
fear (846 n, v):
emotion:
A feeling of fear; the fear of death (27%)
You need have no such fear (13%)
anxiety:
Did not go for fear that he might be hurt; was in fear of his life (8%)
(fear of god 2%; no fear!, no fear of that 0%)
feeling (feel, v 1615)
sense of touch:
I’ve no feeling in my fingers (1%)
pleasure, pain and emotion:
feeling of pleasure, pain, hope
feeling of ease, equality etc
showed such nice feeling in sending these flowers
good/ill feeling (14%)
sensibilities:
You’ve hurt my feelings (6%)
attitude, conviction:
I cannot go against the feeling of the nation
No particular feeling in the matter
A feeling that it’s going to be a success (11%)
shock (174 n, v)
The shock of battle, a blow, or explosion:
earthquake shock
electric shock (20%)
Figurative:
a shock to my feelings
her death was a great shock to me (23%)
Special medical use:
cases of shock after air-raids (8%)
shock -absorber etc (1%)
surprise (470 v, a, n)
I felt some surprise at seeing him there (24%)
Phrases:
What a surprise (3%)
Take a person by surprise (3%)
Adjectives happy (721)
contented, joyful:
be happy at school, feel happy (82%)
fortunate:
Happy New Year, many happy returns of the day
apt
a happy answer (9%)
sad (226)
feel sad, a sad look , a sad event (65%)
It is sad for/that (26%)
anxious (203)
uneasy:
feeling very anxious about the future; an anxious moment (48%)
Nouns

20

Adverbs

Verbs

eager:
I am most anxious to please you (52%)
angry (155)
irate:
I’m very angry; angry at/with (67%)
of mood, easily angered:
He’s in an angry mood (8%)
figurative:
angry sky, sea, wound (25%)
happily
in happiness:
lived happily every after (35%)
luckily (50%)
aptly (15%)
love (2059, n, v)
have friendship (16%)
between the sexes:
he loves her; ‘Tis better to have loved and lost (13%)
appreciate:
love reading, good wine etc (8%)
with a verb:
I love to hear you sing, you singing (2%)
worry (94 v, a, n)
cause minor irritation to:
Does the noise of my typewriter worry you? (19%)
cause anxiety:
I’m very worried about my son’s health (28%)
give way to anxiety:
Don’t worry (23%)
enjoy (434)
enjoy a book/reading/oneself etc
care (1134 n, v)
feel anxious for:
He cares only for his own interests; I don’t care (11%)
idea of responsibility:
The child has been well cared for (8%)
wish:
Would you care to read this?
Do you care to come out for a walk?
I don’t much care for dancing (14%)
hate (242 n, v)
I hate him
I hate to trouble you
I hate to hear good music badly played (78%)
admire (140)
observe with pleasure and surprise (96%)

However, because over 1000 emotion terms were analyzed, such a level of detail was not possible. Instead, decisions were made with respect to meanings/usages of each emotion term,
simply as to whether to include or exclude them from the frequency count (see Appendix A
2.3 online for examples). Such decisions necessarily entail some element of subjectivity,
especially where the difficult nature of evaluative/emotion meaning is concerned (compare
also Wallace & Carson 1973: 6, Moore et al 1999: 541, Whitelaw et al 2005).
The software tool for analyzing emotion terms was the Zurich BNCweb interface
(http://escorp.unizh.ch/), which allows different types of searches of the BNC and the BRC.
21

The type of search that was most frequently used was the ‘lemma query’, permitting a part-ofspeech sensitive search for emotion terms (e.g. love as verb or noun). For the analysis of lexical distribution the correctness of the lemmatization was not checked. The non POS-sensitive
‘standard query’ was also used in some cases. For instance, in order to consider spelling
variations/mistakes, the search was for the forms bad-tempered (lemma query) und bad tempered (standard query); for the lemma overwrought search terms were overwrought, overwrought (both lemma query) and over wrought (standard query) and so on. With standard
queries only the correct POS occurrences were included in the frequency count. Standard queries were also always necessary whenever more than one orthographic word was involved or a
specific form of a word was looked for (e.g. champing at the bit, guilty conscience, second
thoughts, willies).
As said, occurrences that did not refer to emotion were excluded. For example, depress
has four different meanings (from OALD – Oxford Advanced Learner’s Dictionary):
1
2
3
4

‘to make sb sad and without enthusiasm or hope’: Wet weather always depresses me.
‘to make trade, business, etc. less active’: The recession has depressed the
housing market.
‘to make the value of prices or wages lower’: to depress wages / prices
‘to press or push sth down, especially part of a machine’: to depress the clutch
pedal

Needless to say, only occurrences of meaning 1 were included in the frequency count. In
cases where it was easy to differentiate between meanings of affect, judgement and appreciation (Bednarek 2008: Section 1.4), appreciation and judgement were excluded (e.g. it’s a pity,
what a pity vs. out of pity); in other cases, this was not easily possible, and the figures may
include appreciation/judgement meanings of emotion terms.
Further, on account of the large number of words investigated, such decisions were
mostly based only on the immediate context – usually the sentence in which the potential
emotion term occurred, rather than the source text itself. If in doubt, the instance was included, rather than excluded. In some cases, differences between meanings were very small,
with meanings shading into each other (often metaphorical and literal meanings). The attempt
was made to be as consistent as possible in the analysis, but with so many occurrences, there
is a margin of error because of human fallibility and an incapacity for complete consistency
across the analysis of 1000 words: even if one emotion term only occurred ten times on
average in the entire corpus this would mean looking at 10,600 occurrences. The OALD
(which gives definitions for emotion term meanings and usages) was used as a help in the
22

analysis but the final decision was mine. A problem was also that run-on entries,7 in particular
emotion adverbs (compare Appendix A 2.3 online), are not described very thoroughly in the
OALD.
While it is not possible to detail decisions made with respect to every single emotion term,
some decisions are more general and will briefly be listed:

1. Identical or very similar occurrences were counted only once, though not all repeated
instances may have been detected.
2. Emotion terms in titles and names were excluded (e.g. Lady Happiness, Felicity, Turkish/Angel Delight etc), but emotio n terms in quotes (e.g. He said ‘I’m delighted’) were
included.
3. Different spellings were included (e.g. self-pitying/self pitying; ise/ize).
4. When there were more than 500 occurrences of an emotion term in one of the subcorpora, the figures for this sub-corpus were extrapolated from a random subset of 1020%. 20% (every 5th concordance line examined) were used as a subset if 20% of the
total number of occurrences was itself not more than 500 occurrences; otherwise 10%
were used (every 10th concordance line). If even 10% made up more than 500 occurrences, no manual analysis was undertaken across all sub-corpora, and the resulting
figures are not sense frequencies but word frequencies. This happened only with three
emotion terms: like (V),8 want (V),9 and sorry (A),10 which were in fact excluded from
the analysis of lexical distribution (compare Bednarek 2008: Section 2.2.1).
5. Dispersion (Leech et al 2001: 18) was not analyzed systematically. This is slightly
problematic because ‘[p]articularly for the lower frequency words, a large proportion
of the occurrences might take place in one conversation, and this would skew the results.’ (Rayson et al 1997: 149). This was partly levelled out by the large amount of
words looked at.
6. Occurrences for each emotion term also include negated occurrences. For example,
figures for angry include angry and not angry.

Summing up, readers should be aware that the reported figures are sense frequencies, not
word frequencies, and that there are methodological implications of this fact. All figures
should be interpreted as tendencies rather than precise and stable figures: ‘Preliminary investigations into the stability of outcomes in lexical semantics suggests that it is severely lacking.’ (Kilgarriff 1997b: 98). On the other hand, the analysis does have some advantage over a
23

pure automatic calculation: ‘A human parser, of course, does not attain 100 per cent, because
some instances are inherently indeterminate and humans also make mistakes; but there is still
a critical gap between what the machine can achieve and what the human can achieve.’
(Halliday 2005: 67).

A 1.3

List of emotion terms (alphabetic)

The listed words only have an ‘affective’ meaning in some of their usages, and only these
were counted. Spelling variations were also considered, i.e. for the lemma bad-tempered the
search was for the forms bad-tempered and bad tempered; for the lemma overwrought search
terms were overwrought, over-wrought and over wrought. For more details see A 1.2 above.

abash
abashed
abhor
abhorrence
abjection
abjectly
abominate
ache to
aching (for)
acrimony
addled
addlepated
admiration
admire
admiring
adoration
adore
adoring
adulation
affection
affinity
affronted
aflutter
afraid
afterglow
aggravate
aggravated
aggravation
aggrieve
aggrieved
aghast
agitate
agitated
agitation
agog

Verb
Adjective
Verb
Noun
Noun
adverb
Verb
Verb
Noun
Noun
Adjective
Adjective
Noun
Verb
Adjective
Noun
Verb
Adjective
Noun
Noun
Noun
Adjective
adverb
Adjective
Noun
Verb
Adjective
Noun
Verb
Adjective
Adjective
Verb
Adjective
Noun
Adjective

agonised
agony
alarm
alarm
alarmed
amaze
amazed
amazement
ambition
amorous
amuse
amused
anger
anger
angry
angst
angst-ridden
anguish
anguished
animosity
animus
annoy
annoyance
annoyed
antagonise
antagonism
antagonistic
anticipation
antipathetic
antipathy
antsy
anxiety
anxious
anxiousness
apologetic

Adjective
Noun
Noun
Verb
Adjective
Verb
Adjective
Noun
Noun
Adjective
Verb
Adjective
Noun
Verb
Adjective
Noun
Adjective
Noun
Adjective
Noun
Noun
Verb
Noun
Adjective
Verb
Noun
Adjective
Noun
Adjective
Noun
Adjective
Noun
Adjective
Noun
Adjective

24

apoplectic
appal
appalled
appetite
appreciate
appreciation
appreciative
apprehension
apprehensive
apprehensiveness
approval
approving
ardour
ashamed
aspiration
aspire
astonish
astonished
astonishment
astound
astounded
attachment
aversion
avid
avidity
awe
awe-stricken
awe-struck
awkwardness
bad
bad blood
bad mood
bad temper
bad-tempered
baffle
baffled
bafflement
bask
be burning to
be spoiling for
bedazzle
bedazzled
befuddlement
begrudge
bemusement
bent on
besotted
bewilder
bewildered
bewilderment
bewitched
bile
bitter
bitterness

Adjective
Verb
Adjective
Noun
Verb
Noun
Adjective
Noun
Adjective
Noun
Noun
Adjective
Noun
Adjective
Noun
Verb
Verb
Adjective
Noun
Verb
Adjective
Noun
Noun
Adjective
Noun
Noun
Adjective
Adjective
Noun
Adjective
Noun
Noun
Noun
Adjective
Verb
Adjective
Noun
Verb
Verb
Verb
Verb
Adjective
Noun
Verb
Noun
Adjective
Adjective
Verb
Adjective
Noun
Adjective
Noun
Adjective
Noun

blackness
bliss
blissfully
blow away
blue
blues
bother
bothered
bowl over
bowled over
brightness
broken-hearted
broodily
broodiness
broody
browned-off
bruised
brutalise
bug
bullishness
buoyancy
buoyant
buoyantly
burdened
bursting (to do)
butterflies
caginess
care
carried away (get/be)
cast down
chafe
chagrin
chagrined
champing at the bit
chariness
charmed
chasten
cheer
cheer up
cheerful
cheerfully
cheerfulness
cheeriness
cheery
cheese off
cheesed off
cherish
chill
chill
chipper
chirpy
choler
choleric
chuffed

Noun
Noun
adverb
Verb
Adjective
Noun
Verb
Adjective
Verb
Adjective
Noun
Adjective
adverb
Noun
Adjective
Adjective
Adjective
Verb
Verb
Noun
Noun
Adjective
adverb
Adjective
Adjective
Noun
Noun
Verb
Adjective
Verb
Verb
Noun
Adjective
Adjective
Noun
Adjective
Verb
Noun
Verb
Adjective
adverb
Noun
Noun
Adjective
Verb
Adjective
Verb
Noun
Verb
Adjective
Adjective
Noun
Adjective
Adjective

25

circumspection
cock-a-hoop
compassion
compulsion
compunction
concern
concerned
confound
confusion
conscience-stricken
consternation
contempt
content
content
contented
contentment
contrite
contriteness
contrition
covet
covetous
cow
cowed
crabbed
crabbiness
crabby
crankiness
cranky
crave
craving
crazed
crazy
crestfallen
cross
crotchety
crush
cut up
daunt
daunted
daze
dazzle
defeatism
deflated
deify
dejected
dejection
delectation
delight
delight
delighted
delirious
demoralisation
demoralise
demoralized

Noun
Adjective
Noun
Noun
Noun
Noun
Adjective
Verb
Noun
Adjective
Noun
Noun
Adjective
Verb
Adjective
Noun
Adjective
Noun
Noun
Verb
Adjective
Verb
Adjective
Adjective
Noun
Adjective
Noun
Adjective
Verb
Noun
Adjective
Adjective
Adjective
Adjective
Adjective
Noun
Adjective
Verb
Adjective
Noun
Verb
Noun
Adjective
Verb
Adjective
Noun
Noun
Noun
Verb
Adjective
Adjective
Noun
Verb
Adjective

demotivate
deprecatory
depress
depressed
depression
desire
desire
desirous
desirously
desirousness
desolate
desolately
desolation
despair
despairing
desperate
desperately
desperation
despise
despondency
despondent
detest
detestation
devastate
devastated
devotedly
devotion
disaffect
disappoint
disappointed
disappointment
disapprobation
disapproval
disapprove
disarm
discombobulate
discomfit
discomfited
discomfiture
discomfort
discomposure
disconcert
disconcerted
disconsolate
discontent
discontented
discontentment
discourage
discouraged
discouragement
disdain
disdain
disenchant
disenchanted

Verb
Adjective
Verb
Adjective
Noun
Noun
Verb
Adjective
adverb
Noun
Adjective
adverb
Noun
Noun
Adjective
Adjective
adverb
Noun
Verb
Noun
Adjective
Verb
Noun
Verb
Adjective
adverb
Noun
Verb
Verb
Adjective
Noun
Noun
Noun
Verb
Verb
Verb
Verb
Adjective
Noun
Noun
Noun
Verb
Adjective
Adjective
Noun
Adjective
Noun
Verb
Adjective
Noun
Noun
Verb
Verb
Adjective

26

disenchantment
disfavour
disgruntle
disgruntled
disgust
disgust
disgusted
dishearten
disheartened
disillusioned
disillusionment
dislike
dislike
dismay
dismayed
dispirited
displease
displeased
displeasure
disquiet
disquieted
disrespect
dissatisfaction
dissatisfied
dissatisfy
distaste
distraught
distress
distress
distressed
disturb
disturbed
doldrums
doleful
dolefulness
doting
dotty (about)
down
down in the mouth
downcast
downhearted
dread
dread
drive insane
drive round the bend
drive up the wall
druthers
dying for
dying to
eager
eagerness
ecstasy
ecstatic
edgily

Noun
Noun
Verb
Adjective
Noun
Verb
Adjective
Verb
Adjective
Adjective
Noun
Noun
Verb
Noun
Adjective
Adjective
Verb
Adjective
Noun
Noun
Adjective
Noun
Noun
Adjective
Verb
Noun
Adjective
Noun
Verb
Adjective
Verb
Adjective
Noun
Adjective
Noun
Adjective
Adjective
Adjective
Adjective
Adjective
Adjective
Noun
Verb
Verb
Verb
Verb
Noun
Adjective
Adjective
Adjective
Noun
Noun
Adjective
adverb

edginess
edgy
elate
elated
elation
electrified
embarrass
embarrassed
embarrassment
embitter
embittered
enamoured
enchanted
enchantment
enjoy
enjoyment
enmity
enrage
enraged
enrapture
enraptured
enthral
enthused
enthusiasm
enthusiastic
entrance
envious
enviousness
envy
envy
esteem
esteem
euphoria
euphoric
exaltation
exasperate
exasperated
exasperation
excited
excitedly
excitement
exhilarate
exhilarated
exhilaration
expectancy
expectant
expectantly
expectation
exultant
exultation
fall for
fanatical
fancy
fascination

Noun
Adjective
Verb
Adjective
Noun
Adjective
Verb
Adjective
Noun
Verb
Adjective
Adjective
Adjective
Noun
Verb
Noun
Noun
Verb
Adjective
Verb
Adjective
Verb
Adjective
Noun
Adjective
Verb
Adjective
Noun
Noun
Verb
Noun
Verb
Noun
Adjective
Noun
Verb
Adjective
Noun
Adjective
adverb
Noun
Verb
Adjective
Noun
Noun
Adjective
adverb
Noun
Adjective
Noun
Verb
Adjective
Verb
Noun

27

faze
fear
fearful
fearfully
fearfulness
fed up
feel for
feeling
felicity
fervent
fervid
fervour
fever
fevered
feverish
fired-up
flabbergast
flabbergasted
flap (in a)
floor
flummox
flummoxed
fluster
flustered
flutter
fondness
forlorn
forlornly
fractious
frantic
fraught
frenzied
frenzy
fret
fright
frighten
frightened
frown on
frustrate
frustrated
frustration
fulfilled
fuming
furious
furiously
furiousness
fury
fuss
gaiety
gaily
gall
gasping for
get on sb’s nerves
get sb’s back up

Verb
Noun
Adjective
adverb
Noun
Adjective
Verb
Noun
Noun
Adjective
Adjective
Noun
Noun
Adjective
Adjective
Adjective
Verb
Adjective
Noun
Verb
Verb
Adjective
Verb
Adjective
Noun
Noun
Adjective
adverb
Adjective
Adjective
Adjective
Adjective
Noun
Verb
Noun
Verb
Adjective
Verb
Verb
Adjective
Noun
Adjective
Adjective
Adjective
adverb
Noun
Noun
Verb
Noun
adverb
Verb
Adjective
Verb
Verb

glad
gladden
gladdened
gladness
glee
gleeful
gleefulness
gloom
gloomily
gloominess
gloomy
glory in
glum
glumness
gnaw (at so)
gob-smacked
good humour
good temper
goodwill
grate
grateful
gratefully
gratefulness
gratification
gratified
gratify
gratitude
green-eyed
grief
grieve
grind down
grouchy
grudge
grudge
grumpiness
grumpy
guilt
guiltiness
guilt-ridden
guilty
guilty conscience
gusto
gutted
hacked off
hanker/hanker after
hankering
happily
happiness
happy
harrassed
hate
hate
hatred
haunt

Adjective
Verb
Adjective
Noun
Noun
Adjective
Noun
Noun
adverb
Noun
Adjective
Verb
Adjective
Noun
Verb
Adjective
Noun
Noun
Noun
Verb
Adjective
adverb
Noun
Noun
Adjective
Verb
Noun
Adjective
Noun
Verb
Verb
Adjective
Noun
Verb
Noun
Adjective
Noun
Noun
Adjective
Adjective
Noun
Noun
Adjective
Adjective
Verb
Noun
adverb
Noun
Adjective
Adjective
Noun
Verb
Noun
Verb

28

have it in for
head over heels in love
heartache
heartbreak
heartbroken
heavy-hearted
heavy-laden
heebie-jeebies
heedful
hero-worship
het up
high spirits
hold dear
homesick
homesickness
honour
honoured
hope
hopefulness
hopeless
hopelessly
hopelessness
horrified
horrify
horror
horror-stricken
horror-struck
hostility
hot under the collar
huffy
humble
humbled
humiliate
humiliated
humiliation
hung up
hunger
hungry (for)
hurt
hurt
hurt
hurting
idealise
identification (with)
identify with
idolise
ill at ease
ill humour
ill will
ill-disposed
impatience
impatience
impatient
impress

Verb
Adjective
Noun
Noun
Adjective
Adjective
Adjective
Noun
Adjective
Verb
Adjective
Noun
Verb
Adjective
Noun
Verb
Adjective
Noun
Noun
Adjective
adverb
Noun
Adjective
Verb
Noun
Adjective
Adjective
Noun
Adjective
Adjective
Verb
Adjective
Verb
Adjective
Noun
Adjective
Noun
Adjective
Noun
Adjective
Verb
Adjective
Verb
Noun
Verb
Verb
Adjective
Noun
Noun
Adjective
Noun
Noun
Adjective
Verb

impressed
Adjective
impulsion
Noun
in a funk
Adjective
in a huff
Noun
in a state
Adjective
in disbelief
Adjective
in good spirits
Adjective
in high dudgeon
Adjective
in seventh heaven
Adjective
in the dumps [down in ...] Adjective
incense
Verb
incensed
Adjective
inclination
Noun
incomprehension
Noun
inconsolable
Adjective
incredulity
Noun
incredulous
Adjective
indebted
Adjective
indignant
Adjective
indignation
Noun
infatuated
Adjective
infatuation
Noun
inflame
Verb
infuriate
Verb
infuriated
Adjective
inhibition
Noun
insecurely
adverb
insecurity
Noun
intent
Adjective
intimidate
Verb
intimidated
Adjective
irate
Adjective
ire
Noun
ireful
Adjective
irk
Verb
irked
Adjective
irritability
Noun
irritable
Adjective
irritate
Verb
irritated
Adjective
irritation
Noun
itch
Noun
itch
Verb
itching
Adjective
jar
Verb
jauntiness
Noun
jealous
Adjective
jealousy
Noun
jitteriness
Noun
jitters
Noun
jittery
Adjective
joie de vivre
Noun
jolliness
Noun
jollity
Noun

29

joviality
Noun
joy
Noun
joyful
Adjective
joyfully
adverb
joyfulness
Noun
joylessness
Noun
joyous
Adjective
joyousness
Noun
jubilant
Adjective
jubilation
Noun
jumpily
adverb
jumpiness
Noun
jumpy
Adjective
keen
Adjective
keenness
Noun
keyed-up
Adjective
knock your socks off
Verb
knocked out
Adjective
lap up
Verb
lather (in a)
Noun
leaning
Noun
let down
Verb
light-hearted
Adjective
lightheartedly
adverb
like a cat on a hot tin roof Adjective
like a cat on hot bricks Adjective
liking
Noun
livid
Adjective
lividly
adverb
loathe
Verb
loathing
Noun
long
Verb
longing
Noun
look up to
Verb
love
Noun
love
Verb
lovesick
Adjective
love -struck
Adjective
low
Adjective
low spirits
Adjective
lugubriousness
Noun
lust
Noun
lust
Verb
luxuriate
Verb
mad
Adjective
madden
Verb
maddened
Adjective
make sb’s blood boil
Verb
make sb’s hackles rise Verb
malaise
Noun
malcontent
Adjective
malice
Noun
manic
Adjective
melancholic
Adjective

melancholy
merriment
merry
miff
miffed
mind
mirthful
miserable
miserably
misery
misgiving(s)
miss
mistrust
mistrustful
moonstruck
morose
moroseness
mortification
mortified
mortify
mournful
mournfulness
mystification
mystified
mystify
nag
narked
needle
nerves
nerviness
nervous
nervousness
nervy
nettle
nettled
niggle (at so)
nonplus
nonplussed
nostalgia
obeisance
obligated
odium
offence/offense
offend
offended
on cloud nine
on edge
on tenterhooks
on the warpath
oppress
optimism
out of sorts
outrage
outrage

Noun
Noun
Adjective
Verb
Adjective
Verb
Adjective
Adjective
adverb
Noun
Noun
Verb
Verb
Adjective
Adjective
Adjective
Noun
Noun
Adjective
Verb
Adjective
Noun
Noun
Adjective
Verb
Verb
Adjective
Verb
Noun
Noun
Adjective
Noun
Adjective
Verb
Adjective
Verb
Verb
Adjective
Noun
Noun
Adjective
Noun
Noun
Verb
Adjective
Adjective
Adjective
Adjective
Adjective
Verb
Noun
Adjective
Noun
Verb

30

outraged
over the moon
overawe
overawed
overexcite
over-excited
overjoyed
overstrung
overwhelmed
overwrought
pain
pain
panic
panic
panicked
panicky
panic-stricken
paranoid
partial to
partiality (for)
passion
passionate
passionately
peeved
peevish
peevishness
penchant for
penitent
perplex
perplexed
perplexity
perturb
perturbation
perturbed
pessimism
petrified
petrify
petulance
petulant
pine
pique
pique
piqued
pity
pity (so)
please
pleased
pleasure
possessiveness
predilection
prepared
pressured
prey (on)
pride

Adjective
Adjective
Verb
Adjective
Verb
Adjective
Adjective
Adjective
Adjective
Adjective
Noun
Verb
Noun
Verb
Adjective
Adjective
Adjective
Adjective
Adjective
Noun
Noun
Adjective
adverb
Adjective
Adjective
Noun
Noun
Adjective
Verb
Adjective
Noun
Verb
Noun
Adjective
Noun
Adjective
Verb
Noun
Adjective
Verb
Noun
Verb
Adjective
Noun
Verb
Verb
Adjective
Noun
Noun
Noun
Adjective
Adjective
Verb
Noun

pride yourself on
prize
protectiveness
proud
psych
psyched up
punch-drunk
put off
put sb’s back up
puzzle
puzzled
puzzlement
qualm
queasy
rabid
rabidly
rage
rancorous
rancour
rancourousness
rankle
rapture
rattily
rattle
ratty
regard
regret
regretful
regretfully
relish
relish
reluctance
remorse
remorseful
repel
repentance
repentant
repugnance
repulsion
resent
resentfulness
resentment
resignation
respect
respect
respectfulness
revel in
revere
reverence
revolt
revulsion
rile
riled
romanticise

Verb
Verb
Noun
Adjective
Verb
Adjective
Adjective
Verb
Verb
Verb
Adjective
Noun
Noun
Adjective
Adjective
adverb
Noun
Adjective
Noun
Noun
Verb
Noun
adverb
Verb
Adjective
Noun
Noun
Adjective
adverb
Noun
Verb
Noun
Noun
Adjective
Verb
Noun
Adjective
Noun
Noun
Verb
Noun
Noun
Noun
Noun
Verb
Noun
Verb
Verb
Noun
Verb
Noun
Verb
Adjective
Verb

31

rotten
rub up the wrong way
rueful
ruffle
sad
sadden
saddened
sadly
sadness
satisfaction
satisfied
satisfy
savour
scandalise
scandalised
scare
scare
scared
scourge
seething
self-contempt
self-disgust
self-dislike
self-doubt
self-hatred
self-loathing
self-pity
self-pitying
self-reproach
send over the edge
shake up
shaken
shame
shame
shamefaced
shattered
sheepish
sheepishness
shock
shock
shocked
show up
sick
sicken
sickened
smitten
soft spot
sold on
solicitude
sombre
sombrely
sombreness
sore
sorrow

Adjective
Verb
Adjective
Verb
Adjective
Verb
Adjective
adverb
Noun
Noun
Adjective
Verb
Verb
Verb
Adjective
Noun
Verb
Adjective
Verb
Adjective
Noun
Noun
Noun
Noun
Noun
Noun
Noun
Adjective
Noun
Verb
Verb
Adjective
Noun
Verb
Adjective
Adjective
Adjective
Noun
Noun
Verb
Adjective
Verb
Adjective
Verb
Adjective
Adjective
Noun
Adjective
Noun
Adjective
adverb
Noun
Adjective
Noun

sorrow
sorrowful
sorrowfully
sorrowfulness
sorrowing
sour grapes
sourness
spite
spitefulness
spleen
spook
squirmy
stagger
staggered
starry-eyed
startle
startled
stirred up
straining at the leash
stress
stress out
stressed
strop
stroppiness
strung up
stumped
stun
stung
stunned
stupefaction
stupefied
stupefied
stupefy
suicidal
sulk
sulkiness
sulky
surprise
surprise
surprised
suspense
swear by sth
sweep so off their feet
sympathy
take a dim view of
take a fancy to
take a shine to
take aback
take exception
taken aback
taken with
tear apart
teed off
temptation

Verb
Adjective
adverb
Noun
Adjective
Noun
Noun
Noun
Noun
Noun
Verb
Adjective
Verb
Adjective
Adjective
Verb
Adjective
Adjective
Adjective
Noun
Verb
Adjective
Noun
Noun
Adjective
Adjective
Verb
Adjective
Adjective
Noun
Adjective
Adjective
Verb
Adjective
Noun
Noun
Adjective
Noun
Verb
Adjective
Noun
Verb
Verb
Noun
Verb
Verb
Verb
Verb
Verb
Adjective
Adjective
Verb
Adjective
Noun

32

tense
tensely
tension
terrified
terrify
territorial
terror
terrorise
terrorised
terror-stricken
testy
tetchy
thankful
thankfully
thankfulness
thirst
thirst
thirsty
thrill
thrilled
thrown off balance
thunderstruck
ticked off
tickle
tickled
tired
tizzy
torment
torment
tormented
torture
torture
touched
traumatise
traumatized
treasure
trepidation
triumph
triumphant
trouble
troubled
twitchily
twitchiness
twitchy
umbrage
unease
uneasiness
uneasy
unglued
unhappily
unhappiness
unhappy
unhopeful
unnerve

Adjective
adverb
Noun
Adjective
Verb
Adjective
Noun
Verb
Adjective
Adjective
Adjective
Adjective
Adjective
adverb
Noun
Noun
Verb
Adjective
Verb
Adjective
Adjective
Adjective
Adjective
Verb
Adjective
Adjective
Noun
Noun
Verb
Adjective
Noun
Verb
Adjective
Verb
Adjective
Verb
Noun
Noun
Adjective
Verb
Adjective
adverb
Noun
Adjective
Noun
Noun
Noun
Adjective
Adjective
adverb
Noun
Adjective
Adjective
Verb

unnerved
unquiet
unsatisfied
unsettle
unsettled
up in arms
upbeat
uplifted
upset
upset
uptight
urge
value
venerate
veneration
vengeful
venom
vex
vexation
vexed
vindictive
vindictiveness
wallow in
wanderlust
wariness
warm/warm to
warmness
waspish
weakness
weigh down
well-disposed
whim
wholehearted
wilfulness
will
willies
willing
willingness
wired
wish
wistfulness
with bated breath
woe
woeful
wonder
work up
worked-up
world-weariness
worried
worry
worry
worship
worship
worshipful

Adjective
Noun
Adjective
Verb
Adjective
Adjective
Adjective
Adjective
Noun
Verb
Adjective
Noun
Verb
Verb
Noun
Adjective
Noun
Verb
Noun
Adjective
Adjective
Noun
Verb
Noun
Noun
Verb
Noun
Adjective
Noun
Verb
Adjective
Noun
Adjective
Noun
Noun
Noun
Adjective
Noun
Adjective
Noun
Noun
adverb
Noun
Adjective
Noun
Verb
Adjective
Noun
Adjective
Noun
Verb
Noun
Verb
Adjective

33

wound
wounded
wrath
wrathful
wretched
wretchedly
wretchedness
wrought-up
yearn
yearning
yearningly
yen (have a for)
zeal
zest
zing

Verb
Adjective
Noun
Adjective
Adjective
adverb
Noun
Adjective
Verb
Noun
adverb
Noun
Noun
Noun
Noun

34

Notes

1

For research on affect/emotion and gender see Lutz (1990), Gallois (1994: 306-307),
Anderson & Leaper (1998), Planalp (1999: 36), Goldshmidt & Weller (2000), Galasinski
(2004), Oatley et al (2006: 246-248); on affective communication and social groups see
Besnier (1990), Irvine (1990), Gallois (1994).

2

Oakey (2002) discusses some of the advantages/disadvantages of this approach: ‘The
sampling methods used in its [the BNC] construction meant that partial texts were included, and thus some areas of discourse may not be represented equally. Nonetheless, it
is fair to say that a subset of this size yields sufficient occurrences for meaningful comparisons to be made between sub-genres.’ (Oakey 2002: 115-116)

3

Abbreviations: Hu = humanities, me = medicine, ns = natural sciences, sos = social
sciences, as = applied science, ple = politics, law, education; bs = broadsheets, art =
arts/cultural material, fin = finance/commerce, nw = home/foreign news, sc = science, lst
= lifestyle etc, sp = sports, tbl = tabloid, uk = unknown.

4

Crucially, there is no assumption that one ‘emotion term’ corresponds to one ‘emotion’
as defined in psychology or biology. Rather, emotion terms are ‘ways of speaking’
(Galasinski 2004: 6) or ‘discursive phenomena’ (Edwards 1999: 279), and ‘[l]exis provides categories of affect in the form of folk taxonomies.’ (Downes 2000: 108).

5

‘There is certainly a core of prototypical emotions on which subjects will largely agree
that they are emotions. But there is a large border area where judgements differ.’ (translation mine)

6

In retrospect, some of the words in the categories involving ‘interest’ could have been
added to the list, as they can be dealt with under the affect sub-categories of
dis/satisfaction: interest and ennui, even though these categories ‘take us to the borders of
affect’ (Martin & White 2005: 50, emphasis in original).

7

‘A run-on entry is […] a word morphologically derived from a dictionary headword
which is not itself defined, but is printed, along with its word class and possibly an
example, usually in a bold typeface, at the end of the entry for the word it is derived
from’ (Kilgarriff 1997a: 153).

8

6353 occurrences in conversation, 5347 occurrences in fiction.

9

13,060 occurrences in conversation, 10,461 occurrences in fiction.

10 10,787 occurrences in conversation.

35