2
Key aspects
Colouring COMPARA
contrastive and monolingual
colour studies in
English and Portuguese
Rosário Silva
Susana Inácio
Diana Santos
COMPARA
„ Colouring COMPARA
„ English-speaking authors
„ Contrastive analysis (Portuguese vs.
English)
„ Morphosyntactic patterns in Portuguese
and in English
„ Originals vs. Translations
„
Colouring COMPARA
3
4
COMPARA
COMPARA
COMPARA contents
An English-Portuguese publicly available
parallel corpus (since 2001)
„ developed at Linguateca, project leaders:
„
„
Ana Frankenberg-Garcia & Diana Santos
„
manually revised texts, alignment and
grammatical annotation (from PALAVRAS
„
several innovative features
and CLAWS, in progress)
75 pairs, corresponding to 72 fiction works
„ 1.436 million words in Portuguese (Africa,
Brazil and Portugal)
„ 1.549 million words in English (GreatBritain, America and South Africa)
„ 97,728 alignment units
„ see http://www.linguateca.pt/COMPARA/
Colouring COMPARA
5
„
6
Colouring COMPARA
Colour research by:
Language
Text
„ Author
„ Translated or Original text(s)
„ Date of publication
„
… Portuguese
… English
„
Colouring COMPARA
Colouring COMPARA
Colour research by:
COMPARA
Language variety
… Portuguese
(Portugal, Brazil, Mozambique
and Angola)
… English (United Kingdom, United States and
South Africa)
Colouring COMPARA
Several outputs available (concordances
and distributions)
Colouring COMPARA
1
7
Why COLOUR?
„
„
„
Colouring COMPARA
9
Colouring COMPARA
COLOUR categories and groups
„
„
Colour
„
„
colour:race
„ colour:human
„ colour:wine
„ colour:original
„
Colouring COMPARA
What is COLOUR? – Criteria
We wanted to start semantic annotation of
COMPARA
An easy to grasp and relatively-contained lexical
category
The subject of many arguments in linguistics
A frequent category, both literally and
metaphorically
„
8
Colouring COMPARA
„
„
„
„
„
„
Beige
Black
Blue
Brown
Gold
Green
Grey
Orange
Pink
„
„
„
„
„
„
„
„
Purple
Red
Silver
White
Yellow
Other
Multiple
Unspecified
All straightforward words conveying
colour, regardless of word class
„ Words that convey colour but are not in
themselves a colour
„ Words that have colours in them but have
gone beyond the mere colour reference
„ The word “colour” and its derivations.
„
For more information see
http://www.linguateca.pt/COMPARA/DocAnotacaoCorCOMPARA.pdf (continually updated)
10
Colour categories - examples
colour amber-coloured; blackened; lime-green; redand-white; watery-blue; white; yellowing...
colour:race whites; blacks; white-skinned; coloureds...
colour:human blond; blushed; brunette; red-haired;
rosy-cheeked; suntanned; whitebearded...
colour:wine white; red.
colour:original blacklist; bluebirds; brownies;
whitewash; eggwhite...
Colouring COMPARA
11
Colouring COMPARA
Instances marked as colour in COMPARA:
4,609 in PT and 4,548 in EN
0.0032%
„
Colouring COMPARA
12
English-speaking authors
ENGLISH-SPEAKING AUTHORS IN COMPARA
Quantitative data
„
Colouring COMPARA
0.0029%
Number of different colours in COMPARA:
409 in PT (209 lemmas) and 427 in EN
Colouring COMPARA
Julian Barnes
Lewis Carrol
„ Joseph Conrad
„ Nadine Gordimer
„ Joseph Heller
„ Kazuo Ishiguro
„ Henry James
„
„
David Lodge
Ian McEwan
„ Mary Shelley
„ Joanna Trollope
„ Oscar Wilde
„ Richard Zimler
„
„
Colouring COMPARA
2
13
English-speaking authors
English Authors
Texts and Dates
# types
# words
14
Contrastive analysis (EN vs. PT)
EN
frequency
freq.
12
Barnes
EBJB1 (1984) e EBJB2 (1989) e EBJB3
(1991)
46
121
2,02
Carrol
EBLC1 (1871)
7
11
0,99
Conrad
EBJC1 (1902)
19
41
3,37
ESNG1 (1990) e ESNG2 (1979) e ESNG3
(1981) e ESNG4 (1974)
92
617
7,01
EUJH1 (1990)
12
25
1,89
2
EBKI1 (1995) e EBKI2 (1989)
23
69
0,81
0
James
EUHJ1 (1880) e EUHJ2 (1895) e EUHJ3
(1893)
22
65
1,38
Lodge
EBDL1 (1995) e EBDL2 (1989) e EBDL3
(1975) e EBDL4 (1980) e EBDL5 (1991)
e EBDL6 (1984)
95
534
2,90
McEwan
EBIM1 (1992) e EBIM2 (1998) e EBIM3
(1978)
44
103
2,47
10
Wilde
Zimler
Trollope
Poe
Shelley
McEwan
James
Heller
Ishiguro
Gordimer
Carrol
Conrad
Lodge
PT
frequência
12
10
Poe
EUEP1 (1837)
7
20
0,82
Shelley
EBMS1 (1818)
6
12
0,57
101
312
3,70
EBJT1 (1996) e EBJT2 (1993) e EBJT3
(1995)
8
6
4
15
Soares
Sena
Saramago
Sá-Carneiro
Rey
Melo
Lins
Jorge
Colouring COMPARA
Machado de Assis
Fonseca
Dourado
0
Eça de Queirós
3,46
Couto
212
Coelho
63
Carvalho
Zimler
EURZ1 (1998) e EURZ2 (2000)
Cardoso Pires
6,47
Castelo Branco
166
Buarque
52
Alencar
EBOW1 (1890)
Almeida
2
Wilde
Azevedo
Trollope
Barnes
Ishiguro
4
permilagem
Heller
6
Agualusa
Gordimer
per thousand
8
Colouring COMPARA
16
Most colourful authors
English-speaking authors
Colour categories (1)
(no. of different colour types)
Zimler
102
Joanna Trollope
51
blue, white, black, red,...
scarlet, hazel, peacockblue, saffron-yellow,
purplish, red-and-ochre,
emerald-green...
Wilde
Trollope
preto, branco, negro,
verde,... esbranquiçado,
pratear, verde-garrafa,
azuleleca, multicor,
rubro, alvaiado,....
Autran
Dourado
Shelley
Poe
McEwan
Lodge
James
Least colourful authors
Ishiguro
(no. of different colour types)
Heller
Gordimer
6
6
black, blue, green,
white, red, blackness.
negro, verde,
amarelo, azul, cor,
preto.
Camilo
Conrad
Carrol
Barnes
0,0%
Castelo Branco
Mary Shelley
0,1%
0,2%
0,3%
0,4%
% colour / author total words
% colour:human / author total words
0,5%
0,6%
Colouring COMPARA
17
Contrastive analysis (EN vs. PT)
EN
Zimler
Wilde
Sena
Saramago
Poe
Sá-Carneiro
18
English-speaking authors
Authors’ favourite colour groups
PT
Soares
Shelley
(Each author contributed only with his/her favourite colour group)
Blue
7%
McEwan
Rey
Lodge
James
Melo
Ishiguro
Machado de Assis
Heller
Red
14%
Lins
Gordimer
Conrad
Jorge
Carrol
Fonseca
Barnes
0,0%
0,8%
Colouring COMPARA
Colour categories (1)
Trollope
0,7%
% colour:race / author total words
% colour:wine / author total words
0,1%
0,2%
0,3%
0,4%
% colour / author total words
% colour:human / author total words
0,5%
0,6%
% colour:race / author total words
% colour:wine / author total words
0,7%
0,8%
Eça de Queirós
Dourado
White
43%
Couto
Coelho
Castelo Branco
Carvalho
Cardoso Pires
Buarque
Azevedo
Almeida
Alencar
Agualusa
0,0%
% cor / total palavras autor
0,2%
0,4%
% cor:raça / total palavras autor
0,6%
0,8%
% cor:humana / total palavras autor
1,0%
1,2%
% cor:vinho / total palavras autor
Colouring COMPARA
Black
36%
Colouring COMPARA
3
19
English-speaking authors
20
Contrastive analysis (EN vs. PT)
Author with the highest individual percentage of the colour:
Red
9% Green
9%
Unspecified
18%
White
8%
Blue
25%
Blue
7%
Brown
9%
Pink
9%
PT
EN
Red
14%
Black
42%
Red
8%
White
45%
Lewis Carrol
Favourite colour groups
(Each author contributed only with his/her favourite colour group)
Vermelho
5%
White
43%
Green
17%
Mary Shelley
Não
Especificada
9%
Black
36%
Blue
2%
Unspecified
11%
Red
29%
Gold
8%
Purple
1%
Green
6%
White
16%
Henry James
Black
8%
Red
13%
Pink
5%
Beige
2%
Preto
27%
Yellow
5%
Black
11%
White
12%
Brown
4%
Branco
49%
Blue
15%
Unspecified
4%
Other
2%
Silver
0,3%
Grey
6%
Grey
13%
Pink
2%
Multiple
6%
Gold
2%
Amarelo
5%
Azul
5%
Brown
8%
Orange
0,3%
Green
8%
Joanna Trollope
Colouring COMPARA
21
Colouring COMPARA
Portuguese-speaking authors
22
English-speaking authors
Author with the highest individual percentage of the colour:
Múltipla
5%
Não
especificada
13%
Azul
14%
Vermelho
5%
Amarelo
13%
Branco
43%
14%
13,2%
12,7%
14,0%
12%
Preto
50%
Preto
5%
José de Alencar
16%
Verde
13%
Verde
29%
All colour groups
(The authors contributed with all their colours)
Azul
13%
Camilo
10%
Castelo Branco
8%
9,0%
7,3%
7,2%
6,8%
6,0%
5,2%
5,1%
6%
3,5%
4%
2,3%
2,6%
1,4%
1,3%
1,3%
1,1%
2%
Jorge de Sena
Marcos Rey
24%
44%
low
red
31%
gre
en
ora
ng
e
br
ow
n
b la
ck
wh
ite
pu
rp
le
be
ige
p in
k
gre
y
silv
er
go
ld
o th
er
mu
ltip
Un
spe le
cif
ied
José Eduardo
Agualusa
26%
yel
b lu
e
0%
Mia Couto
Colouring COMPARA
23
Colouring COMPARA
Contrastive analysis (EN vs. PT)
24
All colour groups
English-speaking authors
Races (colour:race)
(The authors contributed with all their colours)
EN
25%
Zimler (2)
20%
15%
13,2%
12,7%
14,0%
Trollope (2)
9,0%
10%
PT
7,3%
7,2%
6,8%
5,2%
5,1%
6,0%
5%
Poe (3)
3,5%
2,3%
1,1%
1,3%
2,6%
1,4%
1,3%
Uns
gold
McEwan (5,5)
oth
er
mu
ltip
le
pec
ifie
d
k
grey
si lv
er
te
pin
bei
ge
pu
rp
le
ge
whi
bla
ck
bro
wn
oran
e
red
gree
n
blu
21,1%
yell
ow
0%
25%
20%
Lodge (22)
15,4%
15%
11,8%
James (0,5)
10,7%
10%
9,0%
7,8%
Gordimer (336)
5,0%
4,2%
5%
0,7%
2,8%
1,9%
0,6%
1,7%
2,3%
1,3%
3,5%
Conrad (6,5)
0,2%
Ou
t ra
s
Mú
lt ip
Nã
la
oe
sp e
ci fi
cad
a
Do
ur a
do
to
zen
ate
ad
o
Pr
Ro
sa
C in
Ro
xo
Cr
em
e
Pr
eto
Br
an
co
ran
ja
Ca
sta
nh
o
Ve
rd
e
La
Am
are
lo
Az
ul
Ve
rm
elh
o
0%
Colouring COMPARA
Barnes (2)
0%
9 out of 14
10%
20%
30%
White race
40%
50%
Black race
60%
70%
80%
90%
100%
Other races/Undefined
Colouring COMPARA
4
25
26
Contrastive analysis (EN vs. PT)
English-speaking authors
Races (colour:race)
Soares (8)
Does colour quantity change with time?
Number of colour types per text dates
Zimler (2)
Rey (6)
Trollope (2)
Melo (19)
Poe (3)
60
EN
McEwan (5,5)
Machado de Assis (11)
Lodge (22)
50
Osman Lins (1)
James (0,5)
Jorge (20)
Gordimer (336)
40
Fonseca (17)
Conrad (6,5)
Eça de Queirós (1)
Barnes (2)
0%
10%
EN
Dourado (145)
20%
30%
White race
40%
50%
Black race
60%
70%
80%
90%
100%
30
Other races/Undefined
Couto (15)
20
Buarque (7)
Azevedo (54)
10
Alencar (9)
Agualusa (5)
0
0%
10%
20%
PT
30%
40%
Raça Branca
50%
Raça Negra
60%
70%
80%
90%
100%
1815
14 out of 22
Outras raças/Indefinida
1825
1835
1845
1855
1865
1875
1885
1895
1905
1915
1925
1935
1945
1955
Colouring COMPARA
27
1965
1975
1985
1995
2005
Colouring COMPARA
Portuguese-speaking authors
28
Does colour quantity change with time?
English-speaking authors
Does colour quantity change with time?
Number of colour types per text dates
Number of colour types per authors’ birth date
40
PT
120
EN
35
1943 - Trollope
100
1935 - Lodge
30
1923 - Gordimer
80
25
1956 - Zimler
20
60
1854 - Wilde
15
1946 - Barnes
40
1948 - McEwan
10
1954 - Ishiguro
1843 - James
20
5
1857 - Conrad
1832 - Carroll
1797 - Shelley
1923 - Heller
1809 - Poe
0
1850
1860
1870
1880
1890
1900
1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
0
1790 1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960
Colouring COMPARA
29
Portuguese-speaking authors
Colouring COMPARA
30
Morphosyntactic Patterns in English
Grammatical Category of Colours
Does colour quantity change with time?
Number of colour types per authors’ birth date
60
PT
1926 - Dourado
50
„ NOUN: 932 words
(...came back in the afternoon gold.)
(...hand-painted in orange.)
1890 - Sá-Carneiro
40
1845 - Eça de Queirós
1857 - Azevedo
1924 - Lins
30
1925 - Fonseca
1944 - Buarque
1962 - Melo
1922 - Saramago
1919 - Sena
1839 - Machado de Assis
20
1925 - Cardoso Pires
1925 - Rey
1944 - Carvalho
1938 - Soares
10
1946 - Jorge
1947 - Coelho
1955 - Couto
1831 - Almeida
1960 - Agualusa
0
1820
1830
1840
1850
1860
1870
1880
1890
1900
1910
1920
1930
1940
„ ADJECTIVE: 3,721 words
(red-hot needle, white image, blue eyes,...)
1950
„ VERB: 108 words
(...she smiled and blushed.)
(...eyes blackened and engorged by fatigue...)
„ ADVERB: 7 word
(...she smiled rosily and said...)
1960
Colouring COMPARA
Colouring COMPARA
5
31
Syntactic Patterns in Portuguese
32
Grammatical Category of Colours
„
COLOUR + NOUN (1,138)
ADJECTIVE: 3,193 words (161 lemmas)
(carro verde, fato azul)
„
NOUN: 1,349 words (93 lemmas)
(...o verde era uma cor que..., ...estou em pleno azul, suspenso...)
VERB: 107 words (15 lemmas)
„
(...muito além daquela serra, que ainda azula no horizonte...)
(...as saladas esverdinhavam à hora do jantar.)
„
ADVERB: 1 word (1 lemma)
(purpureamente)
„
„
„
„
„
„
„
„
„
ADJn: 29 words (16 lemmas)
(-- No entanto gosto do azul -- prosseguiu, a cliente -- e o verde é muito bonito. )
ADJ_N: 27 words (5 lemmas)
(Mas Isidoro, mesmo preto, não era capaz de uma coisa dessas.)
Morphosyntactic Patterns in English
Morphosyntactic Patterns using Colour
„
„
hair 49 (brown hair, reddish hair, wine-red hair...)
eye 45 (red eye, watery-blue eyes,...)
silk 22 (crimson silk, amber-coloured silk,...)
paper 16 (yellowing paper, brown paper,...)
light 16 (purple light, greyish-blue light,...)
velvet 11 (black velvet, purple velvet,...)
sky 11 (indigo sky, grey sky,...)
plastic 11 (pale-grey plastic, white plastic,...)
bear 10 (brown bear, black bear,...)
Colouring COMPARA
Colouring COMPARA
Data from EN originals only; without colour:original, colour:race, colour:human and colour:wine
33
Morphosyntactic Patterns in Portuguese
34
Morphosyntactic Patterns using Colour
Morphosyntactic Patterns in Portuguese
Morphosyntactic Patterns using Colour
NOUN + COLOUR (914)
„
„
„
„
„
„
„
„
COLOUR + NOUN (45)
cabelo 41 (cabelos brancos, cabelo negro,...)
olho 32 (olhos azuis, olhos castanhos, olhos verdes,...)
luz 21 (luz vermelha, luz anilada,...)
seda 20 (seda verde, seda roxa,...)
papel 19 (papel colorido, papel amarelo,...)
mármore 13 (mármore negro, mármore verde,...)
vestido 10 (vestido azul, vestido prateado,...)
céu 9 (céu azul, céu cinzento, céu negro,...)
„
„
„
„
„
„
„
mar 6 (verdes mares...)
praia 2 (alvas praias...)
areia 2 (brancas areias...)
cor 2 (alva cor...)
rede 2 (alva rede...)
fio 2 (alvos fios...)
pelúcia 2 (verde pelúcia...)
Colouring COMPARA
Data from PT originals only; without cor:original, cor:raça, cor:humana and cor:vinho
35
Colouring COMPARA
Data from PT originals only; without cor:original, cor:raça, cor:humana and cor:vinho
Morphosyntactic Patterns in English
36
Morphosyntactic Patterns in Portuguese
Morphosyntactic Patterns using Colour
Morphosyntactic Patterns using Colour
COLOUR + “and” + COLOUR (98)
COLOUR + “e” + COLOUR (32)
pink and (colour) 17
black and (colour) 10
red and (colour) 9
green and (colour) 7
white and (colour) 6
orange and (colour) 6
blue and (colour) 6
Data from EN originals only
branco e (colour) 7
azul e (colour) 5
preto e (colour) 3
ouro e (colour) 2
verde e (colour) 2
vermelho e (colour) 2
amarelo e (colour) 2
Colouring COMPARA
Data from PT originals only
Colouring COMPARA
6
37
„
„
„
„
„
„
„
Morphosyntactic Patterns in Portuguese
Morphosyntactic Patterns using Colour
Morphosyntactic Patterns using Colour
COLOUR (hyphenated words)
COLOUR (hyphenated words)
(219)
(396)
„
38
Morphosyntactic Patterns in English
Colour can appear on both sides
of the hyphen
colour-[borrowed from something] 85 (cherry-red,...)
colour-[hue modifier] 66 (deep-blue,...)
colour-[applied to something] 66 (green-shirted,...)
colour-[“colo(u)red”] 61 (carrot-colored,...)
colour-[colour] 46 (pink-brown,...)
colour-[hair-related] 44 (red-haired,...)
colour-[skin-related] 20 (olive-skinned,...)
colour-[eye-related] 8 (gray-eyed,...)
39
„
[applied to something] 1 (saliva-laranja)
„
„
Colouring COMPARA
EN-side of the corpus
„
[hue modification] 104 (verde-vivo, vermelho-escuro,...)
cor-de-... 62 (cor-de-rosa, cor-de-rato,...)
[borrowed from something] 46 (verde-musgo,...)
[several colours] 6 (verde-azul, rosa-e-azul,...)
„
PT-side of the corpus
Morphosyntactic Patterns in Portuguese
Colouring COMPARA
Hyphenated form of clitics in PT not included
40
Morphosyntactic Patterns using Colour
Colour - Originals vs. Translations
160
141
COR + DE + ...
(105)
140
120
100
80
60
„
„
„
„
„
laranja 12
azeitona 8
areia 7
cobre 4
prata 4
40
20
0
8
1
0
0
-6
-20
-22
-40
Cor geral
-6
-5
-14
-6
T (LC)
T (JC)
T (NG)
T (JH)
T (KI)
T (HJ)
T (DL)
T (IM)
0
8
-6
-22
0
-6
-14
141
-5
T (EP) T (MS)
1
-6
Colouring COMPARA
PT-side of the corpus
41
Originals vs. Translations - EN texts
-22
-26
-7
42
Originals vs. Translations
Example: Julian Barnes
colour
46
50
-26
T (JT) T (OW) T (RZ)
Colouring COMPARA
Colour categories - Originals vs. Translations
60
-7
-22
T (JB)
40
30
colour:human
20
10
Julian Barnes
0
-20
-1
-2
-10
-1
Translator (JB)
-6
-6,5
-12,5
-13
-18
-18,5
-40
-40
-41
-50
-38
T (JB) T (LC) T (JC) T (NG) T (JH) T (KI) T (HJ) T (DL) T (IM) T (EP) T (MS) T (JT) T (OW) T (RZ)
Colour
-18
-2
-6,5
-41
-1
-13
-18,5
46
-12,5
-1
-6
-40
-38
No translation
-24
8
10
0
0
0
0
0
0
0
-5
4
-0,5
-2
-1
-2
-2,5
2
-4,5
0
branco
louro
louras
louro
bronzeado
Ø
Ø
white-haired
blonder
blond
blond
tan
silver-haired
blushes
0
0
0
0
-15
-1
-2
-1
-1
-2
-20
colour:race
-20
-25
-4
-3
-4
-4
-5
-6
-6
-30
T (JB) T (LC) T (JC) T (NG) T (JH)
colour:race
Translator (JB)
EBJB1(228):
EBJB1(240):
EBJB2(613):
EBJB3(494):
EBJB3(865):
EBJB3(881):
EBJB3(888):
Translations fall into other
colour categories
0
-10
colour:human Julian Barnes
colour:human
6
6
5
4
-3
-24
-30
7
-2
0
-2,5
-20
0
T (KI) T (HJ) T (DL) T (IM) T (EP) T (MS) T (JT) T (OW) T (RZ)
0
-0,5
0
-4,5
0
0
-2
0
-1
-8
T (JB)
colour:human
-3
T (LC) T (JC) T (NG) T (JH)
0
-1
-6
0
T (KI)
-2
T (HJ) T (DL) T (IM) T (EP) T (MS) T (JT) T (OW) T (RZ)
-1
6
-4
0
0
-4
Colouring COMPARA
-1
-5
Colouring COMPARA
7
43
Originals vs. Translations
44
Example: Julian Barnes
Red
Julian Barnes
16
Translator (JB)
9
-7
No translation
Translations fall into other
colour categories
Which language uses more colour?
Red
Julian Barnes
Translator (JB)
EBJB1(363):
EBJB1(364):
EBJB1(490):
EBJB1(913):
EBJB1(929):
EBJB1(936):
EBJB2(448):
EBJB2(449):
EBJB2(457):
EBJB2(461):
EBJB2(616):
EBJB2(717):
EBJB2(774):
EBJB3(169):
EBJB3(290):
EBJB3(709):
red
red
red
reddish
reddish
reddish
red
red
red
red
reddish
red
red
burgundy
encrimsoned
red
vermelhas
vermelhas
Ø
avermelhados
avermelhado
avermelhado
ruivo
ruivo
ruivos
ruiva
vermelhusca
vermelho
vermelhas
cor-de-vinho
corei
vermelha
Colouring COMPARA
45
COLOUR DENSITY (per 1,000 words)
„ English: 3.041
„ Portuguese: 3.315
COLOUR DIVERSITY (per 1,000 words)
„ English: 12.569
„ Portuguese: 7.980
without colour:original
Colouring COMPARA
46
Which language uses more...?
We presented a large repository of aligned
coloured texts available to every researcher
interested in Portuguese or English.
COLOUR:RACE (per 10,000 words)
„ English: 5.026
„ Portuguese: 4.997
COLOUR:HUMAN (per 10,000 words)
„ English: 2.055
„ Portuguese: 1.941
COLOUR:WINE (per 10,000 words)
„ English: 0.145
„ Portuguese: 0.079
COLOUR:ORIGINAL (per 10,000 words)
„ English: 2.055
„ Portuguese: 1.414
Colouring COMPARA
COMPARA is also open to the semiautomatic annotation of other semantic fields,
in collaboration with interested researchers
who volunteer to revise the final annotation.
Colouring COMPARA
47
Acknowledgment
„
This work was done in the scope of Linguateca,
contract no. 339/1.3/C/NAC, project jointly
funded by the Portuguese Government and the
European Union.
Colouring COMPARA
8