Kernerman Dictionary News • Number 10 • July 2002



Analysis of sexy in the 450-million-word Bank of English corpus

By Ramesh Krishnamurthy

It may interest readers to know more precisely how corpus data can be used in linguistic analyses. Here are some of the features of the word sexy, as revealed by some typical corpus outputs.

1. Frequency

1.1 There are 8698 examples of sexy in the corpus. This gives us a rough idea of how common or rare the word is, and can help us to decide whether to include it in a dictionary or in our vocabulary teaching, how much space or time to devote to it, etc.

1.2 Compare this with the frequencies for some of the most common adjectives: new 676403; first 580674; good 349425; old 302953; little 229915; best 209462; big 175660; black 129727.

1.3 Look at the adjectives with a similar frequency to sexy: stupid 9951; generous 9852; naked 9849; clever 9836; accurate 9416; visible 9323; incredible 9169; silly 8856; strict 8104; mature 7901; bare 7884; eager 7841; nasty 7840.

1.4 Some people may be surprised at how commonly sexy is used, but it has increased in usage quite substantially in recent years: it was ranked at 5522 in the 1996 (323 million word) corpus, and is now ranked 4743, a rise of 779 places in 6 years.

2. Distribution by subcorpus

The corpus is held as several subcorpora, each of which has some degree of homogeneity, for example regional variety, text-type, etc.

2.2 "Ephemera" covers junk mail, advertising leaflets, tourist brochures, etc.

Bank of English: 01.2002 SUBCORPORA % No of Words No of Texts Dates
American Books 7.23% 32.4m 327 1990 >
NPR American Radio 4.96% 22.2m 726 1990-93
BBC World Service Radio 4.15% 18.6m 143 1990-91
British Books 9.67% 43.4m 578 1990 >
British Ephemera 1.03% 4.6m 2359 1991-6
British Magazines 9.84% 44.2m 1113 1992-00
British Spoken 4.48% 20.1m 2670 1991-96
Economist 3.50% 15.7m 229 1991-99
Independent 6.26% 28.1m 240 1995-99
Times 11.57% 51.9m 322 1999-01
Guardian 7.20% 32.3m 332 1995-99
New Scientist 1.76% 7.9m 188 1992-99
Australian Newspapers 7.79% 34.9m 406 1995-99
American Ephemera 0.78% 3.5m 2786 1995-96
American Newspapers 2.23% 10m 281 1994-96
Sun and News of the World 9.98% 44.8m 597 1997-01
American Academic Textbooks 1.41% 6.3m 31 1990-96
American Spoken 0.45% 2m 16 1994-97
Strathy Canadian Corpus 3.55% 15.9m 234 1980 >
Wolverhampton Business Corpus 2.15% 9.7m 23 1999-00
TOTAL 448.5m 14,813

2.3 Here is the distribution by subcorpus for sexy. This indicates which writers and speakers are using the word, and in what type of texts.

Corpus Total Number of Occurence Average Number per Million Words
sunnow 3037 67.9/million
brmags 2324 52.6/million
oznews 652 18.7/million
indy 452 16.1/million
times 811 15.6/million
guard 474 14.7/million
usephem 51 14.5/million
usbooks 259 8.0/million
brbooks 274 6.3/million
brephem 26 5.6/million
brspok 92 4.6/million
newsci 28 3.5/million
strathy 56 3.5/million
usnews 33 3.3/million
econ 43 2.7/million
npr 59 2.7/million
usacad 14 2.2/million
usspok 1 0.5/million
bbc 8 0.4/million
wbe 4 0.4/million

2.4 The examples of sexy are most frequent in UK lower-register journalism (Sun and News of the World; British Magazines), then in Australian and UK higher-register journalism (Courier Mail, Brisbane; Independent, Times, Guardian). In books and ephemera, they are slightly more frequent in US than in UK. They are generally more infrequent in spoken data than in written data; and very infrequent in academic and business texts.

3. Collocation

Collocation is the tendency for some words to co-occur in close proximity to each other, either loosely (e.g. The work demanded of him was initially quite hard) or in more-or-less fixed units (e.g. It was hard work). Various statistics (T-score, MI-score, log likelihood, etc) are used to measure significant co-occurrence (because the very frequent words like the, of, and, to will obviously co-occur with all the other words).
The standard corpus output shows the top 50; but only the top 10 collocates of sexy are given here:

Top collocates by T-score co-occurrences t-score
and 2551 17.920569
very 404 16.729363
she 491 14.856371
look 238 13.508434
s 1093 12.842582
feel 203 12.821835
a 2094 12.108906
her 377 11.718177
is 1026 11.654412
woman 152 10.893944

3.2 With adjectives, and is often a significant collocate, because we tend to use adjectives in pairs (sexy and attractive, sexy and funny; beautiful and sexy, feminine and sexy); very indicates that sexy is a gradable adjective. Just from the top 10 collocates, we can see that sexy is strongly associated with females rather than males: she, her, woman (the top 50 also includes women, girl, girls, female, but no male gender-specific nouns or pronouns). We also note that sexy is more strongly associated with look (i.e. appearance) rather than feel (i.e. emotion) (confirmed in the top 50, where looking and looks are more significant than feeling).

3.3 The top 50 collocates include some gender-neutral nouns (star, singer, blonde, supermodel), many of which actually refer to women when the contexts are examined.
Some of the collocates are other adjectives: black, young, blonde, slim. Several are nouns referring to clothes (also associated with women): underwear, dress, lingerie, dresses.

3.4 The significance of the strong female gender-association of sexy is highlighted if we look at the most common gender-specific words in the corpus: male terms are usually more frequent, e.g. he 2729706, she 1026560; his 1855728, her 951899, hers 4443; Mr 517655, Mrs 75254, Ms 35211, Miss 27876; man 273109, woman 114686 (the main exception is women 204605, men 181933)

3.5 The corpus software can also display collocates in their exact positions in relation
to the word being examined (called the NODE), e.g. three words before, two words before, one word before, one word after, two words after, three words after, etc (up to a maximum of six words away). At Cobuild this is called the collocate "picture" (as opposed to the "list" shown above). It is slightly more difficult to interpret, as you have to look down the columns, not read horizontally as in normal reading.

Collocate "picture" by T-score
think is a NODE and and she
she </dt> very NODE but star her
2001 find and NODE new but i
makes 1999 s NODE underwear i t
is s feel NODE football she black
make feel be NODE singer with you
s with so NODE women girl woman
1999 m as NODE black look has
i are too NODE lingerie <p> deidre
</dt> she her NODE as fun star

3.6 Note that all words are lower-cased in analytical outputs; some corpus codes can also be seen here: </dt> indicates the end of the date of a newspaper article, <p> indicates the end of a paragraph in any text.

3.7 This "picture" shows that the most significant adjacent combinations are: a sexy, very sexy, and sexy, etc and sexy and, sexy but, sexy new, etc. The most significant nouns modified by sexy are: underwear, football (see 5.2.3 below), singer, women, lingerie.

4. Concordance Lines

The corpus software can show us all the examples in the corpus texts for any word, with a screen-width of context for each example (but this can be increased by the user), and additional information if required (e.g. the grammatical wordclass and the source text). The word being examined is at the centre of the screen (this is called a KWIC - Key Word In Context - concordance). Here are 20 concordance lines for sexy, selected at random by the software, and sorted alphabetically by the immediately following word (note that codes and punctuation come first; then and, change, choose, enough, etc down to until):

<hd> WHEN YOUR BODY SAYS SEXY </hd> WHAT is your body saying to
let's put it on together and make it sexy # <p> He insists: `They cost too
I find his peace plans fantastically sexy. <p> 3 The Pope <p> Censored-Ed.
and Javier Bardem are irresistibly sexy and stylish, when they actually
Heavy kohl and parted glossy lips-sexy and just a little kitsch-can
of what is considered beautiful and sexy change, too. Therefore, to come
far left. <p> Or for something more sexy choose the crochet-strapped pink
with that. Is my style of music sexy enough? I think it is. As far as
all my hair gone # Kathy Burke is too sexy in her skirt Kathy Burke, with
fired by the sinister Cigar Girl, sexy Italian actress Maria Grazia
accept her responsibilities. But now sexy Julie-Ann has formed a friendship
a rented mansion in California. <p> Sexy Latin singer Jennifer, known as
ditch big T-shirts in favour of sexy little combos that look at home
the occasion. Dave dressed up as a `sexy nurse", and Robbie wore a French
not that. They are just there to be sexy or set dressing or the victim.
comfortable as summer gets soft and sexy. Photographs by Adrian Wilson.
who don't have college degrees, sexy professions or fat paychecks.
a Mintel survey says. <hl> Caprice sexy scenes too hot for telly </hl>
DOLL. The bizarre gift was left with sexy underwear at the Cancer Research
It used to be that women were sexy until they had a baby and that

4.2 Concordance lines can also be sorted by the word occurring two words after the NODE, one word before the NODE, etc.

4.3 By closely examining the contexts, we can gradually build up an analysis of the usages of sexy.

5. Usage analysis of sexy

The most frequent use of sexy is to describe people, usually women, as attractive and arousing sexual feelings. The nouns directly modified by sexy include woman, women, girl, girls, star, singer, blonde, actress, lady, model.

5.1.1 Sexy also describes clothes, usually women's clothes: underwear, dress, lingerie, clothes, dresses, undies, outfit, costume, swimwear.

5.1.2 Some of the following nouns are gender-neutral or abstract, but still clearly relate to sexually attractive appearance or behaviour: thing, image, stuff, look, beast, mood, voice, fun, curves, antics, pout

5.1.3 Sexy is also used to describe books, films, advertisements, music, dances, etc that are sexually explicit or suggestive: film, photos, ads, campaign, play, thriller;
e.g. Chinese censors rap Swedish rock'n'roll for being too sexy (Guardian 21/2/95).

5.1.4 Some objects have traditional, indirect associations with sexual attractiveness, especially cars and motorbikes:
e.g. PEUGEOT'S sexy 206 coupe cabriolet is now available (Sun 3/8/01);
and more recently mobile phones:
e.g. Anxious not to alienate men, Nokia denies its sleek and sexy 8810 (from 350) was designed with women in mind (Guardian 11/3/99);
and, despite recent anti-smoking campaigns, cigarettes:
e.g. Cigarettes are sexy, companionable, and make for great romantic gestures in cinema (Economist, 15/7/95).

5.1.5 The 1960s and 1970s are often described as sexy decades:
e.g. Astonishingly beautiful, she was a social butterfly who epitomised the irreverent, promiscuous days of the sexy 70s (Antique Collector, Sept 1992; British Magazines).

5.2 More recently, sexy seems to have lost its connection with sex almost completely, and has taken on a much more general meaning like "exciting" or "fashionable". The earliest examples are from US texts, but they now occur widely in texts from the UK and most other English-speaking countries.

5.2.1 This usage may have originated in the area of product design, as an extension of women's fashion design (mentioned above), and is perhaps also connected with the 1960s and 1970s:
e.g. It was a German gun … and it lacked the sexy compactness of the Israeli Uzi (Tom Clancy, Clear and Present Danger, 1989; US books);
And let's face it: Sofas and end-tables aren't as sexy as state-of-the-art computer equipment (US Newspapers, 11/3/96);
It seemed that everything was re-designed in the Sixties … Everything was sexy. The jumbo jet was sexy, being a stewardess was sexy, everything was eroticised. And the design of everything reflected that. You didn't have kettles, you had `sexy kettles" (Times, 3/7/99);
… the sexy brushed-aluminium Wine Tube which has been launched by a Danish manufacturer (Times, 22/9/01);
What's sexy about a thing like a chair if it is not comfortable? (Australian newspapers, 25/10/98).

5.2.2 The media and advertising industry obviously seized on this usage very quickly:
Commerce is sexy and creative careers have been transformed (Times, 8/1/00);
And if female creatives do stick it out, they may be stereotyped and have to fight to land the `sexy" briefs such as cars and beer (Independent, 5/6/99);

5.2.3 The term sexy football was coined by the Dutch football star and manager Ruud Gullit, and became very topical in 2001:
e.g. `First there was the tradition, and that is everywhere," he said. `Then the attacking football, sexy football as Ruud Gullit called it, and you have already seen that (Times, 27/10/01).
Here are some of the concordance lines for sexy football, showing the text sources:
times/UK nowadays." When it came to sexy football, Johnstone was a casanova
oznews/OZ ouldn't say we were playing sexy football at the moment ... more
sunnow/UK boss: Ruud wanted to bring sexy football" to the Geordie faithful
guard/UK s while constantly chanting `sexy football" - presumably in homage
indy/UK ith whom he hopes to make his sexy football on Tyneside not because

5.2.4 Cities are now classified as sexy:
e.g. New York, on the other hand, is fast and furious, affluent and sexy (Guardian, 16/1/99);
Edinburgh was the only Scots city to be included in Cosmo's list of sexy cities in the UK (Sun, 17/7/01).

5.2.5 Sexy can now be used to describe almost anything,
from food (e.g. As far as salads go, it's pretty damn sexy (Times, 8/9/01); Who said baked spuds could never be sexy? (Independent, 9/4/99); For a sexy alternative just add a few ripe strawberries or other soft fruits (Sun, 27/10/99)),
to sport (e.g. Yes, cricket -once it was sexist, now it's sexy (Sun, 4/7/01); … teams like the `49ers and the Steelers, the big glamorous sexy clubs (NPR, 27/11/91)),
computers and technology (e.g. A word processor isn't very sexy (New Scientist, 12/12/98); IT isn't as sexy as personnel management (British Magazines); Flight simulators and virtual reality are enticingly sexy (Economist, 5/3/94); The internet has made home delivery sexy again (Times, 23/1/00)),
politics (e.g. `I think the party is lacking in lustre, but I want to make politics sexy' (Times, 19/9/99); He has lovely hair and I find his peace plans fantastically sexy (The Oldie, 25/6/93; British Magazines); In America, cartel busting is sexy (Times, 24/2/00); If there were something sexy about entitlements, now we'd have a story (NPR, 1/4/92)),
science (e.g. Richard Dawkins, do people find you intimidating? Or have you made science sexy? (Independent, 23/12/98); James has reportedly described Beagle's mission to search for life as `sexy" (Independent, 3/12/98)),
the arts (e.g. The new conservative [authors] understand the `need to be provocative, push buttons, be sexy." (British Magazines); I don't find theatre sexy (Talk Radio UK, 12/2/96); Danny Boyle - who made British films sexy again with Trainspotting (Sun, 12/10/99); It seems that, in the world of literature at least, everything Indian is sexy (Australian newspapers, 26/6/99)),
success (e.g. Success is sexy (New Scientist, 3/4/93)),
living in cities (e.g. Living in cities has become sexy again, and it's being reflected in higher prices (Guardian, 17/4/99)),
careers (e.g. Engineering is no longer regarded as `a sexy career choice" among school leavers (Times, 8/9/01)),
the name Colin, Britishness, political incorrectness, homelessness, the First World War, pensions, price stability, trains, and even podiatry!

K Dictionaries Ltd
10 Nahum Street, Tel Aviv 63503 Israel
tel: 972-3-5468102 • fax: 972-3-5468103