|
Learning
and Teaching through Context - A Data-driven Approach
by Ramesh Krishnamurthy
- 2
Frequency
The frequency of a word, phrase, or any other linguistic feature
is the number of times that it occurs in the corpus. Teachers
and students often find it difficult to decide whether a word
or phrase is important or not. Frequency offers a simple guide.
For example, here are the corpus frequencies for some words,
in alphabetical order:
| accident |
22,231 |
| accidental |
2,476 |
| accidentalism |
2 |
| accidentalist |
4 |
| accidentalists |
1 |
| accidentally |
2,698 |
| accidentalness
|
1 |
| accidentals |
8 |
| accidents |
6,012 |
It
is obvious that accident and accidents, accidental
and accidentally are important, because they are very
frequently used. But accidentalism, accidentalist,
accidentalists, accidentalness, and accidentals
are not important, because they are very rare.
Frequency
can also help us with multi-word units (such as compounds
and phrases), grammatical patterns, or any other linguistic
unit. For example, here are the most frequent two-word units
with accident as the first member, in frequency order:
| accident-prone |
312
(75 written without a hyphen) |
| accident
insurance |
202 |
| accident
investigation |
181 |
| accident
victims |
179 |
| accident
inquiry |
126 |
| accident
investigators |
87 |
| accident
victim |
79 |
Concordance
Corpus software usually presents the examples for a word or
phrase in what is called a concordance, in KWIC
(Key Word In Context) format, with the selected word or phrase
in the middle of the screen.
Concordances are rich sources of information and stimulation
for students and teachers. 20 corpus examples for accident,
randomly selected by the computer, are shown below. Remember
that there are 22,231 examples for accident in this corpus.
So if any linguistic feature in the examples is of interest
to the user, it can be investigated in greater detail very
easily, by getting more examples containing that feature.
Each line below is a separate example for accident,
coming from the texts in the corpus. In this case, the length
of each example is just the width of the computer screen,
but corpus software can usually provide longer contexts, full
sentences or even paragraphs, if required.
|
1
|
those
suffered by boxers or road
|
|
accident |
|
victims
to the start of t
|
|
2
|
e
suffered a horrendous motorcycle
|
|
accident |
|
,
needing pins in his dam
|
|
3
|
shaken and distressed" by the
|
|
accident |
|
,
said a police spokesman.
|
|
4
|
a
drugged patient in a hospital's
|
|
accident |
|
and
emergency unit. The 3
|
|
5
|
I
put in for Jasper after his road
|
|
accident |
|
.
. . thank you once agai
|
|
6
|
marriage
was killed in a car
|
|
accident |
|
in NSW last year but Sull
|
|
7
|
s
womb, I have been a clumsy,
|
|
accident |
|
-prone
person and I can't
|
|
8
|
actress
who has suffered a fatal
|
|
accident |
|
.
-- Des Partridge
|
|
9
|
that
Chernobyl was just a minor
|
|
accident |
|
.
Approximately 135,000 pe
|
|
10
|
opulation.
Tragically, it took the
|
|
accident |
|
at
the Chernobyl Atomic
|
|
11
|
<hl>
Crash 2 vow to fly again;
|
|
Accident |
|
;
Air </hl> <dt> 01 May 19
|
|
12
|
up
by a lorry wheel in a freak
|
|
accident |
|
killed
disc jockey Kemi O
|
|
13
|
predicted
his own sister's fatal
|
|
accident |
|
.
HER bedroom was freezing
|
|
14
|
which
could be created by
|
|
accident |
|
or by a small group of pe
|
|
15
|
strength
at the moment, due to an
|
|
accident |
|
."
<p> And have you any ne
|
|
16
|
the
plane's engine had failed. The
|
|
accident |
|
shocked
the mission commu
|
|
17
|
ho take the main burden of nuclear
|
|
accident |
|
risk.
Nuclear Electric n
|
|
18
|
to
aid the woman, despite the
|
|
accident |
|
occurring
within one mile
|
|
19
|
thought
to be among them. <p> The
|
|
accident |
|
came
a day after 75 peopl
|
|
20
|
get
it wrong again </h> <p> THE
|
|
accident |
|
prone
National Rivers Aut
|
Note:
The items in angle brackets are corpus codes: <hl> means
"beginning of a newspaper headline", </hl>
means "end of a newspaper headline", <dt>
means "date of publication of a news item", and
<p> means "end of paragraph". Most public
corpora have some codes of this kind, which the user will
soon get used to.
To
page 3 of 3
Back
to the articles index
|
