Corpus English

Question

Hi,Recently, I happened to be interested in Corpus English.I've found out some, which are, Corpus of contemporary American english(COCA), BNC, Collins, etc.I'd like to learn how to use them. How to find out Collocations is not difficult, but how to search into syntax seems to be difficult.For example, I'd like to see the result of the sentence structure, "If he helped them, they would have been ~~ ".  In Google, I would type in, "helped them, * would have been", and then I would get the result.But the corpus English makes it diffucult for me. Please some one help me.....

AlpheccaStars · Accepted Answer

You have to look under the pull down menu: help/information/contact.
Scroll down and click on "query syntax"
The syntax is very tricky, but very powerful.

for example,
[v*] - matches any verb or any verb form - am, find, kicked.....

[vvg] - matches any present participle
[sing] - matches all forms of "sing" - sing, sung, sang, singing
[j*] - matches any adjective
b?t - matches a single letter between b and t - bit, bat, but, bet...
b*t - matches any string of letters between b and t - (the above plus beat, ballot, blanket, backlight....)

You can match many parts of speech, word combinations, even synonyms and punctuation.

helped them , * would
This matches

even if he hadn't helped them , they would have gravitated towards him anyway as the best among them.

* matches one single word, not many words.
note the space before the punctuation mark - that's important
note that there are no quote marks - it matches on a word-by-word basis.

The data base is large, but not anything like Google. the more words you have in your query, the less likely it is that there will be a match.

There are no results for this:

helped them , * would have been

Pructus · Answer

Thanks so much, AlpheccaStars!!

I didn't expect some one would reply on this question.

Your explanation is so helping....

I still need to know how to use it, though.

The manual is too short and without some specific examples,

so it is difficult, specially for me, a non-native of English language.

I wish I could find some kind of manual for the Corpus.

Isn't your name a little too long?

How about, "Alphecca" or "Alpha" or "AlphaStars"?

AlpheccaStars · Answer

You can call me Stars, or Alph, no problem.

There is no manual that I have found.

This site is primarily for researchers and linguists in the English language. It has a carefully collected database of English-language sources with citations, and a query language designed specifically for English grammar and words.

It is not a general search tool like Google.

The main query is in the box called word(s)
Type in a sequence of words or markers that you are looking for:

Example:
If you type this in:

[j*] bird * [sing]
you ask for
[an adjective] bird [any word] [any form of the verb "sing"]

The results are:
yellow bird ( singing
recorded bird is singing
little bird did sing
greeny-yellow bird that sings
golden bird to sing
actual bird typically sings

If you type in -

to [r*] [vv0*]
to [any adverb] [any base verb]

you will get the occurrances of "split infinitives"
If you write one of these, some strict English teachers will smack your hand with a ruler!
Here are some results:
to best buy (This is not a spit infinitive, though it matches the pattern)
to further reduce
to boldly go
to accurately determine
to always tell
to better assess
to almost throw
to carefully inspect

etc.

What exactly do you want to use it for?

Pructus · Answer

Thanks so much for the info, Alph!!

I'd like to find out if the sentence structure is possible: "If they helped him, he would have succeeded". The structure of [If + Past, would/should/might + Past Perfect].

Some examples might be like....

1. If he did it for them, they would have been able to ~~

2. If you called for me, I might have tried to help you.

3. If he said the words to them, they should have attacked him.

Could you explain, as a native speaker, how you think on these sentence structures, and let me know what words I should type in the Corpus?

AlpheccaStars · Answer

I would probably not use the Corpus for this query. I would consult a grammar reference on conditionals: http://www.englishclub.com/grammar/verbs-conditional_4.htm

Conditionals have the following structure
If (a), then (b).

The tenses in (a) and (b) depend on the possibility that (a) is true or not.
For example, zero conditional is when (a) is true, and (b) is true.
If the nuclear reactor loses its cooling water for more than a day, there will be a meltdown and the earth will be contaminated. with readioactive material.

The situation you describe is "third conditional" when (a) is in the past, and did not actually happen. It is impossible for (a) to be true.

Pructus · Answer

Thanks a lot, Alph!!

I need to work more on the conditionals....

Even though it's before I studied the webpage you linked but .......

I happened to see a sentence like this: "If it were not for his potential disapproval, they would have made ~~~". The structure of this one is: If + Past, Would have P.P. And I'd like to know if this structure could be extended generally.

What should I type in the Corpus?

Is it, "if [nn] [vvd], [nn] would have [vvd]"? This one doesn't work. What should I type in the Corpus?

AlpheccaStars · Answer

pructus I happened to see a sentence like this: "If it were not for his potential disapproval, they would have made ~~~". The structure of this one is: If + Past, Would have P.P. And I'd like to know if this structure could be extended generally.

"If it were not for his potential disapproval, they would have made..

The sentence is not strictly grammatically correct. But since the subjunctive mood is dying out, you will see sentences like this using variant verb forms and structures. These are the formally correct forms:

"If it had not been for his potential disapproval, they would have made...

If it were not for his potential disapproval, they would make...

The English language is so variable, and the Corpus query syntax so rigid, I don't think there is a query string that would return what you are looking for.
Even if the corpus were to return representative sentences, it would not explain them, nor explain why you see discrepencies in usage.

Pructus · Answer

I see, Alph!!

Thanks again....

Corpus English

Top answer

Get the Weekly English Kit 📬

Related Questions