Richard-Business Eng
Professional Teacher
Looking for some help from a knowledgeable n-gram user...
I wanted to compare the frequency of use of "if and the conditional should".
I did not want n-gram to include the usage of the modal should.

Is there a way to specify a particular use, meaning, or part of speech when entering a word in n-gram?
Jun 21, 2020 11:58 AM
Comments · 12
Yes there is and it seems you put a recognised operator at the end, to specify the part of grammar you are interested in.. I need about a day to familiarise myself with the literature. I do not use N-gram myself.

this is an experiment with a result that should never occur which it didn't

I think this shows the word verb with if or should but there is a way to specify the usage.

I believe you have to play around with these tags and the wildcard tag *
if * = if plus common.

the tag list:

_NOUN_These tags can either stand alone (_PRON_)
or can be appended to a word (she_PRON)
_DET_determiner or article
_ADP_an adposition: either a preposition or a postposition
_ROOT_root of the parse treeThese tags must stand alone (e.g., _START_)_START_start of a sentence_END_end of a sentence

[this means type in _uppercase recognised tag stand alone or not, like for computer programming using super globals]
you have to type _VERB whatever you are interested in, with the underline at the front _DET etc

And in any case you will get views now there is one view up here for others to follow, that's what humans do.

This is assuming google n-gram and not Stanford university or other similar n-gram tools.
June 21, 2020

The conditional "should" is the one that you use in place of "if": "Should it rain tomorrow, the match will be cancelled."
June 21, 2020
My first thought is to think about the surrounding words. “Should” as a modal verb is followed by a verb; “should” as a conditional is followed by a noun, pronoun, or determiner. It’s impossible to do a search for “should” with all nouns, but you can look up instances of its usage with each pronoun and determiner: “should I”, “should my”, “should the”, etc.

To create comparable results for “if”, you can add the same controls: “if I”, “if my”, “If the”, etc.

The total results will exclude the incidences of “should" and “if” followed by a plural, uncountable noun, or proper noun. But it should be a good enough representative for comparison.

EDIT: I just realized that “should” as a modal verb can be followed a noun/pronoun/determiner in a question, so that would affect the results.
June 21, 2020
Unfortunately, the n-gram tool is just a word search with no knowledge of grammar.

June 21, 2020
That would be possible but you may have to run several searches and the documentation/literature/the handbook-manual available on the internet unfortunately like all computer algorithm language is not clear.
It would at first glance involve the use of forcing the algorithm to work how you want it by (putting in parentheses) and adding a limiter to the query.

I suggest you use Stanford universities N-gram or other university/ language institute/ AI language learning N-gram tools. Google only scans the books it has scanned into a database for use in its google books service. Therefore google's public N-Gram it has limited corpus compared with the experts in this field.

It will be should? be possible but the expression you need to type in will be complicated and not natural English it will be like something from a computer program. To get your result in one search, I am not familiar with that type of searching and usage of language databases.
June 21, 2020
Show More
Richard-Business Eng
Language Skills
English, French
Learning Language