2

I want to find all words that occur commonly with a given word. For example the words light, sunshine, breakfast, alarm commonly occur with the word "morning". Where can I find such information? I hope this question is relevant on this site.

7
  • Btw, what's it about "Wordnet" here?
    – Kris
    Commented Jan 6, 2015 at 14:30
  • 1
    Morning and breakfast do not collocate; they just happen to be often used near each other. Are you looking for words that just happen to be used around each other, rather than actual collocations? For example, baseball glove are collocations, while baseball Pirates or baseball Ruth are not; it's just that utterances that include the words Pirates and Ruth just happen to frequently also include the word baseball. There is, to my knowledge, there is no even remotely reliable way of extracting information about the latter. Commented Jan 6, 2015 at 14:37
  • @JanusBahsJacquet "information about the latter:" This? (english.stackexchange.com/a/218843/14666) Araucaria's answer below.
    – Kris
    Commented Jan 6, 2015 at 15:11
  • @Kris That is still highly limited compared to what would be needed in the scenario I asked about. “Joey made me the best breakfast I’ve had in weeks this morning”, for instance, would not yield any ‘hits’ for morning + breakfast. Extracting statistics for this kind of ‘semantic collocation’ (for lack of a better term) would require an absolutely enormous amount of manual, multidimensional tagging. Commented Jan 6, 2015 at 15:21
  • @Janus, yes you are right, I am looking for words that just happen to be used around each other
    – Curious
    Commented Jan 7, 2015 at 6:10

2 Answers 2

6

"all words that occur commonly with a given word" – it's called "collocation."

Look up the collocations of the word you have in mind. E.g., morning in the Online OXFORD Collocation Dictionary.

adjectives:

this, tomorrow, yesterday | Friday, Saturday, etc. | early, late The side of the mountain appeared pink in the early morning light. | April, May, etc. | spring, summer, etc. | beautiful, bright, fine, sunny | cold, frosty, grey

3

You can use a concordancer. Here's a link to a concordancer.

It basically searches various language corpuses, and pulls out strings of text with the given word in. You can sort by the word to the left, or to the right. You can also choose to sort be two/three/four etc words to the left or right as well. And you can set the length of the readout from 10-1000 lines.

Different concordances will be able to search in more or less detail (i.e. pull out X + adjective, or X + noun combinations). Many will also give you the frequency of the different collocations in descending numerical order. Here is a sample from a readout for "water [word]". It's relatively short but long enough to give you an idea:

  1. from Frederica to painted signature The arrested WATER shone and danced In the
  2. ained but we had a problem with a little bit of a WATER shortage but it rained for a
  3. e balance Another urgent problem is the country s WATER shortage he added If they do
  4. his Its too dry really there s going to be a real WATER shortage still Yes a
  5. ivity of that task In New Zealand for example the WATER shortages of
  6. l A of Fig shows only the subjects pre exposed to WATER showed neophobia pre
  7. hovered in wavering singing clouds Sometimes the WATER shuddered before them as a
  8. speak but the effort cost her another mouthful of WATER Shut up and hold still a
  9. arina in Philadelphia it isn t as developed as uh WATER Side in Norfolk or the
  10. ements related to it i think well i m i m water a WATER sign and i m a fire sign
  11. d hold of him like this and then straight out the WATER Silly fool I said I don
  12. water Each pound of ice produces about gallons of WATER Since it takes calories
  13. been imprisoned here the last one without food or WATER since the guard had not
  14. a pan tends to rise to the top The colder heavier WATER sinks to the bottom and
  15. hat just sixty five in a bath Well not sixty five WATER Sixty five from the hot tap Oh
  16. ave a time share down there and take the boat and WATER ski and and go fishing
  17. haven t got to use it much but i love to swim and WATER ski and stuff like that
  18. s three years old How bout that That squirrel can WATER ski Man that s hilarious

Here we can see that the word shortage/s seems to have some special relationship with the word water. At the bottom (I've deliberately cut it off because it's too long) there are twenty-seven instances of ski or skiing and so forth. This shows a very strong relationship.

At the bottom of the readout from the concordancer that I've linked to, they also give the frequency of each item within that selection. You need to weed out the accidental stuff, and grammatical words. Here's the reading for water [word]:

and=383 in=139 the=87 to=79 gypsy=76 is=68 on=63 it=59 s=59 was=59 from=54 i=49 you=48 that=46 for=42 but=41 yeah=37 he=36 a=33 or=32 so=32 they=31 as=28 at=28 with=28 heater=27 supply=26 oh=24 she=23 then=23 of=22 out=22 when=21 which=21 if=20 all=19 pump=19 into=18 bottle=16 over=15 there=15 this=14 now=13 skiing=13 well=13 by=12 we=12 had=11 right=11 uh=11 water=11 what=11 where=11 will=11 would=11 comes=10 down=10 rafting=10 tank=10 are=9 before=9 fishing=9 like=9 no=9 off=9 only=9 up=9 aerobics=8 bill=8 colours=8 cress=8 too=8 um=8 until=8 why=8 without=8 yes=8 because=7 below=7 can=7 does=7 gets=7 line=7 look=7 sports=7 around=6 cooler=6 do=6 just=6 landing=6 mm=6 near=6 not=6 one=6 pipe=6 please=6 pollution=6 seal=6 ski=6 some=6 system=6 were=6 after=5 ah=5 between=5 coming=5 dripped=5 even=5 every=5 fish=5 has=5 here=5 his=5 leak=5 okay=5 rat=5 round=5 running=5 rushed=5 snake=5 softener=5 supplies=5 again=4 an=4 barrels=4 base=4 boils=4 bottles=4 colour=4 could=4 did=4 er=4 goes=4 heaters=4 level=4 lilies=4 looked=4 standing=4 them=4 till=4 your=4 almost=3 also=3 baby=3 bath=3 bed=3 boiled=3 came=3 chestnuts=3 color=3 come=3 contained=3 cos=3 doesn=3 dripping=3 fell=3 fountain=3 freezing=3 give=3 go=3 going=3 have=3 hot=3 ice=3 its=3 jack=3 jug=3 last=3 left=3 loss=3 mark=3 molecule=3 my=3 next=3 pipes=3 polo=3 projects=3 ran=3 rates=3 really=3 see=3 shortage=3 slowly=3 soluble=3 something=3 stops=3 systems=3 take=3 taken=3 than=3 though=3 through=3 torture=3 towards=3 two=3 under=3 while=3 within=3

You also need to be aware of which corpus you choose to select from. This one seems to have included a book about water gypsies so this collocation returns a freak number of water gypsies. But from the above list we can pull the following collocations:

heater=27 supply=26 pump=19 bottle=16 skiing=13 well=13 water=11 rafting=10 tank=10 fishing=9 aerobics=8 bill=8 colours=8 cress=8 sports=7 cooler=6 landing=6 pipe=6 please=6 pollution=6 seal=6 ski=6 system=6 coming=5 dripped=5 leak=5 rat=5 round=5 running=5 rushed=5 snake=5 softener=5 supplies=5 barrels=4 base=4 boils=4 bottles=4 colour=4 heaters=4 level=4 lilies=4 looked=4 standing=4 baby=3 bath=3 bed=3 boiled=3 came=3 chestnuts=3 color=3 dripping=3 fell=3 fountain=3 freezing=3 hot=3 ice=3 jack=3 jug=3 last=3 mark=3 molecule=3 pipes=3 polo=3 projects=3 ran=3 rates=3 shortage=3 slowly=3 soluble=3 something=3 stops=3 systems=3 torture=3

If your search includes word families then the results for colour, colours and color would appear together, so you'd see that it would actually be fourth on that list. I find it helps to sort these kinds of results alphabetically.

The linked-to concordancer won't search by part of speech. However, many do, and the Original Poster would need to do a more specific search for morning using parts of speech. This is because morning most frequently occurs as the last word in a sentence. The words following the word morning are likely to have no special relationship with it whatsoever!

Hope this is helpful.

1
  • 1
    The import and purpose of concordance is quite different from what the OP refers to.
    – Kris
    Commented Jan 6, 2015 at 14:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.