Concordance for advanced users

[@priyamathew5608]

Thank you for this video. When I count the frequency of KWICs, I also want to find the range i.e. how many texts within the same corpus contain this particular KWIC word form.

[@SketchEngine]

The frequency tool does not show document frequency. You would need to show the concordance for each line and use Filter - Adavanced - 1st hit in document.

[@elizabethcarson5624]

I am not sure how I did this before so I can't recreate the search. I am trying to search for 2 specific lemmas close to each other. Before, I was able to type in child+woman and got results for any lines containing child and woman close to/next to each other. I do not know how to recreate a search like this.

[@SketchEngine]

Sketch Engine cannot do the search as you tried. The syntax is different. Go to CONCORDANCE - ADVANCED - Simple and type "child" and use - Filter Context - Lemma context for "woman". Alternatively, you can use the CQL command CONCORDANCE - ADVANCED - CQL - CQL BUILDER - "meet" . It is described here: www.sketchengine.eu/documentation/cql-meet-union/ Contact our support for more details [e-mail]

[@juliamakarova6794]

Hi! I'm writing a term paper on the use of double negation in English. I am using the British National Corpus for that. I am struggling with creating a formula though: I basically need a formula for double negation. I tried [word="not|no|ain't"] [word="no|not|nobody|nothing|nowhere"] but I am just getting sentences containing one of all those words. However, I need a combination of one word from the first part AND one word from the second part. I would really appreciate it if you had any helpful tip for me. Thank you!

[@SketchEngine]

Can you please generate the result, even if it is not what you want, and on the result screen, click the Request help icon in the top right of the interface (second icon from the right in the top right-hand corner) and include your message above? This way our support will see exactly what you are doing and which criteria you are using and will be able to help.

[@lovebts774]

Thank you for your valuable video. Could you please show me how to find all the sentences including "make somebody/someone + Verb". I know that we can use CQL, but I don't know the formula. Please show me, Thank you

[@SketchEngine]

Hi, you can use the following query: [lemma="make"] [word="someone|somebody"] [tag="V.*"]

[@lovebts774]

@@SketchEngine thank you. This tool is very useful.

[@SketchEngine]

@@lovebts774 You're welcome! Feel free to contact us at our support email [e-mail] if you need any further assistance.

[@voronasira9460]

Is there any way to filter by document properties? for example, only return results whose doc.date is between such and such dates.

[@SketchEngine]

Yes, it is possible search, filter and generate statistics by any metadata included in the corpus. Metadata are called text types in Sketch Engine. Look for the text type selector, see: www.sketchengine.eu/my_keywords/text-type/ To generate statistics by text types, this video will be helpful: kzbin.info/www/bejne/fYGYoWuOetuqnK8

[@monicalai9566]

Can I use CQL to create a subcorpus with a list of multiple words?

[@SketchEngine]

Yes. Simply construct a CQL query that searches for the words in your list and then use the "subcorpus from a concordance" method (Option2) as described here: www.sketchengine.eu/guide/create-a-subcorpus/ Contact [e-mail] if you need more guidance.

[@Anna_Floyd]

Hello, I am working on Collocations and I have built my own corpus, but can you tell me please, is it possible to extract all collocations in my corpus as a list of collocations, which are used there? Or must I search for it in Wordsketch with a single lemma? Thanks in advance.

[@SketchEngine]

Hi, next to the Advanced tab there is a As a list tab, where you can set criteria relevant to your needs and then it you will get a list of the collocations. It could take a longer time with a larger corpus.

[@irinaalvitez8457]

hi, can I look for several non consecutive words at the same time?

[@SketchEngine]

Hi, can you describe in more detail what you are trying to find? What do you mean by non consecutive words?

[@irinaalvitez8457]

@@SketchEngine hi, yes, I'm looking for different words to make subcorpus. So at the moment, I can only look for one words at a time, which means I need to create a subcorpus per word. I hope it makes sense. thank you

[@SketchEngine]

@@irinaalvitez8457 Hi, unfortunately it is not possible to create subcorpora in bulk if you have a list of words. The options for creating subcorpora are stated here: www.sketchengine.eu/guide/create-a-subcorpus/

[@Yujin847]

Hi, I watched your tutorials but couldn't find a way to search for synonyms using CQL. For instance, I would like to search the pattern of 'very and synonyms of very' occurring near 'probably'. What can I do to accomplish this? I would greatly appreciate your help, thanks.

[@SketchEngine]

You can find it in our user guide ('Searching for similar words with CQL': bit.ly/3qLMRLd). In your case, for example a CQL query [lemma="probably"] ~10"very-a" will find a lemma 'probably' followed by 'very' or other words from top 10 thesaurus results for 'very'.

[@Yujin847]

@@SketchEngine Thank you very much, but I'm afraid I need to ask for your help once more. Is there any way I can tag the auxiliary "do" in my CQL search? Also, I wonder if it's possible to distinguish between auxiliary be/have/do and their main verb counterparts using CQL. Looks like there are no separate tags for them on the web page. I would greatly appreciate your help, thank you in advance.

[@SketchEngine]

The question is too specific for us to help without more details such as the corpus you use. Please contact us via the 'Request help or support' in the interface. You will find it in the top right corner of every screen.

[@SketchEngine]

@@Yujin847 Please contact us via the 'Request help or support' in the interface. You will find it in the top right corner of every screen.

[@teotao5422]

How can I set 2 words to be found within the same sentence, but not necessary together? So that they can occupy random positions and be divided by other words in the sentence

[@SketchEngine]

Hi Slava, this is indeed possible but requires the use of CQL. You can set various advanced conditions such as the distance between words or that the words should be inside the same sentence. We now have CQL Builder in the interface that helps you construct the queries. Please refer to the CQL videos to learn how to do that: kzbin.info/aero/PLPVgp0Xm6hWBLZzD7haCVbhC9pSfWZk1S or use our web guide: www.sketchengine.eu/documentation/corpus-querying/

[@ellendeschacht7743]

How can I find a word that contains for example a minimum of 5 characters and a maximum of 7?

[@SketchEngine]

You need to use the word option on the Advanced tab and then use a regular expression that defines your criteria. Concordance - ADVANCED - word and use this regular expression .{5,7} (notice the dot before the first bracket) You can use the same procedure in the Wordlist tool which might be more practical for your task.

[@abdullahamouna2912]

Hi, I am conducting a study about the functions of "you" in Ted Talks. How can I specify my research to get the results for you as either subject and object pronoun? How can I search for NS and NNS in Ted Talks?

[@ondrejmatuska8945]

You need to use the CQL and specify criteria by using tags and position of tokens, e.g. "you" used before a verb which is not an infinitive and "you" after a verb. you may have to experiment. Please refer to our CQL video to learn to do it. CQL on KZbin: ske.li/cqlvideo CQL web guide: ske.li/fiz

[@soillazze1998]

hHi! Is it possible to filter per year?

[@SketchEngine]

Yes. If you are working with a corpus which contains the metadata - year. These are for example "Timestapmed corpus" or "English Trends" corpus. More information here: www.sketchengine.eu/my_keywords/text-type/

[@soillazze1998]

@@SketchEngine Is there any Italian corpus which allowd to search and filter per year?

[@SketchEngine]

@@soillazze1998 Yes. The Timestamped corpus. Just type Timestamped Italian in the corpus selector.