The point of it…
Suppose you're interested in identifying structures of a certain type (as opposed to a given word or phrase), for example sequences of Noun+Noun+Noun. You can type in the tags you want to concordance on (with or without any words).
How to do it…
In Concord's search-word box, type in the tags you are interested in. Or define your tags in a tag-file.
<w NN1>table finds table as a singular noun (as opposed to as a verb)
<w NN1>* <w NN1>* will find any sequence of two singular common nouns in the BNC Sampler.
Note that <w NN1>table finds table if your text is tagged with < and > symbols, or if you have specified [ and ] as tag symbols, it will find [w NN1]table.
There are some more examples under Search word or phrase.
It doesn't matter whether you are using a tag file or not, since WordSmith will identify your tags automatically. (But not by magic: of course you do need to use previously tagged text to use this function.)
In example 2, the asterisks are because in the BNC, the tags come immediately before the word they refer to: if you forgot the asterisk, Concord would assume you wanted a tag with a separator on either side.
Are you concordancing on tags?
If you are asked this and your search-word or phrase includes tags, answer "Yes" to this question. If not, your search word will get " " inserted around each < or > symbol in it, as explained under Search Word Syntax.
Tags are only case sensitive if your search-word or phrase is. Search words aren't (by default). So in example 1, you will retrieve table and Table and TABLE if used as nouns (but nothing at all if no tags are in your source texts).
After you have generated a concordance you may wish to hide the mark-up. See the View menu for this.