Current selection:

Home    

Contact us  

Advanced Search     Switch Selection     Classical books index     Preferences

[ Previous ]     [ Contents ]     [ Next ]


A "How To" Guide for the WCT Reader

How Do I Add Power to a Simple Search?

How do I search for an exact phrase?
Within a simple search, anything within double quotation marks is treated as an exact phrase. Enter the two words "general election" (with the quotation marks), and click on Search. During the processing, the quotation marks are taken into account to ensure that the two words are right next to each other (no intervening words) and they are in the order given ... general, then election. The result is in Screen 1:

The exact phrase "general election" occurs 12 times in this collection from G.K. Chesterton.
How do I search for an inexact phrase?
Any well designed search engine can find exact phrases. For example, the phrase "great statesman" is present exactly once in the Collection from G.K. Chesterton. Enter the two words between quotation marks, and click on Search. The result is in Screen 2:

However, this "exact phrase" search fails to get at important variations of the phrase. To get an inexact phrase, simply leave out the quotation marks! In other words, do a simple search of the words in the phrase. The result is as in Screen 3:

It turns out that the idea of "great statesman" appears four times, not once. As expected with Words Close Together relevance ranking, the exact phrase shows up at the top of the list. Next come three inexact phrases -- great Unionist statesman, great social statesman, and great and gray-haired statesman. Notice the meaning scores: 100 for zero intervening words, 99 for one intervening word, 98 for two intervening words. The fourth hit is the last inexact phrase; the hits after that are not phrase variations of great statesman.

So there were four significant hits, not just the one that other search engines would find.
Can I mix phrases and words in a simple search?
Yes. Simply put the phrase (or each phrase if there are more than one) within quotation marks. For example, if you wish to learn of Chesterton's view of truth related to the phrase "public schools", input the word and the phrase in quotes (word first or phrase first; it doesn't matter). Screen 4 shows part of the first results page:

Note that the count for the phrase "public schools" alone is 26. The combination of word and phrase focuses the result on a particular aspect in Chesterton's view of public schools.
Is there an easy way to search for variations of a word?
Yes, yes, yes. Wild cards have been a longstanding tradition in computerized search. One wild card is the asterisk, which stands in for any number of characters from 0 upward. The other is the question mark which stands in for exactly one character. Once you are used to wild cards, you will have a hard time not using them; they make a search so much more powerful.

A series of examples is offered to make the use of wild cards clear.

Suppose you want to get at words related to democracy. How many letters are needed to limit the words? Find out by experimenting with a single word search. "Demo*" -- four letters and wild card -- fails; it calls up words such as demon and demonstrate. Try five letters and the asterisk wild card, "democ*". That does it; see the shorter list in Screen 5:

Turn now to a two word search, both with wild cards -- aristocra* democra*. Sure enough, Screen 6 shows variants of democracy and aristocracy found near one another:

With 40 search results, comparison of democracy and aristocracy is obviously a theme that is important to Chesterton.

Chesterton's writing spans the prelude and aftermath of World War I. Let's call up records that get at his views re Germany and Britain. Input German* Brit* to get Screen 7. (In the page on Advanced Search, we will find ways to increase the power of search for this theme even further.)

The question mark can be used to replace a single letter, as in deal? which calls up present tense deals and past tense dealt as in Screen 8:

There are limitations to wild cards. Obviously, simply searching on an asterisk would be an invitation to search on everything. To keep the WCT Reader from spiraling off into hyper space or having a hernia, we set a limit of not more than 300 different words being called up by any search term that includes one or more wild cards. A further limit: a combination of three letters or digits must be present within any term using wild cards. Searching on *ing in the Chesterton collection exceeds the 300 word limit. Searching for *ging is not particularly useful, but it actually works, as in Screen 9:

Wild cards turn out to be especially useful when we want this word OR that word OR another word. More of that in the page on Advanced Search.
How role do headings play in search results?
In the standard Internet search engines, words in headings are generally treated the same as other words. Pity!

Headings help to get at meaning. Here in Screen 10 is a segment of the table of contents within the Chesterton collection:

With the preference set to a maximum of 300 words in each screen layout, Chapter III is displayed as 36 successive screens. Eleven of these screens (apart from the heading) never mention any variation of poem, poems, poet, poetess, poets, or poetry. Example paragraph:
When we come to ethics and philosophy, doubtless we find Swinburne in full revolt, not only against the temperate idealism of Tennyson, but against the genuine piety and moral enthusiasm of people like Mrs. Browning. But here again Swinburne is very English, nay, he is very Victorian, for his revolt is illogical. For the purposes of intelligent insurrection against priests and kings, Swinburne ought to have described the natural life of man, free and beautiful, and proved from this both the noxiousness and the needlessness of such chains. Unfortunately Swinburne rebelled against Nature first and then tried to rebel against religion for doing exactly the same thing that he had done. His songs of joy are not really immoral; but his songs of sorrow are. But when he merely hurls at the priest the assertion that flesh is grass and life is sorrow, he really lays himself open to the restrained answer, "So I have ventured, on various occasions, to remark." When he went forth, as it were, as the champion of pagan change and pleasure, he heard uplifted the grand choruses of his own Atalanta, in his rear, refusing hope.
Is the subject of this paragraph poetry? Surely, yes. Atalanta is a poetic work by Algernon Charles Swinburne. Tennyson and Browning were poets. The discussion is about Swinburne's poetic style. Other search engines searching for discussion of poetry would miss this sample paragraph. Words Close Together picks it up. So too the other ten screens in this chapter will be found in a search for variations of poets -- precisely because the word is present in the heading. That's the power of recognizing headings as headings.

How close together are words in a heading to words in the paragraphs below? Right together? Too far away? That could be argued at length. The solution in the WCT Reader is to place all results that depend on words in the headings at the end of the results list. So whatever is the lowest possible meaning score in a search is assigned to heading-dependent results. Example: Suppose you are searching with the preference set to a ceiling of 25 intervening words. The lowest score will be 75 (100 less 25). All results that depend on one or more words in a heading are assigned meaning score equal 75.

What if all the requested search terms are in one heading? Then the meaning score is 100 less the count of intervening words within that heading. That score is assigned to every screen layout of text under that heading.

Complicated? A bit, especially under the surface. More meaningful results? Yes, definitely yes.
How do I change the number of results per page?
Here is another choice that you may set in the Preferences page:
When results of a search are displayed, how many "hits" would you like to see on each page? The current preference is 10; the standard number is 10. Set your new preference (between 10 and 100):
This is useful if you wish a quick overview of many more results in a hurry. Note that the result page may be quite large if you set the number all the way up to 100.
Can I change the order of results?
Suppose results are coming out in order of reducing meaning scores, and I would prefer to see the results in the order that they appear within the ebook. Can I do anything about it? For example, there are times when the results of a search of a subject in the Bible might be more logically presented in the biblical order. This can be done on the Preferences page. Click Home, then Preferences, and scroll down to Screen 11:

In order that you do not forget that normal relevance order does not apply, a reminder is shown with each search result as in Screen 12:

"12 records were found, presented in bookshelf order" -- The last two words are a link that take you to the Preferences page, so that you can change back to the normal relevance ranking when you are ready.

[ Previous ]     [ Contents ]     [ Next ]


 
Meaningful  precision  search  for text data  is available now.  Learn more.
 
words close together.com The "Research Quality" Search Engine by Marpex, Inc.