Natural Language

The most frequent mistake searchers make is to search on only one or two words. Search engines are fairly good at bringing up the major sites on broad topics such as Canadian law, or French cuisine. But to find something specific, you need more words. Often entering our query with several words and using natural phrasing will get better results.

Natural language queries are queries in plain English. Are you looking for resorts in Bermuda? Enter the question naturally budget resorts in Bermuda. Typically, the search engine will ignore the common words - in; possibly look for singular and plural and for word variants; sometimes make sense of the query - most search engines would recognize this as travel related; and then rank the results according to the relevancy formula.

Because search engines will rank results according to the order and proximity of words, it makes sense to enter them in a natural order. The query most poisonous snake gets slightly better results than snake poisonous. The answer format will help too, such as longest river in the world is.

The search engines vary in their capabilities, but all will respond reasonably well to more words.

Natural Language Exercise

Go to Google and ask it a couple of questions. There will likely be some very good leads in the first 10 results.

1. first woman to climb Mount Everest
2. budget accommodation in Paris
3. convert gallons to cubic meters
4. grey owl's real name

Hakia (www.hakia.com) is a "meaning based" search engine intent on finding results that match the meaning or intent of your question, and not merely the exact words. Its technology analyzes sentences to identify the concepts and connect them with words or similiar concepts. You can use keywords for the query but natural phrasing will make your meaning clearer.

Try it on the four factual questions, and these fuzzier questions.

1. introducing collaboration to the workplace
2. trends in real-time search

Also see what Hakia can do with a health related questions -- diagnosing concussion

Ask (www.ask.com) parses queries to discern the question and match that to answers it can identify through its Web index and databases. Ask seems to work better with natural language queries than a fragmented keyword style.

1. first woman to climb Mount Everest
2. who was tecumseh
3. what are the qualities of an online course

Which of these is your favourite?

Check your experience against ours on the syntax answer page.

Where to Next?

Start refining your question by marking the phrase.