Oops! The input is malformed! Chasing the Elusive Context by Bill Inmon - BeyeNETWORK China


Chasing the Elusive Context

Originally published June 13, 2013

You don’t have to go far to figure out that you can’t make much sense of text without context. Take this simple scenario. Two guys are standing on a street corner when a young lady walks past them. One guy says to the other, “She’s hot.” Now what’s being said here?

One interpretation is that the lady that has passed by is attractive and one of the guys would like to date her. Another interpretation is that it is summer in Miami, Florida, and it is 98 degrees outside. Or it could be that the two guys are doctors, and one of the doctors observes that the lady seems overheated.

In fact, there are many other possible interpretations of what is meant here. We need context to understand what is being said. Without context, we are likely to greatly misinterpret the meaning of the utterance.

Understanding Text

The question then is how do we get context? What allows us to understand the meaning of any piece of raw text?

In the past there was an approach – natural language processing (NLP) – that attempted to derive context from text itself. So what kind of help does the analyst have using context in this case? Not much. When you look at the text, there just isn’t much that allows you to understand what is being said.

In order to understand the context of what is being said, you need to have much nonverbal information. In my opening example, we need to know if it is a hot summer day in Miami. Is it July? Is the lady attractive? Are there other important pieces of information that will allow us to make a proper interpretation of what is being said? The problem with NLP and context is that that most of what is required for understanding context is not verbal.

But nonverbal context is not the only issue with NLP. The way that something is said also greatly affects the meaning of how raw text is to be interpreted. Consider the following simple sentence: “I want to give Jean Schauer a present.” The sentence takes on different meanings depending on the way that the sentence is said.

Consider how the meaning is affected when the speaker places emphasis on different words in the sentence: “I want to give Jean Schauer a present.” When “I” is emphasized, the sentence means that it is me – not someone else – that wants to give Jean a present. When “want” is emphasized – “I want to give Jean Schauer a present” – the implication is that I have a desire to give Jean a present. I am not being coerced into giving her a present. If “give” is emphasized, as in “I want to give Jean Schauer a present,” the sentence means that I want to make a gift to Jean. I do not want to sell her a present. When “Jean Schauer” is emphasized, the implication is that my present is for Jean Schauer and not someone else. Finally, “I want to give Jean Schauer a present” – emphasis on the word present – implies that I do not wish to give Jean anything other than a present.

It can be seen from these examples that something as simple as emphasis and inflection greatly influence the meaning (i.e., context) of raw text. Unless you actually hear what is being said, you cannot divine what is being said. Stated differently, merely looking at the words that are in raw text does not tell you the context.

When it comes to context, you have to have something better than NLP.

SOURCE: Chasing the Elusive Context

  • Bill InmonBill Inmon

    Bill is universally recognized as the father of the data warehouse. He has more than 36 years of database technology management experience and data warehouse design expertise. He has published more than 40 books and 1,000 articles on data warehousing and data management, and his books have been translated into nine languages. He is known globally for his data warehouse development seminars and has been a keynote speaker for many major computing associations.

    Editor's Note: More articles, resources and events are available in Bill's BeyeNETWORK Expert Channel. Be sure to visit today!

Recent articles by Bill Inmon



Want to post a comment? Login or become a member today!

Be the first to comment!