Applying Theories of Word Meaning

Let’s think again about approaches to Word Semantics and consider how they might be applied. Don’t forget, no one has all the answers, so don’t be afraid to play around a little. Don’t get too serious, though, or you might never stop!

Semantic Fields

In Semantic Fields, meaning is defined by relating or contrasting members of a set (field). These fields are called Semantic Fields.

Words can be understood as being part of a Semantic Field. The members of the field are related but not quite the same.

This is a traditional approach to understanding how word meaning works. A thesaurus, for example, provides lists of words that are similar to a particular word as a way of trying to make the meaning of words clear. Think of the words that make up the Semantic Field for the word insect. In this field, we will find words like bee, fly, ant etc. Although the words in the field are related, they will each be different in some way. A fly has wings but no sting. An ant has neither wings nor a sting.

Componential Analysis

Componential Analysis tries to define word meaning with a set of components called Semantic Primitives (Features). Six Semantic Primitives are given for the word “man” in the example above. But is [+noun] a Semantic Feature?

We don’t just want to think about connections, we also want to think about what kind of information is actually in a word. In Componential Analysis, word meaning uses Semantic Primitives (or Features) to build meanings for words. These features should be binary (either “on” or “off,” + or -) and they should apply across all human cultures. Looking at the example above, how would you build meaning for the word girl? How about dog?

[+tatami] would probably not be a good candidate as a Semantic Primitive because such features must apply across all of the world’s cultures. But how could you build a meaning for a word like “tatami” using Semantic Primitives?

Semantic Networks

Semantic Networks offer a useful way of thinking about how concepts are fixed in the mind. Rather than just having a binary [+] or [-] value, concepts can be related to other concepts in different ways. In the above example, we can define a dog as an animal that eats meat. We can also define a cow as an animal that provides meat and say that Kazuo is the owner of Fido.

If we think seriously about semantic information the way it is in our minds, just having a binary [+] or [-] feature is not really enough. Our knowledge about words relates to other words and concepts  in lots of different ways. Semantic Networks allow us to think about these relations.

Lines (arcs) represent the interrelations between concepts. We can say this is a Cognitive approach to Semantics because now we are actually thinking about how words and concepts might actually exist in our minds.

We can now define concepts by using a variety of relations. It allows us to paint a more realistic picture of the information associated with concepts. We are actually thinking about how the mind might really be organized (We are being Cognitive!). Think about what kind of relationships we should allow to exist between concepts.

How do you know when you have finished?! You can go on forever adding concepts that relate to existing concepts in different ways. When do you stop?

The problem with the Semantic Networks approach (like all Cognitive approaches) is that they soon get incredibly complicated with lines all over the place! Don’t give up. As serious language students and future language teachers, you have to be concerned with how the mind really works.