Introductory models & basic concepts: meaning

Semantic differential

The meaning of a word is its use in the language

Ludwig Wittgenstein (1958)

The use of the word in practice is its meaning

Ludwig Wittgenstein (1969)

If you've looked through the section of the Infobase which deals with meaning, you'll have seen that, following Wittgenstein's advice, we've tried to look at the way that people actually use signs. Following Berlo, we looked at four dimensions to meaning:

structural
contextual
denotative
connotative

Charles E Osgood (from Crystal (1987)

The psychologist and communication scholar Charles E. Osgood developed work in the measurement of the last of those dimensions of meaning: connotation (see Osgood (1957)). His concern was with semantics and he devised a method to plot the differences between individuals' connotations for words and thus map the psychological 'distance' between words. Osgood's method is known as the 'semantic differential'.

Subjects were given a word, for example 'car' and presented with a variety of adjectives to describe it. The adjectives were presented at either end of a seven-point scale, ranging from, say, 'good' to 'bad' or from 'fast' to 'slow'. In this way, he was able to draw up a 'map' of people's connotations for a given word. In the window opposite, you can find Osgood's map of people's connotations for the word 'polite'. The graphic shows ten of the scales used by Osgood. The graphic maps the average responses of two groups of 20 people to the word 'polite'.

Factors of judgement

Osgood's method is a development of the Likert scale in that Osgood adds in three major factors or dimensions of judgement:

• evaluative factor (good - bad) - that can be seen in the example as 'Good-Bad', 'Fresh - Stale', 'Cold - Hot')
• potency factor (strong - weak) - seen in the example as 'Weak - Strong'
• activity factor (active - passive) - in the example as 'Active - Passive', 'Tense - Relaxed'

The intention is that this should differentiate between attitudes in a way that other methods don't and should shed more light on the links between attitudes and behaviour. Osgood gives the example of two subjects' connotations for the word 'negro':

Subject 1: unfavourable, strong, active

Subject 2: unfavourable, weak, passive

Both views are unfavourable, but the actual behaviour of each subject towards negroes may well be different. Subject 1 might be inclined to treat negroes in a placatory manner, anxious to avoid conflict; subject 2 might be inclined to treat them exploitatively, being more likely to boss them around.

Practical work

You may well be concerned in your practical work to measure people's attitudes. For example, it could be that your artefact is intended to change their attitudes to a certain subject, service or product. If so, Osgood's semantic differential is one way of measuring their attitudes, both before and after exposure to the artefact you have produced.

You may also wish to produce a measurement of their responses to certain aspects of your artefact. Here again, the semantic differential technique can prove useful.

It could be that you want to create a certain image for your artefact. Marketers use Osgood's semantic differential as part of their battery of tests to determine how a product is perceived - we all know what beans are, but what are the differences in meaning, in image between Heinz's and Sainsbury's? Think of describing your artefact as you would a person. Get a list of adjectives and then set them out as Osgood suggests. Show your drafts to some members of your audience - does their rating of your product correspond to the one you're aiming for?

Pitfalls

There are objections to Osgood's method - one is that it appears to assume that the adjectives chosen mean the same to everyone. Thus, the method becomes self-contradictory - it starts from the assumption that people's connotations for a word differ, but has tor rely on the assumption that, for certain words at least, they don't. Looking at the example shown , how do know that your 'tense/relaxed' is the same as my 'tense/relaxed'?

There are often objections that the correlation between stated attitudes and actual behaviour turns out to be quite low and attitudes are often poor predictors (as shown in La Piere's study). This should be overcome somewhat by the relative sophistication of Osgood's method, but there are still the problems of

• respondents giving socially desirable responses - what they think are the 'right' answers
• respondents developing a response set - giving consistently moderate answers, or always either agreeing or disagreeing

Solutions

You can overcome some of these problems by:

• very careful wording so that agreement with some questions and disagreement with others indicates an attitude - if you get, say, agreement with them all, then you know that the respondent has probably developed a response set
• stressing the total anonymity of the responses - that may militate against the desire to give socially acceptable answers
• emphasising the importance of the work, especially if your practical work seeks to be 'socially useful' - this may encourage honesty
• the inclusion of some questions which are completely irrelevant - in that way you can disguise the purpose of your test

Related articles:

Ang on meaning

Baudrillard

Berlo on meaning

Fish on anti-formalism

Fish on interpretive communities

Interpersonal Communication: Language

semiotics

On the measurement of attitudes, see:

Kelly's Personal Construct Theory

Likert Technique