Political correctness. Woke-ness. Cancel culture! All symptoms of taking too seriously what people write, and how they write it? What happened to freedom of speech? Whence the Voltairian principle of ‘disapproving of what someone says, but defending to the death their right to say it’?
These are the questions that get asked if words are regrded as just words. Things that passively describe the world around us. The traditional ‘sticks and stones’ idea. But therein lies a fatal error. Words are not passive things, and their use has consequences. They don’t merely describe. They cause things to happen in the world. Writing is, in fact, a form of social action. We cannot be ‘free’ to say whatever we damn well please.
Asserting this is one thing. But can we provide empirical evidence to support such a perspective? The answer is yes, at least with regards to the specific effects of words in specific contexts. It then becomes a matter of weighing the evidence from a disparate body of studies to see if the overarching argument holds. In this piece, I want to outline one neat example of an investigation that demonstrated the potential of words to have real-world consequences. It turns out that one person’s freedom to write or say something can be the cause of another person’s loss of freedoms.
In 1989 Kitto¹ undertook a small-scale study of the effects of using different words to label a job applicant. Specifically, she set out to examine whether decisions about the appointment of an adult female to specific roles could be affected by her being described either as a ‘girl’ or as a ‘woman’ in her character reference. Should the tendency to describe adult women as ‘girls’ be seen as endearing and harmless? As mere banter? Or is there something more important and problematic going on when people do this?
The experiment involved subjects choosing which of two candidates was the more suitable for a hypothetical job. Each candidate was presented to the subjects by means of a reference that had apparently been written by her previous employer. One of the references referred to the candidate as a ‘girl’ and the other referred to her as a ‘woman’. The subjects actually performed this task twice: once for a low-status job, and once for a high-status job. Kitto’s hypothesis (her prediction) was that the candidates referred to as girls would be more likely to be thought suitable for the low-status job, and the candidates referred to as women would be more likely to be nominated as suitable for the high-status job. Sixty-four subjects participated: thirty-two women and thirty-two men.
Each subject received two job advertisements. One was for a comparatively high-status job (‘personal assistant for top executive’), the other for a lower status job (‘helper/server in a café’). Accompanying each advertisement were references, ostensibly written by a former employer, for two candidates. For each job, one of the candidates was referred to as a ‘girl’, the other as a ‘woman’. All four references were matched in terms of the personal qualities and abilities that were described for each candidate, the extent and relevance of the candidates’ previous experience, and their age (all were 25 years old). Each reference appeared an equal number of times in each of the two conditions; that is, for half the subjects reference A referred to the candidate as a girl, and for the other half of the subjects reference A referred to the candidate as a woman. This is a classic example of the experimental control technique known as ‘counterbalancing’.
Each subject was required to choose, for each of the two jobs, which of the two candidates they thought was more suitable. The order in which the subjects made these decisions was randomized, so that some saw the high-status job advertisement and references first, and some saw the low-status job advertisement and references first. Subjects were also asked to give reasons for their decisions.
Figure 1 shows that the candidate labelled ‘woman’ tended to be chosen for the high-status job, while the ‘girl’ was more often selected for the low-status job.
Figure 1: The selection of ‘girl’ or ‘woman’ for high- and low-status jobs (figure courtesy AP Grayson)
About 40 per cent of the subjects reported having been aware of the usage of the embedded terms in the different references that they saw. Kitto labelled these subjects the ‘aware’ group and compared their pattern of responses with the 60 per cent of unaware subjects. Both groups showed the same preference for the girl for the low-status job and the woman for the high-status job. ‘Thus the results cannot be explained as an artefact caused by a small number of subjects self-consciously making selections they thought the experimenter wanted’ (p. 186).
What does it all mean?
The results support Kitto’s hypothesis. The candidates referred to as girls were more likely to be thought suitable for the low-status job, whilst candidates referred to as women were more likely to be thought suitable for the high-status job.
The reasons that subjects gave for their decisions suggested that the inferences they made about the candidates were affected by the embedded terms ‘girl’ and ‘woman’. For example, the subjects for whom the candidate was described as a ‘woman’ were more likely to give their reasons for choosing her in terms of her maturity, reliability, and trustworthiness. On the other hand, the subjects for whom the candidate was described as a ‘girl’ were more likely to give their reasons for choosing her in terms of her youthfulness, vivaciousness, and liveliness. The inferences tended to be positive for both embedded terms but appeared to work against the person labelled as a girl when it came down to assessing suitability for a higher status position.
The results lend weight to the argument that the tendency to refer to adult women as ‘girls’ is more than just a harmless linguistic convention. Kitto argues that we should be aware of the subtle effects that language can have on our attitudes to others. The terms ‘girl’ and ‘woman’ are so frequently used in everyday life to label people that we may not notice the effects that they have on our perceptions of those people.
Of course, we could dismiss these findings as being ‘what happens in an experiment’ and assert that such effects don’t carry over into the real world. It would be wrong to do this, of course. There is plenty of evidence of the actual consequences that words have. One extreme right-wing conspiracy theorist’s ‘freedom’ to label those who experienced the Sandy Hook tragedy as ‘actors’ is a bereaved family’s loss of freedom to live in their home.
Words are actions. They matter.
¹ Kitto, J. (1989). Gender reference terms: Separating the women from the girls. British Journal of Social Psychology, 28(2), 185–187.
² Banyard, P. & Grayson, A. (2008). Introducing Psychological Research (Third Edition). Basingstoke: Palgrave Macmillan.