How should we understand implicit bias?
Biased by Our Imaginings
Ema Sullivan-Bissett, University of Birmingham.
Members of marginalized groups are underrepresented in various professional domains: currently in the UK, only five FTSE 100 company CEOs are women1, five are BAME (and just one is Black).2 Of Professors in the UK, 25.5% are women, 8.8% are BAME (and just 0.6% are Black).3 Lots of explanations for these sobering statistics can be given, ranging from explicit prejudice against women and people of colour to sexist and racist institutional and societal structures. Another key component though is what are called unconscious or implicit biases, which are extremely widespread (don’t believe me? You can take an Implicit Association Test here to find out what biases you hold). Implicit biases have been shown to negatively affect the way we make judgements about, and interact with, members of marginalized groups.
What are implicit biases? There are broadly two schools of thought concerning this question: associationism and propositionalism.
We all have stored mental associations, when I say salt, you think pepper, when I say Thelma, you think Louise, and so on. Social psychologists have understood implicit biases as associations, whose existence can be traced back to our environments. For example, I might have an implicit association linking my idea of women to caregiving because of my exposure to Western popular culture – a culture that has traditionally presented women in such roles (and not, for example, in roles of leadership).
Recently philosophers have subjected this associationist picture to scrutiny, and some have argued that implicit biases are propositional. What does this mean? Return to your association between salt and pepper. In this case we think that there’s no relationship that holds between salt and pepper, rather, hearing one just activates (makes one think of) the other. But now suppose you become engrossed in a TV series about two academics, Dr. Salt and Dr. Pepper, engaged in a Shakespearean love affair across University faculties traditionally at odds with one another. Now, the relationship in your mind between salt and pepper is not just a matter of one activating the other. It’s more specific: there’s a relationship between these ideas, namely, that Salt loves Pepper.
What does this have to do with implicit bias? Well, if my implicit bias regarding women and their propensity to caregiving is propositional, it’s not just that my idea of women is somehow linked up to caregiving, it’s that there’s a relationship in my mind between those ideas, namely, women are caregiving.
So which is right: associationism or propositionalism? The evidence is mixed, and many researchers have pointed out the need to recognise that implicit biases come in various metaphorical shapes and sizes. Unfortunately, current models of implicit bias are either fully associationist or fully propositionalist and so are unable to accommodate this diversity.
My account of implicit bias is the exception to this general rule. On my view, implicit biases are made up of unconscious imaginings, and this can happen in an associative or propositional manner. There might be separate associatively linked unconscious imaginings– I might have an imagining of a woman, and an imagining of caregiving, and the activation of the first leads to the activation of the second. Equally, I might unconsciously imagine that women are caregivers.
My unconscious imagination model is uniquely placed to recognize that the category of implicit bias admits of much important diversity. Implicit biases are widespread and harmful, and if we are to effect change, we need to better understand the nature of the problem. The problem, it turns out, is one for which a one size fits all account will not do.
1. Article from EU-Startups.com on the top female CEOs of the FTSE 100 index.
2. 'Only 5% of CEOs in the UK, US and Canada's biggest listed companies from ethnic minority backgrounds.' Blog post from recruitment-international.co.uk
3. Equality + higher education. Staff statistical report 2019.
Department / Institute / Centre