CeADAR researcher Bhawna Singh and UCD associate Prof Paula Carroll. Image: UCD/CeADAR
The researchers claim gendered terms associated with men appear 109 times in the Irish Constitution, while terms associated with women only appear seven times.
Researchers from the CeADAR research centre and University College Dublin are calling for more inclusive language in the Irish Constitution, after their study suggests there is gender bias within the text.
The team used machine learning and natural language processing tools to run a Word Embedded Association Test against a list of gendered terms, in order to detect biases in the text.
The researchers claim that they found male gendered terms at a much greater frequency in the Irish Constitution than female gendered terms. A frequency count of gendered terms in the Constitution found that the words ‘he’, ‘him’, ‘his’, ‘man’ and ‘father’ appear collectively 109 times.
In contrast, the words ‘her’, ‘woman’ and ‘mother’ only appear collectively seven times, while the word ‘she’ does not appear in the text at all.
The study also suggests that certain authoritative words are much more closely associated with men in the Constitution, while family-orientated words were more associated with women.
Some examples of the authoritative words include ‘successor’, ‘genius’, ‘chief’, ‘command’, ‘head’, ‘authority’ and ‘ownership’. Meanwhile, the list of family-associated words include ‘birth’, ‘diversity’, ‘guardianship’, ‘parent’, ‘child’ and ‘spouse’.
This study was published after Ireland’s recent family and care referendums, which took place on Friday 8 March. Both of these referendums aimed to change and redefine certain text and language used in specific articles of the Irish Constitution.
The first referendum concerned the Constitution’s definition of the ‘family’ unit and its relation to marriage in articles 41.1.1 and 41.3.1, while the second referendum sought to remove gendered text in articles 41.2.1 and 41.2.2 relating to women and mothers and insert new text recognising care given between family members using neutral terms instead. Both of these changes were ultimately rejected by a majority of the Irish public.
The study was conducted by UCD Prof Eleni Mangina, associate Prof Paula Carroll and Bhawna Singh, a researcher with CeADAR. Singh said the results show that the language in the Irish Constitution is “outdated” and that there is a need for language in future policy documents to be inclusive.
“When we have the opportunity to choose words, we should choose the gender-neutral option when we can,” Singh said.
Carroll said that the research demonstrates a need to increase visibility of women across society and said close associations between words and gender are important because “ultimately that is how humans learn”.
“If the word ‘president’ appears alongside the word ‘he’, an association is automatically formed,” Carroll said. “We might hear in the discourse that the Constitution is sexist and biased, but this research quantifies it for the first time.”
One study in 2019 claimed that Word Embedded Association Tests were the most common association test for word embedding but that they “systematically overestimates bias”.
In 2022, researchers at the Science Foundation Ireland Adapt centre claimed they found a way to reduce gender bias in natural language AI more efficiently.
Find out how emerging tech trends are transforming tomorrow with our new podcast, Future Human: The Series. Listen now on Spotify, on Apple or wherever you get your podcasts.
