How Oxford Languages is responding to user feedback
At Oxford Languages, we are continually working to provide accurate and meaningful accounts of languages in our dictionaries.
Our dictionaries reflect, rather than dictate, how language is used. This is driven solely by evidence of how real people use language in their daily lives and drawn from huge banks of real-world examples. That real-world evidence is drawn from a variety of sources. For example, Oxford Languages editors utilise our extensive corpora to track and record current language developments across a vast variety of publications, covering everything from specialist journals to newspapers to social media posts. Our expert team of lexicographers analyse this data and use it to inform our dictionary entries and examples. Language is always changing so our lexicographers are constantly reviewing and updating this information. In doing this, they work to ensure definitions or sentence examples that repeat factually incorrect, prejudiced, or outdated statements are updated or eliminated in our periodic content updates.
We also work continually with a network of trusted partners across the world to review the dictionaries' content and feedback from readers, seeking to ensure that information is accurate and up to date. We constantly collect feedback on the way our dictionaries and datasets are written, and we use this feedback to inform our content update process. We do this in two ways.
Our partners develop products that make use of our dictionaries and datasets. They collect their own user’s feedback to pass on to us at regular intervals. We also collect our own feedback through our contact form and social media. Using this two-pronged approach, we work closely with our licensing partners to make sure we achieve a shared, high-level of visibility over all potential issues with our datasets. We feed this information into our content update process to ensure our datasets remain of the highest standard.
How users can report an issue with a dictionary entry:
We welcome feedback from users, especially in cases where users feel we may have fallen short of our rigorous quality standards, whatever the reason. To report an issue with an entry, please use our contact form, and select I have a question about your dictionary products, followed by I have a comment or concern about a dictionary product.
Reviewing our Brazilian Portuguese dataset
Oxford Languages provides language data for a wide variety of languages for the Dictionary feature in Google Search. This provides the opportunity for people across the world to discover high quality, evidence-based language insights. However, we have recently been made aware of some concerns in Brazil about the definitions people are seeing for words like ‘patroa’ and ‘mulher-solteira’. We take these concerns seriously and have carried out an extensive review of these definitions. In addition we are undertaking a more systematic review of several culturally-evolving subject areas within the dictionary to ensure they reflect up-to-date usage and perspectives.
So how and why did these definitions appear in the dictionary? Our dictionaries reflect, rather than dictate how language is used, which means that we include words that may be considered offensive but that are in common use, even where we wouldn’t use them ourselves. These words should be clearly labelled as offensive. We also include regional terms and terms that have fallen out of contemporary usage but which people may still come across to ensure we can give an accurate and detailed picture of the language as a whole. For example, if you’re reading a novel that’s from the 1900s, we want you to be able to easily find and understand unfamiliar terms and, more importantly, understand how and when they should be used, especially if they might be considered offensive.
And what have we done as a result of our review? We’ve implemented a number of changes to help make these definitions clearer and more helpful to people. In the case of ‘patroa’, the definition no longer reflects modern use by Brazilian Portuguese speakers today and has been updated to be more accurate. In the case of ‘mulher-solteira’, although our research showed that the definition displayed was still an accurate reflection of how language is used in specific regional cases, the way in which the definition appeared was misleading and unhelpful so it has been removed from display. These changes are also reflected in Google Search.
Over the coming months we are also launching a more in-depth review of the language directly related to gender, race, and social justice to ensure the definitions are appropriate and accurately reflect current usage. Creating and maintaining a dictionary is a never-ending task which must be anchored in the need to truly capture and reflect language as it’s being used, and feedback from real language users across the world plays a vital role in this pursuit.
Mapping 'woman' in the Oxford Dictionary of English and Oxford Thesaurus of English
Oxford University Press (OUP) publishes many different dictionaries, and we welcome input from members of the public to inform their ongoing revision. A recent public petition has expressed concerns about the treatment of the word woman in one of our dictionaries; the petition raises some useful points for consideration, but it also presents an opportunity for us to provide some context about the methods of lexicography. Often—and this case is no exception—there is a nuance to how words are defined and presented in our dictionaries, which is important to take into account.
Firstly, the petition refers variously to ‘Oxford’s English Dictionary online’ and ‘the Oxford English Dictionary’. It is worth noting that none of the content mentioned in the petition appears in the Oxford English Dictionary (oed.com), which is a scholarly dictionary covering the full history of English. The content discussed in the petition derives from the Oxford Thesaurus of English and the Oxford Dictionary of English, which aim to cover contemporary English usage and are accessible online in a variety of formats.
These texts are based on the methodologies of descriptive, corpus-based lexicography, meaning that editors analyse large quantities of evidence from real-life use to determine the meanings of words. If there is evidence of an offensive or derogatory word or meaning being widely used in English, it will not be excluded from the dictionary solely on the grounds that it is offensive or derogatory. Nonetheless, part of the descriptive process is to make a word’s offensive status clear in the dictionary’s treatment. For instance, the phrase the little woman is defined as ‘a condescending way of referring to one’s wife’, and the use of ‘bit’ as a synonym for woman is labelled as ‘derogatory’ in the thesaurus. Sensibilities regarding language are constantly changing, and our editorial team is always grateful for feedback to ensure that the status of offensive or denigrating terms is clear to our readers.
Because our dictionary is based on evidence of actual usage, entries for two different words will only be perfectly parallel if the words are genuinely used in a perfectly parallel way. Our editors are investigating whether there are senses of woman which are not currently covered but should be added in a future update. Nonetheless, the dictionary entries cannot be completely parallel unless evidence shows that the words man and woman are used in identical ways; at present that is not the case. For instance, sense 4 of man is defined as ‘a figure or token used in playing a board game’, but that meaning is not evidenced for woman. Statistical analysis of large digital text databases, or corpora, shows that there are significant differences in how the words man and woman are used—people speak of a ‘man about town’ but rarely of a ‘woman about town’; of a ‘ladies’ man’ but not a ‘gentlemen’s lady’; of ‘womanly’ curves and wiles, but ‘manly’ handshakes and jawlines.
The relationship between the dictionary and the living language is more like a map than a set of directions; it can tell you the contours of the landscape, but not direct you on where to go or how to get there. As the usage of English speakers changes over time, the dictionary changes to reflect that new lexical terrain. The current cultural moment has seen an increasing acknowledgment of the real-life impact that words can have on individuals and groups. As this awareness leads to changes in linguistic behaviour, the dictionary will seek to record them.
Katherine Connor Martin, Head of Content and Data