This workshop brings together project researchers and Bradford partners connected to several local oral history collections/projects. In preparation for the event, the project team are running some of the digital tools and processes they have developed on the oral histories that partners are bringing. In the workshop we will give the results a good look.

Inspired by these case studies and a bit of tinkering with digital tools like Memory Mapper, we will discuss what we think in terms of potential, ethics, and practicalities of these technologies for usefully linking oral histories.

Content Notes

Language – dialects, accents, multilingual

• Silence is important (pauses)
• Accents inform cultural identity
• Not to lose the importance of sayings and dialects
• ‘Flattening of language’
• How is dialect, slang, accent, emotion lost?
• How do you encapsulate dialectal variation? People choose words for reasons
• Different connotations for same words, works at different pace to tech
• Implicit knowledge that comes in for reading humour – generative process - is that where the social machine enters? Human time to add context etc, while AI does transcription, indexing, etc.
• Subtle nuances & able to detect jokes/humorous self-referential remarks? Dry humour – can AI detect that?
• AI opens up chances to mirror someone’s accent/words/ways of speaking
• Technology provides opportunity to document changing language -> youth culture over time
• Transcribing services need human intervention, who is the right human to do this?
• Is anyone double-checking Whisper transcripts? Dialect and slang must be preserved
• Intonation – how something is said, not just what is said - ‘reclaiming of words’
• Is language linear? Can we communicate with each other with clusters of meaning, rather than sentences?
• Closed captioning + creative captioning – automated – who has responsibility for this?