A signed track does not retire the captions
Aliases: captions and sign · interpreter window · sign does not replace captions
What it is
An interpreter in the corner does not tick captions off. The two media cover overlapping, not identical, people. The signed track speaks to users of that sign language. Captions write the source speech for hard-of-hearing people, people deafened later, Deaf people who do not know this sign language, and hearing people who cannot turn sound on right now. Ship one without the other and a group falls out.
Why it happens
The channels differ. Sign is video in another language. Captions are print in the source spoken language. Signers whose print is not their strong suit need the first. Literate people who do not know this sign language need the second. The interpreter frame is usually small; low-vision viewers cannot resolve handshape and may still need type. In conformance terms, pre-recorded captions sit on the mid-level baseline; sign language sits higher. Using the higher increment to retire the baseline inverts the stack.
Situations split the same way. A hearing person in a quiet carriage who mutes wants type, not a language they never learned. Captions alone ask sign-L1 users to work in a second language. Both tracks are a union. Calling one a substitute shrinks the union to a subset.
Studying it
Cross the groups: sign L1, print-dominant Deaf or hard-of-hearing, hearing viewers in silence. Three stimuli: sign only, captions only, both.
Independent variables: media combination, language background. Dependent variables: comprehension by group. Expect a cross — one group high on sign-only, another high on captions-only; either single channel leaves a group near zero.
Do not pool a “deaf mean.” Pooling makes “sign is enough” look true.
Where it stops holding
When the entire audience is one sign community and the work is born in sign, captions become a translation into print and the priority shifts. Even then, public redistribution still meets people who want type. If the interpreter is the main picture and captions are offered as a second-language translation, that is bilingual distribution — still two tracks, not substitutes. Live interpreter windows and live captions both lag; dropping one leaves that group with no channel for those seconds.
Applying it
- Sign is an increment, not a reason to drop the caption track. Both rows stay on the release checklist.
- The interpreter frame must not cover captions; captions must not cover hands and face.
- How to check: mute and cover the interpreter — can captions carry the piece? Then cover captions and watch only the interpreter. Each path needs people who can follow it alone. Missing a path misses a group. It is not “we already have a video alternative.”