Machines & Meaning · Capstone brief
HUM-05 · Translation and loss
Compare human and machine translation on a short literary or historical passage; annotate what is lost.
The question
Compare human and machine translation on a short literary or historical passage; annotate what is lost.
Texts / materials
Short source passage (public domain or licensed). Human translation (student’s, published, or class-produced). Machine translation from a school-approved system. Annotation scheme for loss (sense, register, allusion, rhythm) declared first. Non-claim list: MT errors ≠ proof about mind.
Expected failure modes
Mocking MT without a scheme. Declaring the human version automatically “correct.” Inventing a translation-studies citation. Using a passage the student cannot actually read in the source language without help they fail to disclose.
Done looks like
Parallel text; annotation set; loss memo; disclosure of the student’s source-language competence; non-claim list.
Five C's
CT: loss as analytic category. CR: pre-declared scheme. CO: peer checks one annotation. CM: loss memo. CZ: source-language community / author.
Mentor role
Language teacher reviews competence disclosure and scheme. School-supervised.
Rubric calibration
R1: one passage, two translations. R2: scheme applied. R3: human baseline named. R4: competence limits stated. R5: memo clear. R6: non-claims on mind/meaning.
Two ways this goes wrong
(a) Gotcha screenshots only. (b) “MT has no soul” as the thesis.
Credit lane fit
Lane A immediately. No verified credit claim.