One of those truths we hold to be self-evident as archivists and oral historians is that we don't mess with an original audio or video recording. If it is born digital, preferably in a non-lossy format, we create a clone; and if we feel it necessary to edit, we then edit the clone. If the recording is carried on older tape or film media we certainly don't cut and splice. We leave the original, the master recording, alone.
Why? With physical media, cutting and splicing undermines the physical integrity of the medium. In materials which are sandwiched together only through the force of human ingenuity and the strength of molecular bonds, we break those bonds with a blade, force the stumps together, and overlay them with the assault of a kind of over-arching sticking plaster which doesn't really want to be there, and which immediately begins laying chemical plans to come away, plans which are aided and abetted by the bump, grind and friction of the mechanical playback mechanism, and possibly by atmospheric dust and organisms which slipped in between stump edge and stump edge or got trapped in the glue of the splice during surgery. The digital equivalent I guess would be translating a non-lossy format file, like .wav, into a lossy format file, like .mp3, and throwing away the original: It's not an exact equivalent, because in the latter case it is more like taking an algorithmic brillo pad to the entire recording, mathematically scrubbing out audio data that experience says the ear will not miss. But in both cases the fullness and integrity of the original is irreversibly destroyed; a part of the world that used to be is no more, and can not be restored.
But editing as such is more than an assault on the integrity of the original; it eliminates context, it eliminates accountability, and it undermines the evidential authority and authenticity of the interview. Editing the original/the master means that it is impossible to know what has been lost, and how and why the interview has been manipulated. The editor will have one set of reasons for eliminating material, which they may or may not be able to articulate; but there may (indeed, almost certainly will) be losses the editor is not aware of (the history of archaeology illustrates how this works); or bias, or unconscious processes at work which the editor can't, won't, and/or doesn't put into the record. The edited recording becomes an account; it is no longer an authoritative document.
{[Orignally published January 20, 2013]