Lady, EVOLVED 
is an animated gif an image?
why is an animated gif an image?
is an animated GIF, where each frame is a page of text, an image?
ok, well, is a JPEG of a page of text an image?
ok well, if you print out a JPEG of a page of text, surely that is still an image then?
@Lady indeed, a jpeg of a page of text is an image, but a printout of that jpeg is text. My reasoning: On a computer, plaintext can be copy-pasted and have assistive tools used on it, therefore an image of text is not the same as text. But you can't do that with text on paper, so both a printout of plaintext and a printout of a jpeg of a page of text is text.
@unspeakablehorror wouldn’t the fact that you can’t copy-paste text on paper imply that it isn’t text, not that it is, if that is part of your definition of “text”?
@Lady No, because my definition of text depends on the context. On a computer, that is part of my definition of text, but in a non-computer environment it is not.
@unspeakablehorror but is not(computer‐text) not still not(computer)‐text? or do you think there is something fundamentally different between printing to an e‐ink screen and printing to paper?
@Lady Yes and no. They both have a partial overlap of definition in terms of being symbols with the ability to record and impart meaning to people, but the functional capabilities are part of the definition of text to me, and those change depending on whether text is on the computer or not. I'm viewing this in the context of a more formalized version of coercion (https://socialsci.libretexts.org/Bookshelves/Linguistics/Analyzing_Meaning_-_An_Introduction_to_Semantics_and_Pragmatics_(Kroeger)/05%3A_Word_Senses/5.04%3A_Context-dependent_extensions_of_meaning) in that I think of the formal definition as literally changing depending on the context.
@unspeakablehorror sure, but it seems like a lot of privileging of “computer” to me. what makes computers so special?
@unspeakablehorror (also, for the record, an iphone can select and copy text from a printed document (or a JPEG (or a rendering of a JPEG on a screen)), so i question your context delineation. people aren’t without computers just because they are dealing with physical entities. we live in an age of cyborgs)
@Lady Yes, computers that aren't iphones can sometimes do that as well, but image-to-text doesn't have perfect accuracy anymore than text-to-speech does. So from an accessibility context, that is absolutely not functionally the same. That action also cannot be considered the same as copy-pasting from plaintext to plaintext in terms of fidelity. So on both counts, it means the source document is definitionally not a text document with respect to the destination.