Etymologically, ‘literal’ means ‘Of, relating to, or of the nature of a letter, or the letters, of the alphabet’. To be precise about what I mean by ‘etymologically’ here, I’m referring to the earliest English use of word as recorded in OED3 (John Trevisa, a1398). I don’t mean the prior senses of Middle French literal, litteral,: ‘of or relating to literature (14th cent.), of or relating to the ‘letter’ of a text … (late 14th cent.)…’ (OED3, etym), nor do I mean the earliest Latin sense of litteralis, ‘relating to letters (3rd cent.)’ (OED3, etym), though I might as well, since this first sense overlaps pretty well with Trevisa’s sense.
‘Literal’ no longer carries this etymological sense, that of relating to letters. But it does carry the sense of ‘etymologically’ in some contexts:
Of, relating to, or designating the primary, original, or etymological sense of a word… (OED3, 5c.)
… and has done since 1597. Of course, ‘etymology’ is literally–I mean etymologically–a discourse on true meanings, as the etymon of ‘etymon’ bears out:
< Latin etymon, < Greek ἔτυμον (originally neuter of ἔτυμος true): (1) the ‘true’ literal sense of a word according to its origin; (2) its ‘true’ or original form; (3) hence, in post-classical grammatical writings, the root or primary word from which a derivative is formed. (OED2 ‘etymon’, etym)
But see how OED uses quotation marks to indicate a restricted (non-literal?) use of ‘true’ here? That’s because the etymology is no more true than the current sense or senses of a word, and may be a good deal less true. No wonder ‘literal’, ‘literally‘ is being used to mean ‘truly’, ‘actually’ out there in the real world, as well as ‘etymologically’, since ‘etymologically’ already etymologically encodes a claim to truth.
Etymological discourse is pervasive. Often etymologies are invoked playfully or suggestively, e.g.:
We do not seem to be in control of our emotions. They just happen to us. This common observation is etymologically enshrouded in the term ‘passion.’ [“Recalcitrant Emotions and Visual Illusions“]
Nothing rests on this link; it’s there for texture. And it textures our understanding of ‘passion’ as much does our understanding of emotion.
But there exists a stronger form of etymological rhetoric, which derives authority of argument from the persisting association of etymology with ‘true meaning’. And this in our post-Saussurian world. Here it is at work in academia:
It was mentioned above that engineering produces artefacts. Etymologically, ‘art’ and ‘artefact’ share the Latin root ars, meaning craftsmanship or skill. The second part of ‘artefact’ is from the Latin facere, meaning ‘to make’. Thus software and SD share features of the sciences, engineering and the arts… [“Towards a Philosophy of Software Development: 40 Years after the Birth of Software“]
Talking out of his ars, you say? Do the shared features of software and the sciences, engineering and art depend in some vital way on the historical meanings of a word they happen to share? Here’s a subtler use in the realm of politics, or civics:
Civility is neither a small nor inconsequential issue. The word comes from the French civilite [sic] which is often translated as ‘politeness.’ But it means much more. … The words ‘civilized,’ ‘civilité,’ and ‘city’ share a common etymology with a word meaning ‘member of the household.’ To be civilized is to understand that we live in a society as in a household. [Congressional Record, 11.19.2009]
I have no idea what etymological relationship the speaker, Lyndon Olson, is thinking of here (unless via PIE *kei-1, ‘bed’, ‘couch’ > OE ‘hind’? A stretch). But it’s clear that the claim of etymological relation is here underwriting the analogy of society to household (which would be a perfectly decent analogy even without this relation).
Now, observe the same kind of argument, this time with ‘literally’ carrying the etymological authority:
Even if the teacher’s purpose is ‘to lead out’ from that home (the literal meaning of ‘educate’) the teacher must still know from where he or she is trying to lead the student. [“Toward a New Pluralism in ABE/ESOL Classrooms”]
And here it provides the structure for an enhanced concept of a word:
Parody literally means to ‘stand beside’ so that, for Hutcheon, it has a structural as well as a mocking relationship to the predecessor it models. [“An Ethical Model in a Postmodern Faust“]
What I think is special about the way that the ‘truly’ senses of ‘literally’ and ‘etymologically’ overlap is that in most of these cases the rhetorical intention is to restrict meaning to a preferred (‘true’, ‘original’) sense, while the practical effect is to broaden meaning in a figurative turn.
Thinking of parody as a ‘standing beside’ is highly metaphorical, whether or not there was ever an ancestor of that word form that literally meant ‘standing beside’. When I think of my profession as a ‘leading out’, I’m adding to the connotative range of ‘education’, not limiting or delimiting its denotative range. I don’t need to know that ‘civil’ and ‘household’ are related (and I don’t know that) to appreciate the analogy between society and family, but the purported relationship helps rhetorically to bring the ideas into proximity. Similarly, the etymological tree that branches into software engineering and the arts promotes thinking of one in terms of the other.
Those are all metaphors – they all think of something in terms of something else. That the link between those things is in each case contingent, an accident of language, doesn’t take away from the imaginative value. Rather it adds value. The etymological turn is anything but literal, and gets less literal the more it claims literalness. Similarly, ‘literal’ used in this way gets more and more figurative the more it lays claim to etymological truth. This may partially account for why Mark Liberman thinks Watkins’s Dictionary of Indo-European Roots ‘has value as a source of found poetry‘, and maybe also why etymologies are so interesting in the first place.
By contingent I don’t mean random or arbitrary.