(On the semantics and pragmatics of deontic should.)
I have a real-life example in mind here, from the NYT Magazine on the 17th, but I’m going to inch up to it, starting with these simpler examples:
(1) I should talk to my father.
(2) I should have talked to my father.
Both examples have the modal verb should, in its deontic sense, indicating obligation, duty, or correctness, incumbent upon some person, persons, or human institution; this is to be contrasted with its epistemic sense, indicating grounds for a judgment of truth — compare (1) and (2) with
(3) A sample this size should weigh about 10 kilograms.
(There are various ways to represent this difference, but that’s not my concern here.)
Then it turns out that deontic should can be used in (at least) two ways.
Examples (1) and (2) differ in tense (and perhaps aspect): crudely, (1) is present (embracing also a futurate present), and (2) past. But then both of them can be used in different contexts, for different effects. Again crudely, they can be positively oriented (deontic+) — “used to give or ask advice or suggestions”, as NOAD2 has it (or, more generally, to simply affirm a duty or obligation) — or they can be negatively oriented (deontic-), used to suggest a failure to carry out such a duty or obligation, “typically when criticizing someone’s actions:”, as NOAD2 notes.
(1) and (2) can be taken either way:
(1a) deontic+: I should talk to my father, and indeed I do, every day. (I talk / have talked to my father, and that’s the right thing to do.)
(1b) deontic-: I should talk to my father, even though I’ve been avoiding it for weeks. (I don’t talk / haven’t talked to my father, but I should.)
(2a) deontic+: I should have talked to my father, and indeed I did, every day. (I talked to my father, and that was the right thing to do.)
(2b) deontic-: I should have talked to my father, but he died before I got to it. (I didn’t talk to my father, but I should have.)
NOAD2 lists the deontic- use first, as the head of its should entry, suggesting (almost surely correctly) that this pessimistic use is the more common.
The deontic- use tends to be favored when the should is accented, no doubt because the accent on its own can convey contrast. This effect is very striking in the yes-no question versions, because there the should has to be accented:
(1c) Should I talk to my father? (I don’t talk / haven’t talked to my father; should I now do so?)’
(2c) Should I have talked to my father? (I didn’t talk to my father; should I have?)
Now we’ve come to the NYT example, from Kwame Anthony Appiah’s “The Ethicist” column, where the NYU philosophy professor struggles weekly with readers’ queries about everyday ethics, a substantial number having to do with family secrets. The head for this one:
Should I Have Talked to My Father About His Cross-Dressing?
My initial reaction to the head was that it conveyed that the writer ddn’t talk to his father (now dead) about the father’s cross-dressing and is asking whether he should have (that was in fact what the writer was asking), but then I saw it could be taken to be saying that the writer did indeed talk to his father about the father’s cross-dressing and is now wondering whether that was the right thing to have done.
Boiling all this down to its simplest terms, there are two very nearly opposite uses here:
‘I didn’t talk to my father about his cross-dressing; was that a mistake?’
‘I talked to my father about his cross-dressing; was that a mistake?’
You need to know more about the writer’s intentions, in this context, to figure out whether things roll deontic+ or deontic-.
Note: I’m not a semanticist, so maybe there’s already a literature explaining how the NYT head can go either way, and I’m just ignorant of it (or have forgotten it). If so, stifle your impulse to bellow, “Arnold, you ignorant slut!”, and just sketch the story (’cause there are plenty as ignorant as I am); I welcome a guest posting.
Leave a Reply