“Imagine if I hadn’t of been there!” said someone in an email to my brother, Richard. He regarded the sentence coldly, as if it were a slimy creature emerging from under a rock. What’s that of ? A misspelled extra have ? Why? Doesn’t had suffice? He turned to the grammarian in the family, and asked me what had gone wrong.
It’s an interesting puzzle that teaches us something about drawing the subtle distinction between intralinguistic slips and interlinguistic variation.
Let’s start by setting aside the spelling error. It’s quite common, and not irrational (in piece of cake the spelling of stands for exactly the same sound as a reduced have with its “h” dropped). The standard spelling for reduced have is ’ve, as in I should’ve known (or -a in some representations of colloquial speech: I shoulda known).
We can also ignore the ungainliness of hadn’t’ve. The puzzle holds for had, as in [1] — which I prefix with ‘%‘ to mark it as grammatical for some speakers of standard English but not others.
[1] %Imagine if I had’ve been there.
The puzzle is, why the extra ’ve? I would have written [1] as [2] (which I think everyone accepts, hence no percent sign).
[2] Imagine if I had been there.
You may feel like saying that the people who write [1] are just wrong. But we need a reason why it isn’t me that’s wrong, for inexplicably leaving out a have that at least some people feel is needed.
The extra-have people have been around for a long time, possibly since the 15th century (see this paper [open access]). Though Henry Fowler (1926) called the construction “no better than an illiterate blunder, and easily shown to be absurd,” the similarly prescriptive Eric Partridge in 1942 called it “an error by no means confined to the illiterate.” I think they’re both wrong; error, in the sense of making a slip, is the wrong way to look at it.
The main problem with calling it a “blunder” is that it doesn’t explain the data. Those who use [1] aren’t random blunderers. They don’t insert arbitrary extra syllables in miscellaneous places. It’s just one extra have (usually reduced), and in only one kind of context.
What context? Not ordinary declarative main clauses: Even speakers who accept [1] don’t write things like [3] (which I prefix with an asterisk to mark it as flatly ungrammatical).
[3] *Previously we had’ve always locked the car.
Instead, had’ve only occurs in what I’ll call irrealis clauses: counterfactuals with if, or after verbs like wish or imagine:
[4] %If only we had’ve locked the car.
[5] %I wish we had’ve locked the car.
That separates had’ve from sequences like could’ve or will’ve, where a reduced have occurs after a modal auxiliary verb. That is found in plain positive declarative main clauses: He could’ve been killed; By now she will’ve told him. And it’s just the ordinary perfect tense have, as in He has been killed; She has told him.
The had that gets the extra have, by contrast, is limited to irrealis contexts. And it’s a special had: It has the past tense shape, but doesn’t refer to past time. In if I had a hammer, the time of me having a hammer is not in the past; it’s in a hypothetical state of the world that doesn’t exist yet. The Cambridge Grammar of the English Language calls this use of a past tense form the remote preterite. (By the way, only auxiliary had gets the extra have, not the main verb signaling possession: Nobody says *if I had’ve a hammer!)
Two key features are found only in irrealis clauses: the peculiar first- or third-person irrealis form of be, as in if I were a carpenter; and the remote preterite, as in if I went home.
So I think the solution to the puzzle is this: People who accept [1] have subconsciously formed the view that the correct form for the auxiliary verb have in irrealis contexts is (or can be) had’ve (or had have ), not just had. They’re not blundering. They’re following the rules of their variety of standard English. And it differs from mine (and perhaps yours, too) only trivially: For me the right form of auxiliary have in irrealis clauses is had; for them it’s had’ve.
Please don’t ask me why they tacitly formed their view; that’s a whole different layer of mysteriousness. I’m not trying to explain the historical evolution or the acquisition at this point. My goal is nothing more than a description in terms that relate the phenomenon sensibly to what we already know. Richard deserves at least that much when he consults his grammarian brother.