Should Kyle be corrected, and if he doesn't, why?
In a recent blog entry, Jeff Atwood quotes his sysadmin Kyle:
"Should the developers have access to the production environment, and if they do, to what extent?"
My understanding is that this sentence is only perfectly grammatical if it's parsed as:
"Should the developers have access to the production environment, and if they do [already have access to the production environment], to what extent [is it okay for them to have it]?"
However, that is not what Kyle is actually asking. It's fairly obvious that he expects the question to be parsed as:
"Should the developers have access to the production environment, and if they
do[should], to what extent?"
And sure enough, that's exactly how I parsed it when I first read it. As in, I didn't even blink at the "do" kicking out a modal verb. It was only upon reading the sentence for a third or fourth time that it struck me as odd. It should also be noted that Kyle is a native speaker. So the construction can't be that ungrammatical after all, does it? Must we correct Kyle, and if we don't, why?
Edit: some of the people who have answered/commented say that they do, in fact, read Kyle's question as:
- Should the developers have access to the production environment?
- If they already have access, to what extent is it OK for them to have it?
It looks like I have to explain why this reading makes little sense — at which I only hinted above — even when no context whatsoever is provided (and Kyle's post is well over a thousand words long).
This interpretation means that if developers already have access, the first question doesn't apply. We will throw it away just for them. But the first question is universal, it must not be thrown away. It is more general, more important. If we come to the conclusion that developers, in general, should not have access to the production environment, it also applies to those developers who already do have access. In fact, it applies to them more than to anyone else.
So, Kyle cannot possibly be asking: "Should the developers have access to the production environment? Oh, and if your developers already have access, never mind, forget that question, let's just discuss how much access is okay." Much rather, he is asking: "Should the developers have access to the production environment? And if and only if the answer to that question is 'yes', how much access should they have?"
I'm trying to capture a vague thought flitting around in my head; apologies if the result is incomprehensible.
I think we accept the should-does construction because there are situations where "do" is the appropriate question word, even though the verb isn't (or doesn't appear to be) "do":
He loves hiking, doesn't he?
The affirmative of this is "He does love hiking" --> hence the "do" in the question.
So I think what happens is that our brain tries to autocorrect the non-parallel constructions by substituting "do" for the modal verb, and if the result is grammatical, we accept it.
Should Kyle be corrected, and if he doesn't, why?
"Kyle does be corrected" doesn't work, so the autocorrect fails.
Should the developers have access to the production environment, and if they do, to what extent?
"Do the developers have access" is perfectly reasonable grammatically, so we don't even blink at this unless someone points it out. [Note that the autocorrect only applies to the grammar: I interpreted this as Kyle intended, i.e. "should they have access, and if yes, to what extent?"]
So the construction can't be that ungrammatical after all, does it?
"Does be ungrammatical" - nope.
Must we correct Kyle, and if we don't, why?
"Don't correct Kyle" - sure, no problem.
Now we just need to wait for some pedant to come up with a suitable counterexample, and the whole house-of-cards theory will come crashing down... <grin>
I think this statement suffers in part from being able to be interpreted in more than one way. I don't know the context in which it was said, but it is feasible that some or all of the developers already had access, and that access (or the extent of it) was being questioned:
"Should the developers have access to the production environment, and if they do, to what extent?"
Read as:
"Should the developers (even) have access to the production environment, and if (some of them already) do, to what extent (should they have it)?"
I'll admit that was not my initial reading of it, but it does make sense. My initial reading of it was assuming no developers currently have access:
"Should the developers have access to the production environment, and if they do, to what extent?"
Read as (simplest interpretation/better restatement of intent):
"Should the developers have access to the production environment, and, if so, to what extent should they have access?"
However, what I think is implied by the way it is phrased is:
"Should the developers have access to the production environment, and if they do (at some future point have access because we have decided they should and have given them access), to what extent (should they have access)?"
I think it is the shift in time from now to a later point that causes most of the confusion because both points are referred to with present tense verbs. If you assume no developers currently have access to production, the statement is asking in the present if they should. Then, in the same breath, the statement is referring to a future point when they have access or might have access also in the present. I don't think it's technically incorrect, but much of the meaning has to be inferred since it relies on the reader to supply the missing words that could have provided clarification.
Compare the sentence with
(?) If the developers should have access to the production environment, they should have limited access.
If already introduce a conditional sentence, and should is not necessary.
The sentence should be written as
If the developers have access to the production environment, they should have limited access.
I interpret the sentence you wrote as
Should the developers have access to the production environment, and if they do [have access to the production environment], to what extent [is it okay for them to have it]?
In the first part of the question asks if developers should have access to production environment; the second part takes the assumption the developers have already access to the production environment, and asks to which extension should the access be allowed.
I think the problem with this sentence is that, unlike many present tense verbs, "have access" can't be used for the future. For example, if the sentence was
Should the developers go to the meetings, and if they do, should they have a vote?
then the implied "do go" fits the future fine.
But "have" is exceptional; you can't use the present tense of "have" in the future; it's replaced by "get". For instance, "tomorrow I go to school" is fine, but "tomorrow I have a car" is not; you should say "tomorrow I will have a car" or "tomorrow I get a car".
The question becomes whether the "do" permits the sentence to shift into the future, even though present tense "have" isn't used for future events. If it does, then the expanded sentence is really something like
Should the developers have access to the production environment, and if they get access, to what extent?
If it doesn't, you would have to use "should" or "will" instead of "do". From people's reactions to this question, it seems many people think this shift is grammatically implied by the "do". However, replacing "do" by "should" is clearly the better option.