Rattigan: If only that was possible.
The Doctor: If only that were possible. Conditional clause.
Of course the Doctor's just trying to annoy Rattigan. But who's right?
If only that was possible is a counterfactual condition - it refers to something that is not real or true. We know this. No one seriously thinks "he used was, not were! He's saying that it really is possible." And yet it's a favourite pasttime of peevologists to find instances of counterfactual was and insist they should be were. For instance, this rant on how counterfactual subjunctive were is an "important linguistic construct," and if we lose it, something terrible will happen.
There are many kinds of conditional clauses, but here are four, to keep things simple (borrowed from here):
1. present possible condition: I wonder if is possible.
2. past possible condition: I wondered if it was possible.
3. present counterfactual condition: If it was/were possible, I would travel back in time.
4. past counterfactual condition: If it had been possible, I would have traveled back in time.
Many speakers and writers use was in both 2 and 3. We can also use were in 3 - it's usually called the past subjunctive or the irrealis.
In using was, we're not "losing the subjunctive", we're simply using a different form. And in doing so, we are bringing be in line with every other verb in English, where we use the simple past tense:
5. If I was in Paris, I would visit the Eiffel tower.
6. If I lived in Paris, I would visit the Eiffel tower.
Why use were in sentences like 3 and 5? One argument is that it avoids ambiguity. But is there ever any real danger of confusing a past possible condition (as in 2) with a present counterfactual condition (as in 3 and 5) because we use the same verb form in both? The Doctor might argue that we don't know if sentence 5 is a possible condition or a counterfactual condition. Except that it's painfully obvious that it's counterfactual - it cannot be anything else. (The would in the main clause is a big clue.) And if there was (were) ambiguity, we would encounter this ambiguity with every other verb in English. But we don't. Sentence 6 is clearly counterfactual, even though it uses the simple past tense lived, which can be used in non-counterfactual clauses:
7. I asked him if he lived in Paris from 1970 to 1975.
Also, be has a special distinct form only used in the first and third person singular: if I were, if he/she/it were. Therefore, if you were, if we were, if they were are all potentially ambiguous - and yet in context there is never any confusion.
Furthermore (as Zwicky notes), the complainers are very good at noticing when was is a counterfactual and so "should" be were, thereby showing that they understand the construction perfectly well. The Doctor correctly identifies "If only that was possible" as a counterfactual condition. So ambiguity is not the problem.
So what is the problem? A good usage book will tell you that while counterfactual was is common and normal in speech, were is common in formal writing. The American Heritage Book of English Usage notes
In fact, over the last 200 years even well-respected writers have tended to use the indicative was where the traditional rule would require the subjunctive were. A usage such as If I was the only boy in the world may break the rules, but it sounds perfectly natural.
And if it sounds natural, you have to wonder about the rules it's supposed to be breaking.
So the difference is one of register. To insist that we should use if it were in conversational speech is to be unhelpful and obtuse. But it is very good for annoying child geniuses.