Standard Formula for Past Counterfactual Conditional (what is the role of -te iru?)

Question

What is the standard formula for making a counterfactual conditional sentence that talks about the past? That is, English conditionals that have this pattern: "If ..... had ..... , ..... would have ...... "

I have checked these similar questions asked previously, but the answers don't seem to match up well (or at least I fail to see the common pattern):

(1) here, the working example is "安ければ買った". Non-past conditional in first clause + plain past second clause.

(2) here, the working example is "あのとき右に曲がっていれば(or いたら)、どうなっていた（の）だろう". Conditional of -te iru form in 1st clause + past of -te iru form in 2nd clause.

(3) here, the working example is "もしお金を持っていたら、食べ物を買っていた" or "もしお金があったら、食べ物を買っていた". Conditional of -te iru form in 1st clause + past of -te iru form in 2nd clause. And the alternative sentence is non-past conditional in 1st clause + past of -te iru form in 2nd clause.

The answers in the above pages made me confused because I don't quite get whether -te iru form is necesssary or preferred when making conditionals of this kind. And whether it is necessary or preferred in both parts of the sentence (first and second clause).

I also tried some examples of my own with online translators:

"If I had eaten those mushrooms, I would have fallen sick."

Both Google Translate and deepL give this: あのキノコを食べていたら病気になっていた (でしょう/だろう)

"If it had been dark, I would not have been able to see."

Google Translate: 暗かったら見えなかったでしょう。
DeepL: 暗くなっていたら見えなかっただろう。

The best rule I can build out of these is:
-te iru conditional + -te iru past is preferred in these kind of conditionals; except for when the verb cannot take -te iru form (like aru) or when the clause is an adjective or noun sentence.
Though I still am at a loss about whether -te iru past is preferred for second clause or not. Or whether using plain past vs. -te iru past in the second clause creates two sentences with different meanings.

It'd be great if someone could correct or fine-tune this rule (or perhaps present the standard rule for counterfactual past conditionals, if there is one).

Answer 1

It's a quite complex topic and there is no general way due to differences between conditional forms, so I would advise you to read a short article by Yukinori Takubo:

https://www.researchgate.net/publication/332249939_Conditionals_in_Japanese_in_Handbook_of_Japanese_Semantics_and_Pragmatics_Ed_by_Wesley_Jacobsen_Harvard_U_and_Yukinori_Takubo_NINJAL

But I will try to explain it shortly. To make something counterfactual, first, it has to happen. Besides exceptional situations with scheduling, we can't make it counterfactual with the future tense. We also need to know it happened, because otherwise it becomes hypothetical guessing. Occurrence and knowing about the result are 2 main factors for counterfactual meaning.

However, と, たら and ば can't change the tense of the verb in conditional clause and general meaning comes from the tense of consequent verb. If we use non-past form with consequent verb, our condition would also follow. As result we get a guessing about what can happen in the future and what will be a following in such case. That limits us significantly and the only way to affect it directly is to use ている form. By using ている form we can turn event into state and get rid of dependence on the tense of consequent verb. Without such dependence we can express higher range of possible scenarios. Because states are talking about something that already happened, that means we can use that either as hypothetical or factual/counterfactual condition instead of the future prediction.

However, that doesn't mean we need to use ている everywhere, because the aim of that form is to get rid of possible future interpretation. If consequent verb is in the past, we already have achieved the same step. For example, look at such sentences:

• あのキノコを食べたら病気になった. Because たら can be used factually with an one-time event, this sentence means "When I ate those mushrooms, I got sick".
• あのキノコを食べたら病気になっただろう. By using だろう we start to assume, so it's not factual anymore. Because we know if we ate it or not, it becomes counterfactual instead of hypothetical. When it's about another person and we don't know if he/she ate it or not, it's hypothetical.

Such way this sentence doesn't need ている to make it counterfactual. あのキノコ refers to those mushrooms, event is already in the past and we make assumptions about following result. That's counterfactual. However, if such sentence were about someone else and hypothetical meaning was possible, then the use of ている is preferred with both verbs similar to how we use past tense in English to make it counterfactual.