Two sentences:

1. He may come tomorrow.
2. He might come tomorrow.

Some would state that "might" indicates less likelihood than "may", in such cases. Some might even attempt to assign probabilities in terms of percentages.

How reasonable is this?

We know that for event E, the sum of the probabilities of all possible outcomes must equal one (or 100%).

For instance, if E = a fair roll of a fair die, there are 6 possible outcomes, each with a probability of 1/6. Their sum = 1 = 100%.

Now take this example:

3. He may come tomorrow. But then again, he may not.

There are two sentences. Each sentence embodies one outcome. Each outcome is presented as of equal probability. Each therefore has a probability of 1/2 = 50%.

Now take another sentence:

4. He might come tomorrow. But then again, he might not.

Again, we have two sentences, and two possible outcomes of equal probability. So again, each outcome has a probability of 50%.

But how can that be, if "might" indicates less likelihood than "may"?

___________

Perhaps we need to be more flexible. Suppose the probability values of "may" and "might" are indeed different, but vary from event to event. In that case, we would expect to see many instances of these two structures:

5. He may do X; but then again, he might not.
6. He might do X; but then again, he may not.

Here, for instance, "may" might express a 67% likelihood, and "might" a 33% likelihood, the sum of which would = 100%.

If that were the case, in #5, X would be twice as likely as not-X; while in #6, X would be half as likely as not-X.

That sounds plausible; and it accords with received "may"/"might" wisdom.

Yet the strange thing is, we don't see many sentences of this kind. If you google on

7. "he may * but then again he might not"

you'll find very few examples, by comparison with

8. "he may * but then again he may not"

The latter (#8) is the common form; the former (#7) is the exception.

If "might" and "may" can express differing likelihoods, why don't we combine them to do so more frequently, after the model of #7?

MrP

<5. He may do X; but then again, he might not.
6. He might do X; but then again, he may not.>

Two levels of liklihood there. "May" can expresses a close possibility and "might" a distant one there, but also it could be expressing the level of commitment the speaker has to his/her statement.

The creation of space in this type of modality is also about the speakers judgement, and not just possibility.

EG

Thinking: "I've said that I think he may come, but I want to cover my ass and state that he might not."

Milky wrote:

<4. He might come tomorrow. But then again, he might not. Again, we have two sentences, and two possible outcomes of equal probability. So again, each outcome has a probability of 50% But how can that be, if "might" indicates less likelihood than "may"?> Because in each case it is the speakers perception of, stance toward, the likelihood (modal) and not the actual likelihood (real world) itself that is being expressed.

MrPedantic wrote:

. And, in your opinion, are attempts to assign a percentage to the likelihood in these cases (e.g. must = 99% certainty, "may" = 50% certainty, etc.) slightly absurd, however well intended? MrP

Absurd indeed.