The word typical is, I think, ill-chosen here. What the writers are saying at 122.2 might be better expressed thus;
We often use can to signify that an eventuality does occur—not always, not necessarily even usually, but from time to time.
At 339.10 there are really two different uses which should be distinguished:
May is used, particularly in academic and scientific contexts, to signify that an outcome is possible but not necessary:
Children of divorced parents may have difficulty with relationships.
May is also used in academic and scientific contexts to describe a range within which possible outcomes lie:
A female crocodile may lay 30-40 eggs.
The flowers may have five or six petals, pink or red in colour.
Either of these may uses can be expressed with can in colloquial registers. Colloquial can has largely (but not entirely) replaced may.
And (as you doubtless know) these do not exhaust the possible uses of either may or can.
An example of "general past possibility" would be:
"In Roman Britain, you could get from what's now Canterbury to what's now St Albans by travelling Watling Street."
As opposed to specific past possibility:
"I could have bought it for half the price yesterday!"
It is common usage, even if incorrect according to some grammars, to use the same syntax as the general possibility for a specific one:
"Yesterday I could buy it for half the price!"
We're actually seeing different senses of could here. In the present (or future) tense, could indicates a sort of hypothetical, where other languages might use a subjunctive form, perhaps. It is used to make suggestions or raise possibilities.
"What shall we do today?"
"We could take the train up to the lakes and do some walking."
It doesn't communicate a strong desire, though tone may indicate some desire. It can also indicate a possibility that one might need to think about.
"Where is she? She said to meet in the library at 3pm!"
"She could be at the other entrance."
Now, when we want to put a hypothetical in the past, we apply that same sense of could to the present perfect:
"Why am I slogging through all this coursework? I could have taken a gap year and be sunning myself on a beach in Thailand."
Used that way, it is expressing a hypothetical about things having been different. It is known not to have happened, but it could have happened.
"Where is he? He said to meet in the library at 3pm!"
"He could have gotten sidetracked. You know what he's like."
That is an unknown hypothetical in the past. You don't know whether it happened or not.
But when we use could in sentence like "in Victorian England, people could die of diseases that we can easily treat today", it's not the same sense of could as above. Let's talk about can.
Can is a modal verb that indicates possibility, rather than a hypothetical. It means be able to, and sometimes you have to resort to be able to because can doesn't exist or work in all tenses. So,
"I can touch my toes!"
But
"I will be able to pay for next week."
But what if you can't touch your toes now, but you were able to in the past? Well, then it becomes:
"I could touch my toes last week..."
Here, could is a past form of can. It means, roughly, "was able to". It says nothing about whether you did or not, merely that had that ability.
The two senses of could are closely linked, no doubt. It probably tells someone more educated in philology than I a great deal about how our language developed from its forebears. But now these two sense are quite distinct - if potentially confusing.
Best Answer
This is an interesting question because logically, it seems like "can" wouldn't necessarily break any grammar rules, but it sure sounds wrong. Willow Rex's comment on the question seems to have the right ring to it. I suspect that there is some context in which "can" would sound right. But the issue seems related to tense. This is my theory:
"Can" is about possibilities. There is a "trick question" used to teach principles of probability. Say you ask someone the question, "I just flipped a coin; what is the probability that it came up heads?" Most people would say 50%. But the answer is actually that it is either 0% or 100%. Probabilities are about the possibility of future events, not things that have already happened and the actual results are final.
That same principle applies to the phone call. A specific person has placed the call and that is the only person it can be, there are no other possibilities. We just don't know who it is. We will discover who it is when we answer the phone (a future event). That also parallels how we think about who it is. We don't think in terms of who theoretically originated the call, we think in terms of who will be at the other end when we pick up.
So knowing who it is is a future condition. "Can" is present tense and "could" is future tense, which makes "could" the appropriate word.