I'm not a neuroscientist, but as I understand it losing a particular ability is usually a sign of brain damage. It is likely that she learned English as an adult (or older than a toddler at least) and it was being stored in a more specific and more constrained area of the brain.

i think the point here was that it might be useful to have language redundancy so that brain damage which robs you of one language may leave you with some of your backups

Someone close to me suffered something similar, she lost Spanish but kept the ability to speak Lithuanian (Her native language) shortly before her death, doctor commented that it was something he had seen before

It made treatment complex, because no one knew enough Lithuanian to communicate to the degree necesary

Yes, depending on where the stroke/aneurysm the effect can be totally different. I had one a few years ago, I was still able to chat with my colleagues and the hospital staff in Japanese and English, as well as in French (my native language) with my family back in France. However when looking at texts in French or English, I was recognizing the letters but putting them together I wasn't recognizing the meaning behind them. It took me a few days before being able to decipher them again, but it took me 3 weeks before recovering a decent reading speed. But some blood that had escaped in my skull at that time was blocking some nerves linked to the vision, so I had a dead angle of around 45 degrees on my right side until the blood naturally disappeared.

I think at that point it's a huge advantage but hopefully none of us have to be in that spot.

I believe there was indeed some research done in the past in which multi-lingual people would generally recover better from strokes compared to people knowing only a single language.

Perhaps additional "connections" are made in the brain that can be used when other "connections" are broken due to stroke.

Anecdotic-ally, my father also recovered quite well from his stroke and he knew Dutch, English and likely a bit of latin languages as well (had Italian and Chinese penfriends when he was young, often used Latin proverbs, ...).

Strokes that effect language come in two forms. One of which is you can understand speech but can't generate it (you can respond to verbal requests but your speech comes out as word salad if any). The other is you can generate speech but not parse it (you superficially participate in conversation, but its entirely one sided as your responses have nothing to do with what was just said). I don't know enough to tell you about multi-lingual cases.

Was your mother a late bilingual? From this paper [1], we know that for late bilinguals, regions for each language are spatially separated in Broca's area. Broca's area is responsible for speaking, while language understanding is attributed to Wernicke's area*. It's possible that the stroke affected only the part of the Broca's region responsible for English speaking. And since language understanding is handled by another region, it wasn't affected.

*For completeness, according to current neurolinguistic models it's more complex than that, more brain regions are actually responsible for speech/understanding.

[1] http://www.nature.com/articles/40623

As a matter of fact, yes, late bilingual!