I wish more linguists were language teachers.
In the last couple of years I've done some tutoring in Latin. Part of the program in teaching Latin usually includes a specific focus on what English words have descended from the vocabulary taught in each lesson. I discovered a small pet peeve on my part when I noticed that the kids' teacher often gave them words that, while related to the Latin word, did not come from them. E.g.: English night from Latin nox, noctis. Yes, we have very many words that come from Latin (let's see, so far I've already used: tutoring, usually, includes, specific, focus, descended, vocabulary, discovered, part, noticed, related). But we do speak a Germanic language, after all, and lots of our words go back from modern English to Old English to Proto-Germanic to Indo-European. (I'm skipping steps here, of course, but the route is clear nonetheless, and doesn't pass through Latin.)
So if Indo-European is the parent language, then the languages that descended from her dialects into their own separate languages are the daughter languages: Sanskrit, Latin, Greek, Persian, Slavic, Germanic, etc. That makes Latin a sister of (proto-)Germanic. She may be an elder sister, but a sister nonetheless. So if an Indo-European root yields words in both Latin and Germanic, then we say they both come from IE, not that the Germanic word comes from Latin, or vice-versa. Let's trace a few random words to get this distinction down.
Night. Not from Latin. There is a direct line from Modern to Middle to Old English (niht, neaht) to proto-Germanic naht, which also yields nahts in old Gothic, Nacht in German, natt in Norwegian, etc. Parallel to the Germanic naht (sisters, again) would be Latin nox, noct- and Greek nyx, nyxt-, as both descend from the Indo-European root nekw-t-.
Cry. From Latin. This word has an interesting etymology, recently posted at Language Hat. It is one of many words that came into Middle English through Old French and Latin. It doesn't appear in any form in Old English, and therefore doesn't come from Proto-Germanic.
Picture. From Latin. Pictus, past participle of pingere, plus the -ura suffix, came directly into Middle English. (The pingere form, having morphed to peindre in old French, finds itself with a new past participle form--peint--that also comes into English as paint.)
At. Not from Latin. The Latin preposition ad is related, but as a sister (not a mother) to the Germanic at. (Ado also comes from Germanic at, but through Norse.)
I'm not sure why it bothers me that people attribute to Latin words that came from Germanic; it's not like the English language has its feelings hurt, and either way kids are learning that languages are connected. And yet, Latin is so obviously important, it doesn't need help from false attribution; whereas I always enjoy pointing out the Germanic character still strong in our language (since it is the most Romancified of the Germanic family). Mostly I guess I just like it when people are precise, and while I know too much about how languages change to expect precision from the average speaker, I would like teachers to be able to make the distinction, or know enough to look it up.