I knew someone who believed the same thing, or rather that there might be two types of knowledge-strong/soft. I'm much more comfortable saying there's knowledge (with requisite understanding) and true opinions.
If I believe someone upstairs is having sex because I hear a bed post thumping and it turns out I'm right, do I really know they're having sex, or was my opinion/guess coincidentally true? For me, I wouldn't know unless I had direct evidence of my belief--thus garnishing an understanding of that thought token.
There definitely is something to the distinction between knowledge and merely true opinion, but perhaps what is lacking in your example is not understanding but a sufficiently strong justification. You do have some justification for your belief that someone upstairs is having sex, but maybe this justification is not strong enough for your belief to be knowledge. So, perhaps knowledge could be understood as strongly justified, true, belief.
Is it really possible to understand a single belief in isolation? Maybe I'm mistaken, but I tend to think of understanding as involving many beliefs and pertaining to their interrelations.
I'm glad you ask! Understanding is where your justification for you belief is the thing that makes the belief true.
I don't like using "strong" and "weak" justification because "Knowledge is strongly justified truth" seems a bit odd. Understanding takes care of this by attaching a causal connection between the justification and the belief.
I'm not sure I understand how the justification for a belief can be what makes it true. In any case, that is not what I mean by "understanding".
What I have in mind is that "strong"/"weak" operates on the justification, and this is nothing more than the strength of an argument or piece of evidence. Some arguments (and pieces of evidence) are stronger than others, and when you have a strong argument (evidence) for your belief, and that belief is also true, then you have knowledge. At least, that's how I understand the "traditional" view of knowledge.
Let's take a basic example, 2+2=4.
If you memorize the sentence "two plus two equals four" but have no idea what makes it true (despite justifiably believing it via authority), I'd be hesitant to call it knowledge.
To have understanding, you're justification for believing 2+2=4 needs to be the thing that makes it true, i.e. that 2 and 4 are numbers, and that + is a mathematical operator for addition, etc. etc.
But if you just memorize the sentence "two plus two equals four" and you don't understand what it means, then do you really believe it? I agree that this is not knowledge, because belief is missing.
Because 2+2=4 is something that you can realise by reason alone, it is hard to see how you can fail to believe it without understanding it and being justified in the belief.
But isn't it otherwise with empirical knowledge?