How does a light-weight consumer through SPV guarantee that the requested merkle path will not be faux?

I do know that the consumer has its personal copy of the block headers, so it may well belief the merkle root from there. It then queries a full node for the merkle path, calculates the merkle root from that, and eventually compares the reference merkle root with the calculated one. If each are the identical, the consumer will be positive the transaction was included, even when it doesn’t belief the total node, since there isn’t any means for the total node to engineer the identical merkle root from completely different knowledge. To this point so good.

Nevertheless, how can it belief the obtained path in case each merkle roots don’t match? The complete node may ship a faux path, which then results in a distinct merkle root upon verification, which let’s the consumer assume the transaction was not included when in actuality it may need been included.

Can somebody please make clear?
Thanks 🙂