TL;DR
A derivation path is the route from your seed to one particular key. For a Native SegWit multisig the standard route is m/48’/0’/0’/2′. If you recover with a different path, you get different keys and an empty-looking wallet.
Why paths exist
One seed can produce billions of keys, organised like folders in a tree. The path says which folder to use. Different wallet types use different folders, so they don’t get in each other’s way.
Reading m/48’/0’/0’/2′
| Part | Meaning |
|---|---|
| m | The master key, from your seed (and passphrase) |
| 48′ | Purpose: multisig, following the BIP48 standard |
| 0′ | Coin: bitcoin (1′ would be testnet) |
| 0′ | Account: your first multisig. Use 1′ for a second vault |
| 2′ | Script type: Native SegWit (P2WSH), the modern address type starting bc1q. 1′ is the older wrapped type |
The apostrophe (sometimes written as h) means “hardened”: keys below that point can’t be linked to keys elsewhere in the tree from public information alone. Below this path, wallets use /0/* for receiving addresses and /1/* for change.
Why it matters for recovery
Your descriptor records the path for each key. If you rebuild a wallet by hand, use exactly the same path; a single wrong digit gives a valid wallet with no coins in it and no error message.
Two vaults, one set of devices
Changing only the account number (0′ to 1′) gives you a completely separate multisig from the same three devices. Build and test them one at a time, as described in step 6.