Traditionally, transactions depend upon a certain level of trust that the agreement in question can, and will, be performed. Typically, this trust is achieved by relying on a centralised intermediary such as a bank (to execute payment instructions and maintain accurate records) or an official registry (to confirm ownership of title). One problem with this model is that it introduces a single point of failure (e.g. through negligence or fraud) and can be costly, inefficient and exclusionary.
Permissionless blockchains (Principle 2) address the challenges of a centralised trust model by enabling parties to rely on (or trust in) the rules of the system itself, rather than on a third party. It does this by creating an append-only ledger that is highly resistant to fraud. Instead of a single database, which can be tampered with, multiple (but synchronised) copies of the ledger are distributed across the whole network, making manipulation extremely difficult and highly visible.
In addition, the way in which data is added to the ledger also generates system-level trust. Before data is added (as part of a data block), it must be verified by the network in accordance with a pre-determined consensus mechanism (see Principle 4). If verified, the new block is then cryptographically linked to the block before it (see Principle 5). Combined, this verification and linking of blocks, renders the blockchain practically immutable, enabling participants to trust in the ledger data.