Defining Blockchain

In the famous parable, several blind people run their hands over different parts of an elephant, each reaching a different conclusion about the nature of the object. There is indeed an elephant in the room. Yet no one has the capacity to perceive it fully.

That is how I think about blockchain. There is a there there; an important new phenomenon. Somehow, though, we can only comprehend pieces of it at a time. One podcast I listen to starts every interview by asking the subject to define blockchain in their own terms. Why is it so hard? Some technical knowledge is required to appreciate how a blockchain operates, and that knowledge is spread across computer science, cryptography, and game theory. Even more than that, however, understanding blockchain requires a conceptual leap which can be difficult to make. And the different parts of the blockchain elephant—like cryptocurrency trading and record-keeping for enterprise consortia—seem far removed from one another.

I've spent years describing blockchain to students, academic colleagues, friends, businesspeople, policy-makers, regulators, and others. No framing works for everyone. The longer and more detailed the definition, the more listeners get lost. A definition isn't an explanation; it's the most compact possible summary that aids understanding.

I've started offering multiple levels of very short definitions. Pick the ones that make sense to you, and work for the applications you're concerned with.

The first is just four words:

A chain of blocks

Yes, this should be obvious. It's in the term itself. But it's actually surprisingly deep. What's the significance of chaining blocks? It's twofold: aggregation and serialization. Blockchains clump transactions together into blocks, and validate each clump. Anything not in a clump isn't recorded. Aggregation means that blockchains aren't one-way streams of information; they are call-and-response systems. Participants in the network must agree on what's in a block, even though no one is in charge. The blocks, however, aren't independent; they are linked cryptographically into a sequence. A blockchain is an unfolding story whose history cannot be changed.

Want an even shorter definition? Try this:

Consensus by design

Blockchains are systems designed to reach consensus. Many structures are designed for consensus; think standards bodies, for example. A blockchain is a community that must achieve consensus, or it falls apart. The magic of the technology is that it reliably produces such a consensus. Imagine that you had to convince a roomful of people from all around the world the come to consensus about something, knowing that some might try to cheat the group, express wildly divergent views, or just be difficult. Just try creating an incentive system that won't be gamed, for example, if there's money at stake! We normally solve these kinds of problems by limiting the room to relatively small, relatively homogenous groups that know and trust each other, or by giving someone the power to impose a solution if the group can't agree. Blockchain removes those limitations.

If that's too cryptic, let's expand to five words:

A database no one owns

A blockchain is a just a database; it's a repository for information. But it's a unique kind of database. It lives in the network, replicated across many participating computers. That's not distinctive today; most databases run on cloud platforms that stretch beyond the boundaries of one location. The difference is that someone owns or controls those cloud platforms, and the databases themselves. Ownership means the power of exclusivity. If it so desires, only the owner can make changes to the database, or dispose of its contents. You don't own your data on Facebook, or the status of your corporate stock on the New York Stock Exchange.

That's not always bad. Sometimes, though, giving up ownership and control is a deal-breaker. If you're a pharmaceutical company, an ocean shipping carrier, or an insurance company, you're not going to trust a competitor or even a supposedly neutral third party with ownership of your transactional data. You'll always believe your records over theirs. Which means you'll always want your own records, even when everyone in the network shares transactions. There are plenty of other white spaces where ownership either grants unreasonable power to an intermediary (think big tech platforms) or prevents efficient data sharing. No country can own a transaction between two sovereign currencies, for example.

Blockchain is the idea that we can have most of the benefits of a shared database, without the limitations of someone having to own it.

If you give me six words, I'd expand that to:

Sharing data without giving up control

There are so many activities that require data sharing, whether that's your age, your medical records, your credit history, the telemetry records from a self-driving car, or the contents of a shipping container on a freighter in the middle of the ocean. Digital information is infinitely replicable, easily combined with other data, and the source of valuable insights when analyzed. Sharing it immediately puts you at risk to whomever you're sharing with. As a result, individuals tend to share too much data (because they don't have a choice) and businesses share too little (because they do). Most people are familiar with the privacy, security, equity, manipulation, and other harms to individuals. The business losses from activities that don't happen, or happen inefficiently, may be even more significant. That's where blockchain comes in: It breaks the link between sharing and giving up control.

Finally, in seven words, I would describe blockchain in the language of my book:

A new architecture of trust and trustworthiness

Trust is itself a difficult concept to get your arms around. Even though, in contrast to blockchain, most people think they know what it means. It's one of the keys to understanding blockchain, because the whole point of consensus (see above) is for the ledger to be trusted. You'll often hear blockchain technology described as "trustless" because it doesn't require trust in a third party, such as an intermediary. That's a misnomer, as I explain at length in the book. Societies and communities without trust are less successful and terrible for their members. We need ways of trusting without the dependency that normally comes with it. Blockchain decouples trusting your bank balance from trusting your bank. It creates a new form of digital trustworthiness, which isn't the same as trust. To change the world, we need systems that are trusted, but also systems worthy of that trust.

The Blind Leading the Blind

I'm under no illusions these definitions will be the last word. They certainly won't help you understand the mechanics of blockchains; you just have to take on faith they can do the things the definitions claim. There are definitions that more directly address the what questions, by talking about cryptography, data structures, and consensus algorithms, but that only gets in the way of answering why questions. If there is more understanding about what what blockchain might do, it won't guarantee success, but it will give us a way to evaluate success.

We may never see the whole elephant clearly; it's enough to make it dance.

This article is subject to refinement. Does it help you understand blockchain better? I welcome your comments and feedback!