If we are going to think lucidly about this topic, we need a more restrictive definition than that presented in the pilot study. From the early guest talks and researcher conversations, it’s already quite clear that treating money systems, religions, and building codes “as protocols” expands the potential range of “protocol” so wide as to be meaningless. Many of the topics under discussion are better thought about in the terms of existing literature on the topic, or in the “naive” industry understandings.

I’d like to present an alternative definition that is largely informed by institutional theory and by technical protocols such as email, TCP/IP, and cryptocurrency mining. Then I’ll explain why I think this theory is better suited to intellectual inquiry than the current approach.

Protocols are a subset of institutions as defined by economist Douglass North. North’s definition of institutions sees them as regulatory bodies which define a field of agency, or in his terms, the choice set.

Institutions are the humanly devised constraints that structure political, economic and social interaction. Institutions consist of both informal constraints (sanctions, taboos, customs, traditions, and codes of conduct), and formal rules (constitutions, laws, property rights). Together with the standard constraints of economics they define the choice set and therefore determine transaction and production costs and hence the profitability and feasibility of engaging in economic activity.

Specifically, protocols are a type of formal rule system, although norms may form around their use.

  1. Protocols define a minimized scope of action that constitutes participation. My idea of “minimized scope of action” can be helpfully compared with legal institutions. Law governing human and corporate persons assumes an unknowable scope of agency, and expressly prohibits only certain behaviors. I think of law as “permissive” in this sense. Protocols are “restrictive” or “minimized” in that only specifically formulated actions or inputs are considered valid participation.
  2. Protocols have a formal specification that is self-consciously adopted by participants. The formal spec is what defines the form of valid participation, and it has to be known before it can be adopted. Lawrence Lessig distinguishes between post facto and ex ante behavioral regulation. Law and norms are examples of the former, where protocols are an example of the latter. Participation in the protocol requires shaping one’s action “to spec” in order to conform. The behavioral regulation takes place before or as action takes place.
  3. Protocols and tolerate a wide range of contentful behavior within the action scope. A valid action may still carry a range of contents. Think of how TCP/IP sends packets which may contain any sort of data. The pilot study is getting at this property when it talks about emergent behavior. This is also where norms may emerge; a protocol can support more types of action or input than are normatively common.
  4. Protocols are participatory and non-coercive. Network effects may still emerge which motivate adoption, but participation is always optional.
  5. Conformance to protocols confers a benefit on participants.
  6. Protocol participants are those who change their behavior to match the scope of defined behavior. This criteria helps focus us on the actual users of protocols, rather than those who take part in downstream effects. Email users are not protocol consumers; email clients are.

In my mind there are several benefits to this definition.

Firstly, it helps us avoid definitional conflicts with well-established research areas. Religious codes, legal systems, norms and traditions, codes of practice, rites and rituals, architectural schema, organizational behavior, and other formal and informal rule systems have been extensively theorized. Technical protocols are a specifically different institutional regime than these other systems; they are the newest invention of the bunch, and they have not been adequately treated. This is the task at hand. If we continue in the current unprincipled way, we are bound to “discover” features of these other rule systems which have already been discussed in existing literature. Attributing these rediscovered properties to protocols instead of their proper institutional domain will only confuse us more.

Secondly, per Kei’s point, a parsimonious definition also helps us focus on where protocols are already in place or new protocols can be useful. By looking at already-existing protocols, we can evaluate how they perform against other other institutional systems, and understand their unique affordances. (Treating other rule domains as protocols, which I’ve been advocating against, is not the same as discovering protocols which are already in use in those domains). At the same time, we can think about protocols as a design tool for creating a mutually beneficial regime of action. Where could protocolization be helpful in areas where other institutions have failed? We can’t answer these questions unless we exclude other rulemaking systems from our definition.