Extracted from Rules: Shaping Behavior and Knowledge, by Ronald G. Ross, 2023, 274 pp, https://www.brsolutions.com/rules-shaping-behavior-and-knowledge-book.html

For groups and communities of people, a rule is a guide for conduct or action. Its expression provides criteria for judging or evaluating that conduct or action. In other words, a rule serves as a criterion for making judgments and decisions. Consider this simple example posted on a sign at the entrance to a restaurant near the beach.

Rule: No shirt, no service.

Using the criterion 'no shirt', a decision can be made (by the restaurant, the customer, or anyone else) about someone dining at that restaurant.

As this example suggests, a rule tends to remove a degree of freedom. (I say 'tends to' because actual results depend on how strictly the rule is enforced.) You can dine in lots of places but, without a shirt, not at that restaurant. A rule always tends to constrain in the sense of narrowing ('ruling out') options or contingencies.

Rules, States, and Actions

Rules and actions are different, fundamentally and unconditionally. If you let the distinction between the two become hazy, your logic is likely to become hazy too.

• By action, I mean performing, doing, or transforming something. In the example above, the act of entering the restaurant as a customer is an action.

The important thing about actions is that they can change the state of things. Before an action, the world is in one state; after an action, the world can be in a different state.

Aside: This is true even for an unsuccessful action — for example, a potential customer being rejected at the restaurant for lack of a shirt. The state of the world has changed because now there is knowledge of at least the unsuccessful action itself. Even if the potential customer didn't get in, they tried.

• Rules, in contrast, are not actions, and not really even about actions, at least not directly. For example, the No Shirt, No Service Rule doesn't imply any action on its own. Instead, it's simply a criterion for whether service can be properly given and received.

Directly, rules are always about the states of things (or more precisely, states of affairs), not actions. The rule above disallows shirtless customers — in other words, it disallows the state of the restaurant in which some shirtless customer is served.

As another example, consider the action Take Rental Car Reservation at a car rental company. If successful, the action will leave the world in a new state. Specifically, a reservation holder will hold a reservation that they did not have before. To be successful, however, the action must leave the world in a state that complies with relevant rules, for example:

In summary, actions do things to try to change the state of the world. Rules, in contrast, simply indicate which states of the world are:

• correct or incorrect
• allowed or disallowed
• legal or illegal
• obligated or prohibited
• necessary or impossible

Applying Rules

As above, actions always try to change states. Rules, in contrast, pertain to the new states that actions seek to create. If the new state is acceptable, there is no need for intervention. If not acceptable, intervention should happen to prevent an undesirable state, ideally in real time. For example, a shirtless person who seeks to dine is turned away at the entrance into the restaurant.

Unfortunately, such timely intervention is not always possible. Consider the Commandment Thou shalt not kill. Unfortunately, murder often cannot be prevented. Victims do die; the rule is sometimes broken.

What is the role of rules in such cases?

• Before the action, the rule sets expectations for behavior. Not murdering is expected of all members of the community.


• After the action, the rule provides the criteria for assessing wrongness. Although the wrong state (a dead victim) could not be prevented, the violation certainly can be detected and should be penalized in appropriate fashion.

Rules, Actions, and Inferences

Inference rules might seem to muddle the distinction between actions and rules. No!

First of all, an inference 'rule' that calls for execution of an action is not a rule at all. Rather, it is an instruction, no matter what some software platform calls it. Consider this example:

Sign Posted on Glass Box Containing an Ax: In case of fire, break glass.

This statement is conditional ("in case of fire") but gives an instruction ("break glass"). It calls for the execution of an action. It is not a rule.

Aside: Maintaining the distinction between rules and actions is critical. Restricting rules to state creates a guidance knowledgebase (rulebook) with clear boundaries within which consistency can be logically verified. Invoking actions can have unknown consequences (side effects). Strict separation of concerns is therefore essential.

So, let's restrict ourselves to inference rules that imply facts from other facts, for example:

Rule: If raining, then streets are wet.

No matter what happens under the covers in some platform, the inferred fact always exists as soon as the conditional fact is true. Nothing 'made' it true; it's simply true that streets are wet if it is raining. Period.

With decision models and decision tables, it may look like there's an action (task) that produces some new outcome, but that's simply a choice some process or application designer has made about when to look at the results of the inferencing. (Many operational business decisions can be optimized by waiting as long 'to look' as possible.) Logically, the inferred facts (inferences) were there all the time — just not looked at.

Expressing Knowledge of States

The language of actions is very different from the language of rules. To specify an action, generally you use verbs as commands. (In natural language, such specification uses verbs in the imperative mood.) For example:

• Do X
• Build X
• Use X
• Sell X
• Buy X
• Perform X
• Read X

These commands can become quite complex and can involve a great many Xs. The Xs, of course, are all labeled, sometimes (but by no means always) with business-friendly labels. The emphasis, however, is on getting things done — that is, on instructions — not on organizing and expressing knowledge about states of the world per se.

Expressing such knowledge also requires many Xs — terms for all the business things of interest. And it requires verbs to relate those things as well — but never in the command form. Instead, knowledge about states of the world is always expressed as facts. Here are some examples, using the same verbs from above:

• People do yoga in the park.
• Beachgoers build sandcastles.
• Carpenters use hammers.
• Supermarkets sell spices.
• Customers buy emergency supplies on credit.
• Orchestras perform concerts.
• Parents read books to children at bedtime.

These are all simple facts; many are far more complex. But use of the verbs in these sentences (and yes, all facts are expressed as sentences) does not impel any action. When we communicate knowledge about states, we use verbs simply to express structural relations between the Xs.

Aside: In natural language, such specification uses verbs in the indicative mood rather than the imperative mood. Using verbs in the indicative mood expresses an action as a statement of fact.

Expressing rules about states goes one small, but important, step farther. A rule doesn't express an actual fact, but rather what should or should not be a fact.

This is a very instinctive thing in natural language. Here are some examples based on several of the facts above, using the auxiliary verb should:

• Supermarkets should sell many spices.


• Customers should not buy emergency supplies on credit.

Voila, rules!

These examples illustrate how closely rules and structured natural language are woven together. For all practical purposes, they are inseparable. Yes, these are simple rules, but expressing more complex ones simply involves referencing more nouns and verbs — that is, more concepts.

Actually, the hard part of expressing any rule lies not so much with its ruleness per se but with expressing all the relevant concepts succinctly. The ability to express facts and rules succinctly is the purpose of a structured business vocabulary based on a concept model.[1]

References

[1] Refer to Business Knowledge Blueprints: Enabling Your Data to Speak the Language of the Business (2^nd ed), by Ronald G. Ross, 2020.

# # #