Linguistic models deal with statements as they are used to express meanings. While fact models focus on simple operational meanings within a given conceptual commitment (existence of objects, discernability of objects as representatives of certain concepts, characteristics of individual objects and relationships between objects), linguistic models focus on meanings corresponding to the conceptual commitment itself, regardless of the actual objects, such as what are the characteristics of concepts, how concepts are defined in terms of other concepts, and what relationships between concepts are necessary, permissible or obligatory. Thus statements in fact models describe individual views while statements in linguistic models describe viewpoints. Linguistic models focus at the definitions of new meanings, therefore they facilitate unconstrained communications where new meanings are defined on-the-fly. The flexibility of unrestricted linguistic communication comes with a price: the intended meaning needs to be unraveled from the expression, and the more complex the meaning, the more complex is the unraveling process. The key challenge of unrestricted linguistic communication is not so much in parsing a natural language sentence (which can be quite complex) but rather in dealing with complex meanings that are defined on-the-fly by some sentences and are used in the follow-up sentences. For example, we can say: “Remember the nuts-and-bolts illustrations from the previous chapter? Did you notice that there are a total of 35 nuts visible in these figures?” While humans routinely deal with the dynamic updates to the conceptual commitment, it is a barrier for machine-to-machine information exchanges.

Fact models, on the other hand, facilitate constrained communications in which the vocabulary is preselected; the forms of expressing this meaning are preselected, and consequently the range of meanings is bounded. Fact models do not define new meanings on-the-fly. Instead, they simply express meanings from the preselected set defined by the vocabulary. These boundaries (the vocabulary, the form of expressing meanings, and the intended range of meanings) become the contract between the participants of the information exchange. This contract is essential for repeatable information exchanges, which is a two-phase process: First, the scope of communications is bounded and the contract is established, and then the actual exchanges start flowing. Linguistic models are efficient in the first phase, but fact models are used for the second phase. Each fact model is tailored to a specific contract. Although it is tempting to say that a fact model is the contract itself, there are several more aspects of interoperability in addition to the exchanged meanings and the agreed upon form of expression of meaning. These additional aspects include locations, responsible contact points, trust, monetization, quality, and so on. It is not surprising that the fact models are also used for conceptual definition of the information systems.

Figure 1 compares linguistic models to fact models and ontologies. Ontology is a branch of philosophy that deals with questions concerning what entities exist or can be said to exist, and how such entities can be grouped, related within a hierarchy, and subdivided according to similarities and differences. However, recently this word has been commonly used in a more applied sense as a formal representation of the knowledge of a particular domain by a set of concepts and the relationships between those concepts. Linguistic models as well as fact models describe entities and their relationships in the real world, or in any possible world, regardless of whether these entities and relationships exist or could potentially exist. Fact models separate concrete (operational) knowledge of the actual things and relationships between them (known as facts), and the general knowledge of what kinds of things and kinds of relationships can be described (known as the conceptual schema). Linguistic models do not make this distinction, instead they focus on concepts and those relationships that are necessary, permissible or obligatory between concepts regardless the instances of these concepts or the instances of the relationships. This is the reason why linguistic models provide the foundation for describing business rules, i.e. the statements of behavior guidance.

Linguistic models involve a body of meanings and a vocabulary to express meanings, as well as a mechanism to construct statements that can define new meanings based on the initial ones. This mechanism makes linguistic models unbounded compared to fact models.

The Semantics of Business Vocabularies and Business Rules (SBVR) is a publicly available specification from the Object Management Group (OMG) [SBVR]. It was developed by 17 organizations in 7 countries and adopted by OMG in September 2005. It was published as a formal OMG specification in January 2008. The word “business” emphasizes the fact that SBVR focuses on conceptual models that directly represent the operational vocabulary of systems, related to the business mission of the system, as opposed to any implementation views used by the downstream stages of the system life cycle. The terms business model and conceptual model are often used as synonyms [Bridgeland 2009][Halpin 2008][Olive 2007][Ross 2009] and [Ross 2003].

SBVR provides the provides the following means for building linguistic models:

The OMG Assurance Ecosystem uses SBVR to analyze and represent cybersecurity knowledge as formal machine-readable content that can be directly used by assurance tools but at the same time allow human readable statements in structured English as well as other controlled natural languages. Another use of SBVR within the Assurance Ecosystem is to define patterns as fragments of interconnected concepts that can be used as queries and matched against fact models.

The SBVR provides the means to build linguistic models, including propositions, questions, rules, and corresponding conceptual schemes and vocabularies. The SBVR specification consists of six major parts (illustrated at Figure 2):

SBVR captures business concepts and business rules that may be expressed either informally or formally. Business rule expressions are formal only if they are expressed purely in terms of the preselected conceptual schema for the business domain and semantic formulations, including certain logical and mathematical operators, logical quantifiers, and so on. Formal rules can be exchanged as content with other rules-based software tools [Leuf 2006]. Informal rules may be exchanged as uninterpreted comments.

The SBVR specification defines a standard protocol for exchanging vocabularies, rules, and patterns, including an XML Schema [XML].

Let's look at some practical uses of SBVR. First, we will see how SBVR is used to define a simple conceptual scheme based on well-known, informally defined terms. Then we will look at how to formalize propositions using the SBVR vocabulary. Finally, we will show how propositions are used to define new elements of the vocabulary.

SBVR itself is defined as several vocabularies. The noun concepts of the SBVR vocabulary for defining elementary meanings are illustrated in Figure 3.

The following fragment illustrates the original definition of SBVR in SBVR:

The concept “characteristic” is an abstraction of some property intrinsic to each instance of the concept; the collection of essential characteristics makes up the concept. Two definitions can describe the same object by producing the same set of incorporated characteristics. Noun concepts were informally introduced in Chapter 9. Unique characteristics of the concept allow discerning the objects of this concept.

Some relationships between the SBVR concepts are illustrated at Figure 4.

The role is an abstraction of a thing involved in an instance of the verb concept. The verb concept with the designation ‘verb concepthasrole’ identifies the set of roles for another verb concept (a fact type is a synonym for a verb concept) the inverse relation is described by the verb concept ‘fact type roleis infact type’.

Roles are related to the characteristics of the concept (an object type, to be precise). The verb concept ‘Roleranges overobject type’ identifies the corresponding characteristic. For example, the role ‘company’ of the verb concept ‘company employs person’ ranges over the object type ‘company’.

The verb concept ‘Concept₁specializesconcept_2’ is the key to developing intentional definitions. This verb concept is defined by the following statement: ‘Concept₁ incorporates each characteristic that is incorporated by the concept₂ plus at least one differentiator.’ A synonymous form for this verb concept has the designator ‘Concept₂generalizesconcept_1.’

For example, the noun concept ‘whole number’ specializes the noun concept ‘integer,’ the differentiator being that whole numbers are nonnegative. As another example, the individual concept ‘Cyber Bricks’ specializes the concept ‘company,’ the differentiator being that Cyber Bricks is one particular fictitious company used in the case study for this book.

Semantic integrations often involve recognizing where different concepts (having different intentional definitions) have the same extensions in all possible worlds. Also, it is possible that concepts employing different methods of conceptualization have the same extension in all cases. For example, the individual concept defined as ‘the owner of the Clicks2Bricks system’ is coextensive with an object type defined as ‘the companies featured in the case study of this book.’ The two companies have the same extension (which includes only ‘Cyber Bricks’), but they are different concepts. The corresponding relation between concepts is represented using the designator ‘Concept1 is coextensive withconcept2.’

Propositions and Questions are also meanings. Proposition is defined as meaning that is true or false. A proposition corresponds to a state of affairs in a possible world defined by a collection of things of interest and possibly a time frame. The same proposition can be true in one possible world and false in another. Note that the word “proposition” has two common meanings: first, a statement that affirms or denies something, and second, the meaning of such a statement. The SBVR concept ‘proposition’ is defined in the second sense and should not be confused with the statement of a proposition.

A Fact in SBVR is defined as a proposition that is taken as true. How one ascertains what is true, whether by assertion, observation, or other means, is outside of the scope of SBVR. However, ascertaining the validity of certain propositions (the assurance claims) is the central concern of system assurance. Note that SBVR is a linguistic model, so the definition of a ‘fact’ is slightly different from the one we introduced in Chapter 9. However, the two definitions are in alignment.

The following characteristics of propositions defined in SBVR are important for assurance:

Question is defined as the meaning of an interrogatory. Note that the word “question” has two common meanings: first, a written or spoken expression of inquiry, and second, the meaning of such an inquiry. By the second definition, a single question can be asked in two languages. But by the first definition, using two languages results in two expressions, and therefore, two questions. The SBVR concept “question” is defined in the second sense (meaning) and should not be confused with the expression or representation of a question.

The following definitions tie together the three major areas of SBVR:

The noun concepts of the SBVR vocabulary for defining representations are illustrated in Figure 5.

A Designation in SBVR is a representation of a concept by a sign that denotes it. (This may include a specific namespace that includes the designation to allow disambiguation of terminology when integrating statements from different communities.) Designation involves a signifier, which is the expression of the designation. For example, the concept ‘nut’ has a signifier, which is a string produced by concatenating the characters “n,” “u,” and “t.” The signifier exists only in the context of a designation, so the string “nut” on its own may be interpreted in many diverse ways.

A Definition in SBVR is a representation of a concept by a descriptive statement (expression), which serves to differentiate it from related concepts. A fully formal definition uses one of the semantic formulations, and its meaning is interpreted unambiguously.

Figure 6 illustrates the organization of the designators for verb concepts.

A Fact type form in SBVR is a representation of some verb concept by a template of expression. For example, the fact type form ‘product of company’ demonstrates a designation ‘of’ and has two placeholders—one using the designation ‘product’ at the starting position 1 and the other using the designation ‘company’ at the starting position 12.

Sentential form (also known as a fact type reading) is a fact type form that is a template that can be used for starting a proposition based on a fact type.

Noun form is a fact type form that acts as a noun rather than forming a proposition. A noun form can have a placeholder for each role of a fact type, in which case the noun form result comes from the role the first placeholder serves. A noun form can have one less placeholder than there are roles, in which case the noun form result comes from the role that no placeholder is for. An example is ‘bolt of the nut-and-bolt assembly’ for the fact type ‘nut-and-bolt assembly has bolt.’ This noun form refers to the bolt. Another example of a noun form is ‘screwing nut’ for the fact type ‘nut is screwed onto bolt.’ This noun form yields the screwing act, which is an action. Gerunds are used in noun forms like this for actions, events, and states. They are used in sentences like this: ‘A nut must be screwed onto bolt after fastening the washer.”

SBVR distinguishes between the following three kinds of designation:

Expression is defined as something that expresses or communicates, but is independent of its meaning (such as a sequence of characters, a sequence of sounds, a diagram, or an XML file). A Signifier is an expression that is a linguistic unit or pattern, such as a succession of speech sounds, written symbols, or gestures, used in a designation of a concept. Two particularly important kinds of expression are Text and Uniform Resource Identifiers (URI).

Earlier in this chapter we discussed the differences between linguistic models and fact models. SBVR provides the means for building linguistic models (including the definition of a conceptual schema for the fact model and stating what is necessary, permissible or obligatory in the world described by the conceptual schema). SBVR does not focus on statements of existence of the objects and relationships between the objects. The concept ‘individual concept’ is a point of alignment between fact models and linguistic models. However, the SBVR vocabulary includes terminology for describing the referents of concepts because this terminology is required to build statements and rules that define the semantics of other SBVR concepts. The linguistic perspective on the fact models defined in Chapter 9 is illustrated in Figure 7.

Extension is the totality of objects to which a concept corresponds. SBVR defines a concept ‘Thing’ (based on the ISO 1087-1 term ‘object’), as anything perceivable or conceivable. Every other concept in SBVR implicitly specializes the concept thing.

The concept ‘State of affairs’ is key to understanding verb concepts and their meaning. A state of affairs is any event, activity, situation, or circumstance. A state of affairs can be possible or impossible. Some of the possible ones are actualities (they occur in the actual world). A state of affairs is what is denoted by a proposition. A state of affairs either occurs or does not occur, whereas the proposition is either true or false. A state of affairs is not a meaning. It is a thing that exists and can be an instance of a concept, even if it does not happen. Conceptual commitment in the form of a vocabulary determines our viewpoint on what state of affairs we can consider.

A reference scheme defines the chosen way of identifying instances of a given concept. Understanding reference schemes is critical for integrating facts from multiple sources during information exchanges. A reference scheme is a way of referring to instances of a concept through related things that are either lexical or are otherwise identifiable. A reference scheme usually uses one or more fact type roles of binary types in order to identify an instance of a concept from facts about the instance. A reference scheme can also use one or more characteristics. A reference scheme can be partial or complete. It is complete if it can always be used to refer to every instance of a concept. An overall complete reference scheme for a concept can result from there being multiple partial reference schemes for that concept, its more general concepts, and its categories. For example, the concept nut involves a characteristic id that is used to identify instances of this concept, for example ‘nut n-2’. Alternatively, a particular nut instance can be identified using its relationship to some other concept—for example, ‘the nut of the nut-and-bolt assembly that involvesboltb-52.’

The semantics of the reference scheme is illustrated in Figure 8.

Communications between humans generally use natural language to express propositions, although some technical propositions are at times expressed in a formal language, such as XML, C, or Prolog, or even graphically, such as a UML model. SBVR provides a means for describing the structure of the meaning of the proposition expressed in the natural language that business people use. SBVR calls this “semantic formulation.” Semantic formulations are not expressions or statements. They are structures that make up meaning. SBVR provides a vocabulary for describing them. By using SBVR, the meaning of a definition or statement is communicated as facts about the semantic formulation of the meaning, not as a restatement of the meaning in a formal language.

Semantic formulations contribute to SBVR Structured English through defining all keywords, as follows:

Semantic formulations make distinctions between expression, representation, meaning, and extension. Note how phrase ‘session hijacking is a threat to a web applications’ is interpreted differently in each of the following examples.

SBVR provides a kind of formulation for nominalization of propositions, which makes a proposition to be an object of discourse as in (3). SBVR also supports nominalization of questions and interrogatives (“what,” “whether,” etc.). A separate kind of formulation accomplishes objectification, which makes the actuality or state of affairs that corresponds with a proposition to be an object of discourse as in (4) above. Number (2) involves a quoted statement, so its formulation identifies the statement as a combination of its text and its meaning (using nominalization).