Institutional Grammar 2.0 Codebook
IInstitutional Grammar 2.0 Codebook
Christopher K. Frantz & Saba N. SiddikiVersion: 1.0 (15th August 2020)
Contents1 Introduction 22 Syntactic Definitions, Institutional Statement Concepts, and Assumptions aboutInstitutional Statement and Syntactic Component Nesting 3 a r X i v : . [ c s . M A ] A ug Introduction
The following instructions are intended to provide guidance for the coding of institutional statements, thefocal unit of analysis in the Institutional Grammar (IG), according to the Institutional Grammar 2.0 (IG2.0) specification. An institutional statement describes expected actions for actors within the presenceor absence of particular constraints, or parameterizes features of an institutional system. Institutionalstatements convey information that contextualizes their applicability. They vary in prescriptiveness andforce, as reflected by the presence of information that more or less strongly compels behavior and bythe presence of information that specifies payoffs for compliance, or non-compliance, with statementsinstructions. Varying in the inclusion of these various kinds of information, institutional statementstypically take two functional forms: constitutive and regulative. Constitutive statements constitutefeatures of a system (e.g., actor positions and roles, processes, venues, etc). Regulative statementsdescribe actions linked to specific actors within certain contextual parameters.According to the IG 2.0, institutional statements are commonly comprised of a set of syntacticcomponents, with individual components associating with unique information, and which combine toconvey a statement’s institutional meaning. Regulative statements are composed of some or all of thefollowing components with the corresponding syntactic labels: (i) an Actor, referred to as an Attribute;(ii) action associated with actor, referred to as an Aim; (iii) action context, referred to as Context; (iv)a receiver of action, referred to as an Object; (v) a prescriptive operator that describes how strongly anaction is compelled or restrained, referred to as a Deontic; and (vi) an incentive linked to action, referredto as an Or else. Constitutive statements are composed of some or all of the following components withthe corresponding syntactic labels: (i) the entity that is being constituted within a statement, referredto as a Constituted Entity; (ii) an action that constitutes the Constituted Entity, called the ConstitutiveFunction; (iii) the constitution context, referred to as Context; (iv) properties that serve as input to theConstitutive Function, called Constituting Properties; (iv) A prescriptive operator that defines to whatextent the action of an institutional statement is compelled, restrained, or discretionary, referred to asa Deontic; and (vi) an incentive linked to action, referred to as an Or else. The operational definitionof an institutional statement is tied to the presence of certain syntactic components, or necessary In addition to Crawford and Ostrom (1995, 2005), the specification draws on the original IG codebook (Brady et al.,2018), and integrates further specific refinements (Siddiki et al., 2011; Frantz et al., 2013).
IG 2.0 Codebook Version: 1.0 2undamental Concepts components. To qualify as a complete regulative institutional statement, the statement must at leastcontain Attribute, Aim, and Context components. The Object, Deontic, and Or else componentsare deemed sufficient components. Institutional statements – regulative or constitutive statements –containing only respective necessary components are referred to as atomic statements.In this codebook, in accordance with the IG 2.0 specification, coding guidance is offered for theencoding of regulative and constitutive institutional statements along the aforementioned syntacticcomponents at three levels of expressiveness: (1) IG Core; (2) IG Extended; and (3) IG Logico. Thecore definitions of syntactic components remain the same across levels of expressiveness. However,the level of granularity with which components are parsed differs across levels. Section 2 provideselaborated definitions of syntactic components that generalize across levels of expressiveness. This isfollowed by a terminological overview of concepts used in IG 2.0. Section 3 provides an overview ofthe pre-coding, or pre-processing steps relevant for document preparation prior to coding. Section 4specifies the syntactic conventions used in this document, followed by the coding guidelines by syntacticcomponent, and by level of expressiveness. Section 5 provides a concluding overview of taxonomies asreferenced throughout the coding guidelines.
IG 2.0 is premised on a set of syntactic definitions, conceptualizations of institutional statements,and assumptions regarding institutional statement and syntactic component-level nesting. While thesedefinitions, conceptualizations, and assumptions generalize across levels of IG encoding, how they arecaptured depends on at which level the encoder is working. In this section, we will thus lay out thesefoundational syntactic definitions, concepts, and assumptions. We start by offering complete definitionsof IG components more generally, and then move to defining key institutional statement concepts, andhighlighting assumptions regarding institutional statement and component level nesting. Concludingthis section, we organize the coding levels based on involved features, along with providing principalguidelines on the coding process.
The IG structure as referred to in this document relies on the elementary syntactic components ofregulative and constitutive statements highlighted in Tables 1 and 2. Definitions of these componentsthat hold across IG 2.0 encoding levels are provided along with each syntactic component. Compo-nents listed in both tables are ordered by their assumed necessity and sufficiency within institutionalstatements. Some components exist within both regulative and constitutive statements, and thus theirdefinitions are presented in relation to each, however with additional information in accordance withdifferent types of statements.
IG 2.0 rests on the basic definition of institutional statements – linguistic constraints or opportunities.The construction of complex institutional statements as put forth in IG 2.0 builds on this basic definitionto include the notion of nested institutional statements of different forms introduced in the following.Motivating the conception of nested institutional statements is the pragmatic observation that state-ments are fundamentally linked in institutions constituted of multiple institutional statements. Lack ofspecificity of these linkages undermines the coder’s ability to comprehensively, and accurately, captureinstitutional content.IG 2.0 accommodates two forms of institutional statement nesting: horizontal nesting and verticalnesting . Generally, horizontal nesting allows for the representation of multiple institutional statements
IG 2.0 Codebook Version: 1.0 3undamental Concepts
SyntacticElement Definition
Attribute An actor (individual or corporate) that carries out, or is expected to/tonot carry out, the action (i.e., Aim) of the statement. The Attributemay also contain descriptors of the actor.Object The inanimate or animate part of an institutional statement that is thereceiver of the action captured in the Aim. Objects can be of director indirect nature. Indirect objects are objects that are affected ortargeted by the application of the Aim to direct objects. Objects canboth be real-world entities, or abstract ones (e.g., beliefs, concepts).Deontic A prescriptive operator that defines to what extent the action of aninstitutional statement is compelled, restrained, or discretionary.Aim The goal or action of the statement assigned to the statement At-tribute.Context The context component instantiates settings in which the focal actionof a statement applies, or qualifies the action indicated in an insti-tutional statement. The former type of Context is referred to as an“Activation Condition.” The latter type of Context is referred to asan “Execution Constraint.” Both can occur in a given institutionalstatement, including multiples of either type. Where no explicit Acti-vation Condition is specified, the context clause is by default “underall conditions”. Where no explicit Execution Constraints are specified,the context clause is by default “no constraints”.It is important to note that
Context in institutional statements reflectsthe context specific to the coded statement (Statement Context), asopposed to capturing context in the wider sense, making reference tothe context of the policy or the domain more generally.Or else A sanctioning provision associated with the action indicated in a partic-ular institutional statement that can exist wholly within an institutionalstatement, or be represented in a nested institutional statement (as de-fined in the following discussion).Table 1: Definitions of Syntactic Elements for Regulative Statementsthat convey co-occurring or alternative actions. Generally, vertical nesting allows for the representationof multiple institutional statements that convey coupled actions that follow from one another in the formof a consequential relationship. It is particularly suited to representing the case of consequentially linkedstatements in which statement A delineates permitted, required, or forbidden activity, and statementB delineates sanctions for non-conformance with statement A. In the IG 2.0 parlance, statement A isconsidered a “ monitored statement ,” and statement B a “ consequential statement .” Horizontalnesting and vertical nesting are described in more detail below.
Horizontal nesting : Horizontal nesting describes a logical combination of two or more statementsto capture institutional content comprehensively. Exemplified in narrative form, a horizontally nestedstatement can combine the two statements, such as “Organic farmers must commit to organic farmingstandards” AND “Organic farmers must accommodate regular reviews of their practices”, but allow formore complex constructs, such as
IG 2.0 Codebook Version: 1.0 4undamental Concepts
SyntacticElement Definition
ConstitutingProperties Delineate properties assigned to a Constitutive Entity.Deontic A prescriptive operator that defines to what extent the action of aninstitutional statement is compelled, restrained, or discretionary.ConstitutiveFunction A verb that constitutes a Constituted Entity,or reflects the functionalrelationship between constituted entity and constituting properties.ConstitutedEntity The entity being constituted, reconstituted, modified or otherwise di-rectly affected within an institutional statement.Context The context instantiates settings in which the focal action of a state-ment applies, or qualifies the action indicated in an institutional state-ment. The former type of Context is referred to as an “ActivationCondition.” The latter type of Context is referred to as an “ExecutionConstraint.” Both can occur in a given institutional statement, includ-ing multiples of either type. Where no explicit Activation Condition isspecified, the context clause is by default “under all conditions”. Whereno explicit Execution Constraints are specified, the context clause is bydefault “no constraints”.It is important to note that
Context in institutional statements reflectsthe context specific to the coded statement (Statement Context), asopposed to capturing context in the wider sense, making reference tothe context of the policy or the domain more generally.Or else A sanctioning provision associated with the action indicated in a partic-ular institutional statement that can exist wholly within an institutionalstatement, or be represented in a nested institutional statement (as de-fined in the following discussion).Table 2: Definitions of Syntactic Elements for Constitutive Statements (“Organic farmers must commit to organic farming standards” AND “Organic farmers must accom-modate regular reviews of their practices”) XOR (“Organic farmers must NOT sell their produce underthe organic farming label”).
Note the use of parentheses to signal the precedence of individual statements. Possible logical oper-ators are
AND (conjunction), OR (inclusive disjunction; colloquially: AND/OR ), or
XOR (exclusivedisjunction; colloquially: EITHER/OR). Where negation is involved, those can be combined with theoperator
NOT (as highlighted in the previous example). An alternative equivalent representation is (“Organic farmers must commit to organic farming standards”
AND “Organic farmers must accom-modate regular reviews of their practices”)
XOR NOT (“Organic farmers must sell their produce underthe organic farming label”).
Vertical nesting : Vertical nesting describes a relationship of two or more statements, in which theleading statement (monitored statement) describes an action that is regulated by a second statementnested in the Or else component (consequential statement). The second statement reflects a con-sequence of violating the instructions captured in the monitored statement. Consequences generallyinvolve some pay-off for non-compliance or compliance respectively. Exemplifying vertical nesting in
IG 2.0 Codebook Version: 1.0 5undamental Concepts narrative form, we can write “Organic farmers must comply with organic farming regulations”,
OR ELSE “Certifiers must revokethe organic farming certification”.
Note that both forms of nesting can be combined, i.e., monitored and consequential statements canembed horizontal nesting. Extending the previous example, we can state (“Organic farmers must comply with organic farming regulations”
AND “Organic farmers must accommodate regular review of their practices”),
OR ELSE (“Certifiers must suspend the organic farming certification”
XOR “Certifiers must revoke the organic farming certification”).
Note the use of parentheses to signal precedence of the respective statements. Vertical nesting canoccur across an arbitrary number of levels (i.e., a consequential statement may be a monitored statementin deeper levels of nesting). Exemplifying multi-level nesting (visually supported by correspondingformatting), we can state (“Organic farmers must comply with organic farming regulations”
AND “Organic farmers must accommodate regular review of their practices”),
OR ELSE (“Certifiers must suspend the organic farming certification”
XOR “Certifiers must revoke the organic farming certification”),
OR ELSE “USDA may revoke certifier’s accreditation”.
The combination of both nesting approaches affords the representation of complex institutional ar-rangements, both in terms of institutional content (horizontal nesting) and enforcement (vertical nest-ing). The principles are schematically highlighted in Figure 1.Figure 1: Nesting Principles
In addition to the nesting of statements, IG 2.0 further assumes the possibility of nesting of individualcomponents as introduced in the following.
Component-level nesting : Specific components may be substituted with entire institutional state-ments that characterize an individual with respect to its Attributes, or to express procedural order, i.e.,statements whose fulfilment is a precondition for the application of a given statement. Substitutingthe conditions component in the IG syntax, we can exemplify this with “Organic farmers may sell theirproduce under the organic label under the condition that organic farmers apply for certification.” , wherethe statement embedded in braces reflects the precondition and may in itself be expressed in (a subset
IG 2.0 Codebook Version: 1.0 6undamental Concepts of) syntactic components of institutional statements. The use case for such component-level nestingfurther includes the articulation of beliefs about individuals’ behaviours (e.g., an official sanctioning anindividual if the official believes that the individual has performed a violation). : In addition to the nesting concepts, advanced coding relieson decomposing actors and objects into core descriptors and associated properties. For this purpose,we rely on the conceptual representation of an
Attribute/Object-Property Hierarchy as exemplified inFigure 2. In this visualization, statements such as “. . . a written notification of proposed suspension orrevocation of certification . . . ” reflect an involved Object hierarchy centered around the “notification” ,that has a property “written” . Looking at the context of the notification we recognize the conceptof “certification” that has the property of being “suspended” or “revoked” , expressed as dependentObjects ( “suspension” , “revocation” ), whereas the latter concepts themselves have a shared propertyof being “proposed” in the first place. However, while the property “written” functionally dependson the “notification” , that is, writtenness alone does not make sense with an Object it refers to, theexistence of a certification does not rely on the notification (i.e., it is functionally independent), andhas a self-contained property hierarchy (suspended, revoked, proposed) as described above.Interpreting complex Object specifications with this decomposition hierarchy in mind affords a uni-form coding approach. The structure of this hierarchy for the specific statement is shown in Figure 2(the dashed line signals relationships between functionally independent Objects). Note specifically thepotential use of logical operators ( “XOR” ), as well as the ability to reflect shared properties ( “pro-posed” ), to disambiguate the logical relationship between the identified properties, an aspect that isimplicit in the textual representation.Figure 2: Attribute/Object-Property Hierarchy for given ExampleCapturing all potential decomposition approaches, we can apply this scheme to functionally dependentproperties ( “written” in the previous example), functionally independent Objects or properties ( “cer-IG 2.0 Codebook Version: 1.0 7undamental Conceptstification” in the previous example), and furthermore afford the substitution of Objects by completeinstitutional statements. Since the latter aspect relies on richer contextualization, it will be exemplifiedin the context of the coding instructions. The stylized general form of the Attributes/Object-PropertyHierarchy is shown in Figure 3. Note that logical operators apply to both functionally dependent andindependent properties and on any level of decomposition.Figure 3: Attribute/Object-Property Hierarchy
A concept central to the coding with IG 2.0 is the action situation (Crawford and Ostrom, 2005).Basically, an action situation is defined as an institutionally governed setting in which two or moreactors interact, in relation to which specific outcomes emerge. Action situations are governed by aconfiguration of seven types of institutional statements, which can be regulative or constitutive inkind, with distinctive functional properties. These seven types of institutional statements, labeled inparentheses in terms of different types of “rules”, convey: positions that actors can occupy withinan action situation (position rules), eligibility criteria for occupying those positions (boundary rules),operational actions linked to actors occupying certain positions (choice rules), situational outcomes(scope rules), channels of information flow (information rules), guidance on collective decision making(aggregation rules), and incentives tied to particular actions (pay-off rules). Each action situation canbe governed by multiple statements of a particular type. Action situations, and key action situationcomponents, are schematically visualized in Figure 4. In the widest sense, action situations describethe context in which institutional statements operate, and in the context of regulative statements,specifically the mapping between actors, actions, outcomes and the associated payoffs.Again, the rule type taxonomy associated with the action situation concept links to the IG insofar aswhole institutional statements can be classified accordingly, depending on their functional properties.
IG 2.0 Codebook Version: 1.0 8undamental Concepts
Figure 4: Schematic Visualization of the Action Situation (as per Crawford and Ostrom (2005))The IG 2.0 specification leverages the action situation concept in recognizing generally that institu-tional statements characterize activity occurring in action situations, and in accounting through syntacticclassification for ways that institutional statement information corresponding to the Context componentcontextualizes intra- and inter-statement activities.Operationalized in the context of regulative statements, Context clauses can instantiate an actionsituation in which an Attribute acts on Objects in a particular manner, and which are governed by someconfiguration of institutional statements. By way of contrast, the Context clauses of other statementsmay simply constrain an Attribute’s behavior in some way within a given action situation. As notedin Section 2.1, context clauses which serve an instantiation function, as well as Attribute or Objectchanges are referred to as Activation Conditions. Context clauses which qualify action are referred toas Execution Constraints. Figure 5 schematically represents how Activation Conditions and ExecutionConstraints situate relative to action situations.Figure 5: Activation Conditions and Execution Constraints PrinciplesNaturally, explicit characterization of how institutional statements relate to action situations neces-sitates understanding of the institutional domain. Without this, the coder may encounter difficulty indetermining as to whether a specific Context descriptor refers to the action situation more generally, or
IG 2.0 Codebook Version: 1.0 9undamental Concepts the action specifically in the form of an action property. While we offer further elaboration as part ofthe coding guidelines in Section 4, we provide a brief example to motivate the distinction at this stage.Inherent to the Activation Condition is reference to a set of exogenous variables; exogenous in thesense that it references states or actions that are beyond the actions that can be qualified within certainExecution Constraints in an instantiated environment (e.g., a new action situation, an environment inwhich Attributes change or take on new roles, or an environment in which Attributes act in an alteredway upon Objects). In other words, Activation Conditions precede the regulated action and activate agiven institutional statement in the first place. Conversely, Execution Constraints describe constraintson actions once enacted (and implicitly on actors and associated pre-/proscription as visualized inFigure 6). Figure 6: Context Relationships in the Action SituationProcedurally, this implies different semantics for Activation Conditions and Execution Constraints.Activation Conditions represent Context external to the action situation an institutional statement isembedded in that activates the non-conditional part of an institutional statement, and possibly leadingto the activation or modification of an action situation. Execution Constraints, in contrast, are directlyattached to the institutional statement and thus reflect context embedded within the action situationitself. The discussed distinction is summarized in Figure 6.Offered here is more operational guidance on the differentiation highlighted above, which starts withthe identification of Context clauses in an institutional statement. To systematize the differentiation, wefirstly provide a terminological basis. Linguistically, context clauses are generally modifiers, specificallyqualifiers ( “usually” , “some” , “annually” ), adverbial clauses ( “When the traffic light turns from red togreen, . . . ” ) and prepositional clauses ( “after midnight” ). Whereas qualifiers reliably signal ExecutionConstraints (action properties in the narrow sense), and adverbial clauses generally indicate ActivationConditions, depending on contextual interpretation, prepositional clauses can fall in either category andselectively signal Activation Conditions or reflect Execution Constraints (action properties in the widersense).The differentiated treatment for prepositional clauses is best described with an example. The examplestatement used here is regulative: “At 8am, farmers may begin selling their goods in the farmer’smarket,” contains two context clauses ( at 8am, in the farmer’s market ), one of which is a conditionsclause ( at 8am ) and one of which is a constraints clause ( in the farmer’s market ). Context clauses maybe implicit, and institutional statements are not constrained in the number of clauses for condition andconstraint type. The remainder of the statement is the non-context clause of an institutional statement.Figure 7 highlights this decomposition of regulative institutional statements with respect to the Contextcomponent.Decision heuristics can be employed to aid in the identification of activation conditions and executionconstraints. The following heuristics are particularly designed to help the analyst determine if a contextclause in question is an activation condition, leaving the resultant classification of the clause as anexecution constraint, if it is determined that it is not. Offered first is a heuristic that generalizes across IG 2.0 Codebook Version: 1.0 10undamental Concepts
Figure 7: Context Clauses in Institutional Statementsregulative and constitutive statements. This is followed by heuristics specific to two different types ofstatements.
General Heuristic for Identifying Activation Conditions:
The clause instantiates a discrete setting (constrained temporally, spatially, or otherwise as shownbelow) and/or event that activates the non-condition clauses of the institutional statement (i.e., non-context clauses along with potential constraint clauses) as a whole. The following example statementscontain activation conditions (underlined) for illustration.: • “Upon receiving final notice of non-compliance, farmers shall cease sale of any product bearing theUSDA organic farming label.” This statement signals the instantiation of a novel attribute-objectlink described by the activity, and is positioned within a discrete temporal setting. • “Starting January 1, the Department of Agriculture is the certifying authority.” Here the state-ment makes explicit reference to an event that leads to the activation of an associated role changein a constitutive statement. • “Upon entry into the house, visitors must remove shoes.” (Event) vs. “At home individuals mustnot wear shoes.” (Discretized setting). Whereas the first statement references a specific event(entry), the second statement describes a general discretized setting in which the statement holdsat all times, i.e., the statement is activated at any time. Heuristics for Identifying Activation Conditions in Regulative Statements:
Attributes : The clause instantiates a) a change in attributes linked to a statement’s activity or b) achange in attribute role. • “Between the hours of 6pm and 6am on Mondays, members of neighborhood watch residing inblocks 7-10 will assume night patrol activities.” This example signals a change in attribute rolewithin a specified time frame.
Objects : The clause instantiates a change of the object(s) linked to the statement’s activity. • “Starting Dec. 15th, inspectors must exclusively use the revised inspection form.” (novel objectuse) IG 2.0 Codebook Version: 1.0 11undamental Concepts
To support the classification more generally, we can further offer practical considerations to aid inthe decision making: • Regulative statements: More generally, in a regulative context with a given specification ofpre/proscriptions, activation conditions constitute a discretized setting in which institutional con-tent can in principle be adhered to or violated. • Context-clause interdependencies: If the application of the clause of concern is contingent onthe prior activation of another context clause, the former is an execution constraint, whereas thelatter describes an activation condition in the context of the analyzed institutional statement. Ifonly the satisfaction of both clauses leads to the activation of the non-context clause, both aresensibly identified as activation conditions. – Example: “When live fish or viable gametes are sold, traded, taken or otherwise disposedof from an aquaculture facility, the permittee or operator shall, at the time of transferof possession, give an invoice to the person receiving such fish or viable gametes.”
Here “when live fish . . . facility” represents the activation condition; the subsequent provision ofan invoice “at the time of transfer of possession” is an execution constraint.Returning to the initial example “At 8am, farmers may begin selling their goods in the farmer’smarket.” , the condition clause “At 8am” signals the instantiation of a discrete temporal setting in whichthe remaining statement is activated as a whole (i.e., permitting the sales of goods on the market).The clause “in the farmer’s market” complements the description of the regulated content and neitheraffects attribute/role, object nor does it define a setting that activates the remaining statement. Theresulting statement would thus be coded as follows:
Coded statement: “At 8am (Activation statement), farmers (Attribute) may (Deontic) begin selling(Aim) their goods (Object) at the farmer’s market (Execution constraint).”
To highlight the distinction between activation conditions and execution constraints more clearly, wecan review the following statement: “Farmers may sell non-organic goods in the organic farmer’s marketonly between the hours of 3 and 5pm.”
Here the time frame “between the hours of 3 and 5pm” signals a distinctively different relationshipbetween attributes (farmers), aim and object, whereas the “in the farmer’s market” complements thecharacterization of the institutional setting in which the permission holds.This is in contrast to the following statement: “Farmers must perform inventory of goods sold at farmers market daily.” (execution constraint)
This statement signals a general obligation to provide inventory information (i.e., is activated at alltimes), but does not establish a specific discretized setting or event that triggers the obligation.Contrasting this, the following example highlights such event, leading to the characterization of theconditional clause as activation condition: “At the close of market each day (activation condition), farmers must perform inventory of goods sold.”
Heuristics for Identifying Activation Conditions in Constitutive Statements: • Entity : The clause instantiates a change in the Entity that is being constituted. – Example: “In the event that the Board Chair position becomes vacant, the Vice-Chair is the chiefexecutive of the Council.” (change in entity specification under event)
IG 2.0 Codebook Version: 1.0 12undamental ConceptsProperties : The clause instantiates a change in the constituting properties of the entity that isconstituted, reconstituted or otherwise affected in the institutional statement. – Example: “Starting Dec. 15th, organic farming is agricultural production that does not involve the useof synthetic chemicals or genetically modified organisms.” (change in constituting properties of constituted entity)
IG 2.0 Codebook Version: 1.0 13.6 IG Coding Levels
The IG 2.0 identifies three levels of encoding to provide flexible accommodation of coding necessitiesbased on the complexity of encoded data, as well as the analytical objectives of the coder:
IG Core , IG Extended , and
IG Logico . IG Core : IG Core, facilitates coding of a basic syntactic structure. This level best accommodatesrelatively simple institutional statements that largely follow basic regulative or constitutive structure,along with analytical objectives that involve the statistical assessment of references to the individualcomponents (e.g., distribution of actor, action, object or deontic references).
IG Extended : The next higher level, IG Extended, focuses on capturing the syntactic structureof institutional statements in greater detail. For regulative statements, this involves the fine-granularencoding of actors and objects, along with complex property relationships. Furthermore, it enablesfor both regulative and constitutive statements,a detailed encoding of context, such as the character-ization of statement dependencies, and categorization based on circumstantial aspects of conditionsand constraints (e.g., temporal, spatial, procedural aspects). Choosing to encode on this level may bemotivated by the complexity of the encoded institution regulation (e.g., complex statements involvingAttribute/Object-Property Hierarchies (Section 2.4), or extensive statement interdependencies), butalso by the analytical objectives, such as the operationalization of the extracted structure in advancedcomputational models that require the explicit representation of actor properties and context charac-terization.
IG Logico : The highest level of expressiveness, IG Logico, aims at enabling the analyst to derivemore sophisticated understanding of semantic relationships embedded in and among institutional state-ments based on institutional statement classification across syntactic categories; for example, improvedunderstanding of actor roles, explicit references between statements, as well as inference of actor obli-gations tacitly expressed in the coded document. As a point of contrast, whereas at the IG Core andIG Extended levels, syntactic classification of institutional statements is a final goal of the encodingexercise, at the IG Logico level the goal is to build on syntactic classification by leveraging this codingtoward identification of institutional semantics that relay functional and/or relational information ofinterest to the institutional analyst.
Shared Assumptions :All coding levels are backward-compatible, i.e., statements coded at higher levels of expressivenesscan be reduced to any lower level of expressiveness. In other words, statement information correspond-ing to different syntactic components is simply more finely decomposed as one moves from lower tohigher levels of expressiveness (e.g., IG Core to Extended), with the effect that moving in the otherdirection, the coder can simply collapse decomposed information. Methodologically, this accommodatesa multi-pass approach towards coding; where coding could: commence at the lowest level, before be-ing incrementally refined to accommodate syntactic parsing associated with the respective next higherlevel(s) of expressiveness, and conclude at the desired level of expressiveness set out in the researchobjectives (which is informed by the nature of the coded document and analytical objectives as discussedabove).
Mapping Prerequisites:
The different coding levels make varying use of the concepts highlighted in Section 2 as outlined inTable 3. Concepts specified at the IG Core level, apply to IG Extended and IG Logico, and conceptsthat apply to IG Extended apply to IG Logico. Statement level nesting applies at all levels. Given themulti-pass coding approach, concepts specified for respective lower levels apply to all higher levels (e.g,statement-level nesting applies to all levels).A high-level overview of the individual levels, along with the discussed objectives is captured inFigure 8. The principles and objectives of the individual codings are discussed in detail in Section 4.
IG 2.0 Codebook Version: 1.0 14.6 IG Coding Levels
Coding Level Relevant Concepts
IG Core • Horizontal and vertical nesting (Statement-level nesting) • Activation conditions, execution constraintsIG Extended • Component-level nesting • Attributes/Object-Property Hierarchy in regulative statements;equally applies to Constituted Entities/Properties in constitutivestatements (both of which may have properties on their own) • Structural Decomposition Patterns (Constitutive Statements);discussed in Section 4.4Table 3: Relevant concepts on different coding levelsFigure 8: Overview of IG Coding Levels and associated objectives
Before discussing the coding of institutional statements in detail, we in this section we lay out “pre-coding” steps that relate to familiarization with the institutional setting and document preparation,commencing with general pre-processing, followed by considerations specific to distinctive levels ofexpressiveness.
IG 2.0 Codebook Version: 1.0 15.1 General Steps
1. Familiarization with institutional setting: Prior to embarking on any coding, the institutional an-alyst should carefully review the institution to be coded. A thorough pre-coding review (e.g.,reading) of the institution to be coded is necessary for gaining a high-level understanding of in-stitutional actors, actions, and institutional statement relationships that can be leveraged in theencoding process.2. Selection of coding platform: One of the first steps the institutional analyst should engage in asshe gains familiarity with the institutional setting is identifying the coding platform in which insti-tutional data will be stored. The selection of a coding platform will be informed by the analyst’sexpectations regarding at which level of expressiveness institutional statements will be encoded,and related assessment of institutional complexity, as certain platforms are better equipped tocapture institutional statement complexity. The selection of a coding platform will also be in-formed by the analyst’s anticipated usage of institutional data; for example, whether stored datawill later be engaged in computational applications. Platforms store data in forms that are moreor less computer readable.3. Initial organization of institutional information: Once the institutional analyst reviews the insti-tution to be coded as part of step 1, she can start to organize its contents. Though variableacross jurisdictional setting, policy content is typically organized according to (i) a preamble, thatdescribes the motivation for the policy; (ii) key definitions, that provide descriptions for actors(e.g., “Secretary’ means the Secretary of Agriculture” ) and other terms (e.g., “‘Prohibited sub-stances’ are substances that have been banned by the Dept. of Agriculture for use in organic foodproduction” ) and abbreviations (e.g., “‘NOSB’ means National Organic Standards Board” ) thataid in the effective interpretation of policy content; and (ii) Policy instructions organized by topicaccording to section and subsection headers.All three types of content (preambles, definitions, and policy instructions) should be coded.Preambles are likely to be comprised of self-referential statements that can convey the purposeof, or contextualize, the institution under examination more generally as well as regulative and/orconstitutive statements. In most cases, definitions are constitutive, and can be useful for encodinginstitutional statements encountered in a policy document; e.g., when some statement clausereferences something that is defined in the definitions section of the policy document. Thisis computationally useful [in the coding process] because it allows the computer to referenceparticular definitions when certain terms inline. It also allows the computer to link statementsthat share common definitional information in the analysis process.The critical aspect of this coding step is to ensure that the institutional analyst identifies allrelevant, codeable information – i.e., all information that is comprised of codeable institutionalstatements.4. Verification and pre-processing of institutional statements: Following the identification of candi-date statements in step 3, the analyst should engage in verification and pre-processing of institu-tional statements to enable their syntactic decomposition in step 5. Verification in this case meansascertaining that candidate statements accord with defining syntactic and semantic features ofregulative and constitutive statements. Principally, this means verifying that statements presumedto be regulative in kind at least contain an Attribute, Aim, and Context, and that statementspresumed constitutive in kind at least contain a Constituted Entity, Constitutive Function, andContext component. Institutional statements often do not align with sentences encountered informal institutions, as a result of writing style (e.g., compound sentences) and punctuation (e.g.,bulleted lists). Examples of excerpts of formal institutions that do and do not accord with the
IG 2.0 Codebook Version: 1.0 16.1 General Steps definition of regulative and constitutive statements are provided below. Importantly, text thatdoes not classify as institutional statements should be retained and annotated as domain specificbackground. This information can be useful for institutional interpretation and implication.Institutional Statements:
Organic farming is hereby established as a practice regulated under the Department of Agriculture.The Department of Agriculture shall promulgate regulations governing the practice of organicfarming.
NOT Institutional Statements:
Organic farming promotes environmental and human health.The Department of Agriculture is committed to marketing of agricultural products.
Pre-processing in this step also means organizing the content of institutional statements to bothremove extraneous content from statements (i.e., punctuation that accompanies statements oftenreflecting institutional style or organization; for example, roman numerals, bullet points) as wellas to begin to arrange statement content to offer additional clarity about how statements are tobe coded in step 5. Provided below is text that has been pre-processed to remove extraneouspunctuation.Unprocessed Excerpt “(a) The producer of an organic livestock operation must establish and maintain year-roundlivestock living conditions which accommodate the health and natural behavior of animals,including:(1) Year-round access for all animals to the outdoors, shade, shelter, exercise areas, freshair, clean water for drinking, and direct sunlight, suitable to the species, its stage of life, theclimate, and the environment: Except, that, animals may be temporarily denied access tothe outdoors in accordance with §§ Pre-processed Text for Institutional Statement Delineation and Punctuation Removal
The producer of an organic livestock operation must establish and maintain year-roundlivestock living conditions which accommodate the health and natural behavior of animals,including: Year-round access for all animals to the outdoors, shade, shelter, exercise areas,fresh air, clean water for drinking, and direct sunlight, suitable to the species, its stage oflife, the climate, and the environment. Except, that, animals may be temporarily deniedaccess to the outdoors in accordance with §§
5. Decomposition of institutional statements: Following the verification and preprocessing of insti-tutional statements, the analyst should commence the syntax-based encoding process accordingto a selected level of expressiveness.Pre-processing is an optional step in the encoding process but can significantly reduce the time andcognitive load associated with subsequent decomposition. Further, the analyst can choose degreesof pre-processing. More extensive pre-processing is particularly useful for encoding at higher levelsof expressiveness. Below are pre-processing guidelines that are useful for encoding at any level of
IG 2.0 Codebook Version: 1.0 17.2 Pre-processing Guidelines for IG Core expressiveness, as well as guidelines that are level specific. The level specific guidelines build upon eachother, rather than being exclusive, meaning that guidelines applicable for IG Core are relevant for IGExtended and IG Logico, and IG Extended guidelines are applicable for IG Logico.Generally, pre-processing, particularly of a more extensive kind, will be easier for analysts with greaterfamiliarity with the IG, as they will likely be able to detect statement structure, components, andrelations without engaging in even a preliminary decomposition of statements. Some degree of formaldecomposition might be required of analysts less familiar with the IG to be able to discern these.
General pre-processing guidelines: • Data cleaning: dealing with extraneous punctuation, fixing typos • Delineation of text into institutional statements • Delineation of nested statements (e.g., Or else statements) • Preliminary classification of institutional statements as regulative, constitutive, regulative or else,constitutive or-else • Preliminary organization of statements by identifiers capturing the institutional structure/orderingof institutional statements in the document. These identifiers can uniquely identify institutionalstatements, and statement linkages (i.e., nested statements), as well as policy sections or partsthat can facilitate understanding of statement context and cross-statement or policy references. • To accommodate encoding of institutional statements at the core level, preliminary decompositionof institutional statements to account for multiple values within individual syntactic fields shouldbe entertained during the preprocessing of institutional documents. Institutional statements oftencontain multiple Attributes, Aims, and/or Objects. For example: “The producer of an organic livestock operation must establish and maintain year-round livestockliving conditions which accommodate the health and natural behavior of animals.”
This statement contains multiple Aims, “establish” and “maintain.”
Neither of these Aims isassociated with unique values in other syntactic fields, and therefore, they can both be capturedwithin a single institutional statement. However, downstream coding is facilitated by capturingmultiple values individually within separate statements. With the recommended decomposition,the statement above is reflected as two: “The producer of an organic livestock operation must establish year-round livestock living condi-tions which accommodate the health and natural behavior of animals.” and “The producer of an organic livestock operation must maintain year-round livestock living condi-tions which accommodate the health and natural behavior of animals.” • In preprocessing institutional statements for coding at the IG Core level, the analyst may considerreformulating statements into active form (where statements are originally captured in passiveform while being careful to retain statement meaning from an institutional perspective). Note,that conversion of statements from passive to active form typically required some implication ofvalues according with different syntactic fields. For example the passive statement: “Notificationsof compliance must be sent to farmers within 30 days of facility inspections” converts to “[Certifier]IG 2.0 Codebook Version: 1.0 18.3 Pre-processing Guidelines for IG Extendedmust send farmers notifications of compliance within 30 days of facility inspection,” promptingthe implication of “Certifier” as the relevant Attribute, or actor in charge of performing action.This implication requires understanding of institutional context obtained through step 1, as wellpotentially of the identification of a convention for notating implied information, such as forexample, the use of brackets ([ ]) in the example included here. • Where actions or actors are implied, those are inferred from the context and additionally specifiedas part of the coding in terms of the institutional statement structure. While found across awide range of statements, this is commonly necessary in the context of statement combinations(combination of two actions performed by the same actor). The same applies to implied logicalrelationships (AND, OR, XOR). • When facing complex sentence structures, statements should be thought of in terms of sequentiallyapplied actions. For example, if statements report outcomes of actions without making referenceto such actions, the coder should reconstruct the action sequence leading to such outcome interms of institutional statements (see Section 2.1). • At this stage, the analyst should flag additional semantic information that she wishes to capturein the syntactic decomposition of statements and associated label.
Additional pre-processing of institutional statements to accommodate their downstream coding at the IGExtended level involves some preliminary characterization of institutional statement linkages, particularlyto capture action sequences. The coding in IG Extended affords richer decomposition of institutionalstatements into action sequences. Composite actions are often represented as exemplified in the fol-lowing: “When an inspection of an accredited certifying agent by the Program Manager reveals anynoncompliance with the Act or regulations in this part, a written notification of noncompliance shallbe sent to the certifying agent.” , where “When an inspection of an accredited certifying agent by theProgram Manager reveals any noncompliance with the Act or regulations in this part” represents aconditional clause that does not overtly reflect an institutional statement due to the expression of ac-tions in terms of nouns (conceptual reification). From an institutional semantic perspective, this clausecaptures two linked action statements, namely the fact that a “Program Manager inspects accreditedcertifying agents” and that the “Program Manager reveals non-compliance in this process” . Retainingthe essential institutional semantics the original statement can thus be rewritten as (with inference ofimplied components) “When Program Manager inspects accredited certifying agents and [the ProgramManager] reveals non-compliance in this process, the Program Manager shall send a written notificationof noncompliance to the certifying agent.”
While possible to identify as part of the coding process, aspecific consideration of IG Extended is to identify such action sequences, and potentially be offloadedto the pre-processing process, subject to the analytical objective, nature of the coded policy, as well ascoder background and algorithmic treatment of reconstruction.
Additional pre-processing of institutional statements to accommodate their downstream coding at theIG Logico level involves some more extensive, albeit still preliminary, capturing of inter-statement re-lationships and embedded actions within institutional statements that the analyst may want to fullyreconstruct in terms of institutional statements during the encoding process. Inter-statement rela-tionships are often indicated with referential clauses that embed within institutional statements. Inpolicy documents, these references are often to statement collections, the coded document at large, orthird-party documents. The example statements below include types of referential clauses that embed
IG 2.0 Codebook Version: 1.0 19.4 Pre-processing Guidelines for IG Logico within statements that the institutional analyst may wish to capture during the pre-processing phase.Embedded actions can generally be thought of as actions ancillary to that represented in the focalaction of an institutional statement (reflected in the Aim or Constitutive Function for regulative andconstitutive statements, respectively). Embedded actions, while referenced, are incompletely described.However, reconstruction of these embedded actions into complete institutional statements can afford amore complete depiction and understanding of the institutional domain being evaluated. The particularreconstruction the analyst pursues will depend on her analytical objectives, but also the specific typesof institutional functions she wants to capture within explicit and reconstructed statements. In the pre-processing phase the analyst might consider constructing a dictionary of terms they observe throughpreliminary review of institutional statements that signal different institutional functions. This promptsconsideration of how different types of observed actions might link to different institutional functionsof interest to the analyst. The example statements below include embedded actions that can be fullyreconstructed during the encoding process.Example Statements with Referential Clauses
Any operation that: (1) Knowingly sells or labels a product as organic, except in accordancewith the Act, shall be subject to a civil penalty of not more than 3.91(b)(1)(xxxvii) of thistitle per violation.A production or handling operation that sells agricultural products as “organic” but whosegross agricultural income from organic sales totals $5,000 or less annually is exempt fromcertification under subpart E of this part.Any agricultural product that is sold, labeled, or represented as “100 percent organic,”“organic,” or “made with organic (specified ingredients or food group(s))” must be: (a)Produced in accordance with the requirements specified in § §§ §§ Example Statements with Embedded Actions
A handler of organic products may use information provided by the certified operation todetermine percentage of organic ingredients. → Embedded action: information provided by the certified operation (i.e., provision ofinformation by certified operation)
A certifying agent must provide an applicant with a copy of the on-site inspection report,as approved by the certifying agent, for any on-site inspection performed. → Embedded action: approved by the certifying agent (i.e., approval of report by certifyingagent)
A certifying agent whose accreditation is suspended by the Secretary under this section mayat any time submit a request for reinstatement of its accreditation. → Embedded action: accreditation is suspended by the Secretary (i.e., accreditation sus-pension by Secretary)
The Program Manager may initiate suspension proceedings against a certified operation,when a certifying agent fails to take appropriate action to enforce the Act. → Embedded action: certifying agent fails to take appropriate action to enforce the Act(i.e., failure to act by certifying agent). In this case, the failure to act is also signallinga violation, or non-compliance of some kind, which could be marked as an institutionalfunction of interest and even used in downstream coding toward the reconstruction of bothdirect statements and their logical inverses.
IG 2.0 Codebook Version: 1.0 20.1 Coding Syntax
Provided next, following a brief overview of conventions we rely on for syntactic notations, are thespecific guidelines for encoding institutional statements at the IG Core, IG Extended, and IG Logicolevels.
In this section, we provide guidelines for coding institutional statements at the IG Core, IG Extended, andIG Logico Levels of Expressiveness. Following the specification of utilized syntax, we specify encodingprinciples for regulative and constitutive statements.
The syntactic coding for examples in the remainder of this document relies on specific symbols, whosefunction depends on the applied context, i.e., grammar component vs. institutional statement, andrespective coding level (IG Core, IG Extended, IG Logico). An overview of all symbols along withapplication context, minimum level of applicable encoding, description and examples is provided inTable 4.Throughout the remainder of this section we use color coding to signal the association/introduction ofspecific symbols for syntactic components or features with specific levels of expressiveness (as introducedin Section 2.6). Symbols associated with IG Core features are held in blue for regulative statements, andin purple for constitutive statements (specifically relevant from Table 8 onwards). Symbols associatedwith IG Extended are held in green, and features associated with IG Logico are called out in orange.Symbols of general relevance across levels and regulative and constitutive statements (e.g., parenthesesto signal precedence or nesting) are held in bold black . Naturally, the examples draw on features notintroduced to this stage, but offer an illustration of the representations used throughout the subsequentguidelines.Symbol/SymbolPairs CodingContext LowestapplicableCoding Level& StatementType Description Example ( )
Component IG Core,Regulative& Con-stitutiveStatements Component classifica-tion: The characteri-zation of an expressionas a component type issignaled through paren-theses that contain thecomponent type.
Certifier ( A ) . . . , where A identifies the certifier asan attribute in a given in-stitutional statement.Where used, compo-nent annotations (e.g.,animate, inanimate)can be appendedto the componentclassification. Certifier(
A;label=animate ). . . , where A identifiesthe certifier as an attributein a given institutionalstatement, and animate isan additional annotation. IG 2.0 Codebook Version: 1.0 21.1 Coding Syntax
Where used to com-bine individual compo-nents of the same type(in addition to anno-tation or the indica-tion of statement com-binations), parenthesessignal component-levelcombinations.
Attendees must not ( eatand drink ) on the train.[ ] Component IG Core,Regulative& Con-stitutiveStatements Tacit components: Theexplicit specificationof implied components(e.g., actor(s)) issignaled with brackets.
They [farmers ( A )] mustcomply with the certifica-tion regulation . . . , where[farmers ( A )] character-izes the inferred actor. ( ) Statement IG Core,Regulative& Con-stitutiveStatements Horizontally nestedstatements are repre-sented using surround-ing parentheses toemphasise the prece-dence of combinedindividual statements. ( stmt AND stmt ) ; ( stmt AND ( stmt OR stmt )) ,where stmt represents aninstitutional statementcombined with otherinstitutional statementsusing logical operators( AND , OR , XOR , andpotentially
NOT ) – moredetails on logical operatorsbelow.Where used to com-bine individual compo-nents of the same type(in addition to anno-tation or the indica-tion of statement com-binations), parenthesessignal component-levelcombinations.
Attendees must not ( eatand drink ) on the train. [ ] Statement IG Core,Regulative& Con-stitutiveStatements Vertically nested state-ments are representedusing brackets thatembedded the re-spective consequentialstatement stmt1 [ stmt2 ] , where stmt1 represents a mon-itored statement, and stmt2 the correspondingconsequential statement. IG 2.0 Codebook Version: 1.0 22.1 Coding Syntax { }
Component IG Extended,Regulative &ConstitutiveStatements Component-levelnesting is repre-sented by embeddingthe component-substituting nestedinstitutional statementin braces. In the caseof component-levelnesting, the compo-nent type specificationfollows the embeddednested statement.
Certifier ( A ) believes( I ) { farmer ( A ) violates( I ) code of conduct( B ) } ( C ex ) In this example, the exe-cution constraint ( C ex ) ofa given institutional state-ment is substituted withanother institutional state-ment. A Component IG Core,RegulativeStatements Identifies the precedingexpression as Attributecomponent.[2]
Certifier ( A ) I Component IG Core,RegulativeStatements Identifies the precedingexpression as aim com-ponent.
Certifier ( A ) monitors ( I )farmers. B dir Component IG Core,RegulativeStatements Identifies the precedingexpression as direct ob-ject component.
Certifier ( A ) administers( I ) certifications ( B dir ). B ind Component IG Core,RegulativeStatements Identifies the precedingexpression as indirectobject component.
Certifier ( A ) registers ( I )certification ( B dir ) for or-ganic farmer ( B ind ). B Component IG Extended,RegulativeStatements Identifies objects thatare neither direct norindirect, but con-tained as functionally-independent objectsin the institutionalstatement (seeAttributes/Object-Properties Hierarchy inSection 2.4). . . . notification ( B dir ) ofsuspension ( B a , B ) or re-vocation ( B a , B ) of certi-fication ( B a ) . . . First-order propertiesare identified by alpha-betic identifiers (e.g., B a , B b , etc.); second-order properties areidentified with numericidentifiers (e.g., B , B , etc.). Here, the functionally in-dependent object certifi-cation ( B a ) is root of aproperty structure consist-ing of two objects as prop-erties (suspension, revoca-tion), both of which areannotated with referenceto the certification. IG 2.0 Codebook Version: 1.0 23.1 Coding Syntax D / D Component IG Core,Regulative& Con-stitutiveStatements Identifies the precedingexpression as deonticcomponent. Regulative:Certifier ( A ) must ( D )monitor ( I ) farmers ( B dir ).Constitutive: From 1st January onwards( C ac ), Council ( E ) shall( D ) be responsible ( F )for adherence with foodproduction standards( C ex ). Alternative example:
From January 1st onward( C ac ), there shall ( D )be ( F ) a National Or-ganic Standards AdvisoryCouncil ( E ) within theDepartment of Agriculture( C ex ). C ac / C ac Component IG Core,Regulative& Con-stitutiveStatements Identifies the precedingexpression as an activa-tion condition compo-nent. Regulative:
Upon accreditation ( C ac )certifier ( A ) must ( D )monitor ( I ) farmers( B dir ). Constitutive:
From 1st January onwards( C ac ), Council ( E ) shall( D ) include ( F ) organicfarming representatives( P ) to review chemicalallowances within organicfood production standards( C ex ). C ex / C ex Component IG Core,Regulative& Con-stitutiveStatements Identifies the precedingexpression as an execu-tion constraint compo-nent. Regulative:
Certifier ( A ) must ( D )monitor ( I ) farmers ( B dir )at any time ( C ex ). Constitutive:
From 1st January onwards( C ac ), Council ( E ) shall( D ) include ( F ) organicfarming representatives( P ) to review adherencewith food productionstandards ( C ex ).IG 2.0 Codebook Version: 1.0 24.1 Coding Syntax E Component IG Core,ConstitutiveStatements Identifies the preced-ing expression as con-stituted entity
From 1st January onwards( C ac ), Council ( E ) shall( D ) include ( F ) organicfarming representatives( P ) to review chemicalallowances within organicfood production standards( C ex ). P Component IG Core,ConstitutiveStatements Identifies the precedingexpression as constitut-ing property
From 1st January onwards( C ac ), Council ( E ) shall( D ) include ( F ) organicfarming representatives( P ) to review chemicalallowances within organicfood production standards( C ex ). F Component IG Core,ConstitutiveStatements Identifies the precedingexpression as constitu-tive function
From 1st January on-wards ( C ac ), Council ( E )shall ( D ) include ( F ) or-ganic farming representa-tives ( P ) to review al-lowances within organicfood production standards( C ex ).IG 2.0 Codebook Version: 1.0 25.1 Coding Syntax prop / prop Attributes,Object,Entity andPropertycompo-nents IG Core,Regulative& Con-stitutiveStatements Identifies propertiesof attributes and ob-jects respective. The prop symbol is usedin conjunction withthe respective com-ponent identifier. Ifcoding on IG Extendedlevel, where multipleproperties for a givencomponent exist, theyreceive a numeric indexsuffix. IG Core:Regulative:
Certified organic ( A , prop )farmers ( A ) must ( D )respond ( I ) to formal( B dir , prop ) certificationrequirements ( B dir ). Constitutive:
The Council ( E ) consistsof ( F ) elected ( P , prop )officials ( P ) resident inthe electorate ( P , prop ). IG Extended:Regulative:
Certified ( A , prop1 ) or-ganic ( A , prop2 ) farmers( A ) must ( D ) respond ( I )to formal ( B dir , prop1 )certification requirements( B dir ). Constitutive:
The Council ( E ) consistsof ( F ) elected ( P , prop1 )officials ( P ) resident inthe electorate ( P , prop2 ). IG Extended furthersupports the explicitencoding of object andproperty hierarchies.Where multiple levels ofobject/properties existin the property hierar-chy, those are contextu-alized with the object-s/properties they referto (i.e., they are ap-pended to the com-ponent specification).Further details on prop-erty coding are pro-vided in Table 6 and il-lustrated below.
IG 2.0 Codebook Version: 1.0 26.1 Coding Syntax prop / prop(ctd.) The example on theright highlights a com-plex object hierarchystructure previously dis-cussed in the context ofthe Attributes/Object-Property Hierarchy(Section 2.4). Asmentioned above,where multiple objectson a given hierarchylevel exist, they areuniquely identified witha numeric index (e.g., B , B , etc.). Wheremultiple propertieson a given hierarchylevel exist, they areuniquely identified witha numeric index (e.g., prop1 , prop2 , etc.). . . . proposed ( B a , B , prop ; B a , B , prop ) sus-pension ( B a , B ) orrevocation ( B a , B ) ofcertification ( B a ) . . . Where a single prop-erty applies to multipleproperties, referencesto both objects/prop-erties are maintainedon this property (sepa-rated by semicolon).
AND , OR , XOR , NOT
Statement,Component IG Core,Regulative& Con-stitutiveStatements The logical operatorsidentify the relation-ship between statementand/or componentsas either conjunction(
AND ), inclusivedisjunction ( OR ), ore-xclusive disjunction( XOR ). Where nega-tion is involved, the
NOT operator is used(e.g., in deontics: mustnot; combination of ex-ceptions:
NOT option1
AND option 2).
Certifiers ( must reviewapplications and ( AND )must not (
NOT ) approveapplications ) by offenders. Table 4: Symbol Reference for IG Coding as applied in this document
IG 2.0 Codebook Version: 1.0 27.2 Regulative Statement Coding
The base syntax of regulative statements (as shown in Figure 9) consist of necessary and sufficientcomponents, and further involves further feature refinements across different levels of expressiveness.Figure 9: Syntax and Features of Regulative Statements by Level of ExpressivenessTables 5 to 7 provide detailed coding guidelines for regulative statements by level of expressiveness,starting with IG Core through IG Logico, leveraging the notation convention captured in Section 4.1. Asthe institutional analyst commences coding at any level of expressiveness, the following general codingprinciples should be entertained. • Where possible, the targeted level of encoding should be clarified at the beginning (see Sec-tion 2.6). • The coder should acquaint oneself with the concepts relevant for the target level of encoding (seeSection 2.6). • Recall that coding can occur iteratively, starting at one level that prompts less granular syntacticexpressiveness (e.g., IG Core) moving with a subsequent coding pass to another level that promptsmore granular coding (e.g., IG Extended).
IG 2.0 Codebook Version: 1.0 28 . R e g u l a t i v e S t a t e m e n t C o d i n g IG Core enables basic, structural analysis of institutional statements. Encoding at this level is designed to be human readable and moderatelycomprehensive in the detail with which syntactic properties of institutional statements are captured.
SyntacticComponent Treatment of Syntactic Componentsby Level of Encoding Relevant Examples Complete Syntactic Classifica-tion of Examples
Attribute
The encoding of Attributes, which caninclude an animate actor (individual ororganizational) only or an animate actorand a property of this actor, differentiatesbetween actor and actor property.Note that in all cases, named entities(e.g., United States Department of Agri-culture) are not decomposed into actorproperties.A heuristic to decide as to whether entitynames are decomposed relies on the insti-tutional context. If, for example, a regu-lation differentiates between organic andnon-organic farmers, the decompositioninto descriptor and associated propertiesis useful. If, however, the regulation is ex-clusively concerned with organic farmers,the decomposition is of little analyticalvalue at this level of encoding. Example statement:
Certified farmer must submit an organicsystem plan annually.
Attribute encoding:Attribute = certified farmer
Actor = farmer
Actor property = certified Certified ( A , prop ) farmer ( A )must ( D ) submit ( I ) an organicsystems plan ( B dir ) annually ( C ex ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Object
The encoding of Objects identifies Di-rect Objects specified within institutionalstatements, along with their respectiveproperties. Where statements are com-prised of both direct and indirect object,it also entails the explicit identification ofindirect objects, i.e., objects that are af-fected by the application of the aim todirect objects, e.g., which the action istargeted to.Rules: • Identify object identifier • Identify object propertiesNote that in all cases, named entities(e.g., United States Department of Agri-culture) are not decomposed into objectproperties. The heuristics for decompo-sition as outlined in the context of At-tributes equally applies for objects. Example statement:
Organic certifier must send farmer noti-fication of compliance within 30 days ofinspection.
Direct Object = notification of compli-ance
Indirect Object = farmer
Object property = organic
Note: The interpretation of property de-pends on the encoded policy (and, ofcourse, the coder’s analytical objectives).If a policy on organic farming exclusivelyrefers to organic farmers as a propernoun, organic is not an attribute prop-erty, but part of the attribute; if the policydifferentiates between organic and othertypes of farmers, capturing the specificcharacterisation as property is suggested.For the example here we highlight thesecond pathway.
Organic ( A , prop) certifier ( A )must ( D ) send ( I ) farmer ( B ind )notification of compliance ( B dir )within thirty days of inspection( C ex ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Object (ctd.)
Pitfalls: • Objects vs. Constraints : The introduction of the indirect object offersthe benefit of capturing the functional interdependence of objects andimplied directionality. This directionality is sometimes explicitly re-flected, as in the following statement. When encountering such state-ments, the coder should be careful to not mischaracterize the indirectobject preceded by a preposition as context. Example: “Parents musttake children to school.”, “school” is sensibly resolved as indirect ob-ject, but based on its prepositional embedding, could be mislabeled asconstraint. For the characterization, we thus require an initial consid-eration of clauses containing objects (other than the direct object) asindirect objects, before characterizing those as a contextual descrip-tor (which primarily make reference to the contextual embedding ofactions). • Object Properties vs. Constraints : Differentiation between the objectproperty and execution constraint can also sometimes be challenging inthe encoding. When such confusion arises, the coder should ask herselfto reflect on whether the statement words or clauses in question arequalifying the object or qualifying the action of the statement. Take thefollowing statement for example that illustrates the referenced potentialconfusion: “Certifiers shall perform audits on product stock two timesper year.” The clause that may potentially give rise to confusion is“on product stock.” This could be confused as an execution constraintrelating to purpose, but in fact it describes the type of audit to beperformed.
Aim
The encoding identifies the focal actionof the statement. Example statement:
Organic certifier must send farmernotification of compliance.
Aim = send
Organic ( A , prop ) certifier ( A )must ( D ) send ( I ) farmer ( B ind )notification of compliance ( B dir ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Deontic
The encoding identifies the prescriptiveoperator that indicates whether the Aim(i.e., action) of the statement is required,allowed, or forbidden. Common Deon-tics indicating varying levels of prescrip-tive force include must, may, and mustnot. Example statement:
Organic certifier must send farmernotification of compliance.
Deontic = must
Organic ( A , prop ) certifier ( A )must ( D ) send ( I ) farmer ( B ind )notification of compliance ( B dir ).Context The encoding identifies the Context ofthe institutional statement. The en-coding differentiates between “ActivationConditions,” which are contextual clausesthat specify preconditions under whichthe Aim is expected to occur or not oc-cur, and “Execution Constraints,” whichare contextual descriptors that qualify theAim by assigning in relation to it tempo-ral, spatial, procedural, and/or other con-straining parameters. Example statement:
Upon entrance into agreement with or-ganic farmer to serve as his/her certify-ing agent, organic certifier must inspectfarmer’s operation within 60 days.
Context clauses:
Upon entrance intoagreement with organic farmer to serve ashis/her certifying agent; within 60 days.
Context encoding:Activation Condition:
Upon entranceinto agreement with organic farmer toserve as his/her certifying agent
Execution Constraint: within 60 days
Note: While coding the essential aspectsof the statement, IG Core is limited withrespect to capturing the nested complex-ity in the activation condition. We willrevisit this statement in the context of IGExtended coding.
Upon entrance into agreement withorganic farmer to serve as his/hercertifying agent ( C ac ), organic cer-tifier ( A ) must ( D ) inspect ( I )farmer’s operation ( B dir ) within 60days ( C ex ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Or else
The encoding of Or else statementsidentifies consequences (e.g., payoffs) ofcompliance/non-compliance with institu-tional statements, or the conduct ofAim (i.e., activities) assigned to spe-cific Attributes (i.e., actors) in institu-tional statements. The encoding cap-tures these consequences, which gener-ally take the form of regulative insti-tutional statements that nest from the‘non-Or else’ (monitored) statements.Sometimes, the statements on which Orelse, or consequential statements, nestconvey a combination of monitoring ac-tivity and the associated payoff attachedto outcomes of monitoring actions.The encoding of Or else statements ac-commodates both vertical and horizon-tal nesting. Vertical nesting is applica-ble when there is one payoff activity thatis specified within a distinct institutionalstatement as a consequence of an actionindicated in another institutional state-ment. Horizontal nesting is applicablewhen there are two or more payoff activi-ties that can be pursued as consequencesof an action indicated in another institu-tional statement. Example:
Certified organic farmers must not applysynthetic chemicals to crops at any timeonce organic certification is conferred, orelse certifier will revoke certification fromfarmer.
Or else clause comprising statement: orelse certifier will revoke certification fromfarmer
Or else statement nests on: certified or-ganic farmers must not apply syntheticchemicals to crops at any any time onceorganic certification is conferred
The above encoding exemplifies verticalnesting. The following example, whichis an extension of the above, exempli-fies horizontal nesting within a verticallynested statement.Example statement:
Certified organic farmers must not applysynthetic chemicals to crops at any timeonce organic certification is conferred, orelse certifier will revoke certification fromfarmer or fine farmer.
In the following example, horizontalnesting is signaled using parenthe-ses ( ( and ) ) around statements (asopposed to individual components),and vertical nesting is expressed us-ing brackets ( [ and ] ).Vertical nesting: Certified ( A , prop1 ) organic ( A , prop2 ) farmers ( A ) must not ( D )apply ( I ) synthetic chemicals ( B dir )to crops ( B ind ) at any time ( C ex )once organic certification is con-ferred ( C ac ), or else [ certifier ( A )will ( D ) revoke ( I ) certification( B dir ) from farmer ( B ind ) ] . I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Or else (ctd.)
These statement combinations can signal • alternative exclusive action options– XORs – (e.g., either suspendingXOR revoking the certification), • inclusive action options – ORs –(e.g., sanctions apply if a driver iscaught speeding AND/OR on thephone) or • co occurring action options – ANDs– (e.g., fining a transgression ANDreporting to authorities). Or else clauses comprising statements: or else certifier will fine farmer or revokecertification from farmer. Vertical and horizontal Or else nesting en-coding:
Certified organic farmers must not applysynthetic chemicals to crops at any timeonce organic certification is conferred [Vertical nesting] or else certifier will revoke certificationfrom farmer [Horizontal nesting]
XOR (signaling exclusive or)
Or else certifier will fine farmer
Horizontal nesting within vertically-nested statement:
Certified ( A , prop1 ) organic ( A , prop2 ) farmers ( A ) must not ( D )apply ( I ) synthetic chemicals ( B dir )to crops ( B ind ) at any time ( C ex )once organic certification is con-ferred ( C ac ), OR ELSE [( certifier ( A ) will ( D )revoke ( I ) certification ( B dir ) fromfarmer ( B ind ) ) XOR ( certifier ( A )will ( D ) fine ( I ) farmer ( B dir ) )] . Note: • Where component-level combinations exist (. . . fine farmer or revoke. . . ), those have to signaled explicitly by parentheses, or be decom-posed into separate logically-combined complete atomic institutionalstatements. Further details are provided under “General IG ExtendedInstructions” (Table 6), Item “Decomposition of component-level com-binations”. • Ambiguities with respect to the linguistic use of logical operators (ex-clusive and inclusive or) are to be resolved as part of this process. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g General IG Core Instructions for Regulative Statements
Additional anno-tations for At-tributes, Objects,and Context (Note: this isequivalent toadditional an-notations forConstitutedEntity, Consti-tutive Functionand Constitut-ing Propertiesin the contextof constitutivestatements) In addition to the identification of proper-ties embedded in the original statements,components can further be annotated us-ing additional annotation labels. Suchlabels can follow the categories listed inSection 5, or be specific to the projectobjectives.A systematic approach to labelling enti-ties is discussed under “IG Logico Instruc-tions” (Table 7), Item “Cross-componentSemantic Annotations”. This is particu-larly recommended if annotations are ofstrong relevance for the coding and of di-verse nature. Example statement:
Organic certifier must send farmer notifi-cation of compliance.
Subject to analytical necessity, additionalannotations can for instance relate to theidentification of aspects such as the char-acterisation of encoded objects with re-spect to their animacy as either animateor inanimate – signified in brackets inthe coded example. Where indicated,the annotation should be separated fromthe component specification by semicolonand have the structure “ label= ”, followedby the annotation.While exemplified here for regulativestatements, this equally applies to con-stitutive statements.
Organic certifier(
A;label=animate ) must( D ) send ( I ) farmer( B ind ;label=animate ) no-tification of compliance( B dir ;label=inanimate ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Decomposition ofcomponent-levelcombinations (Note: This ap-plies to regula-tive and constitu-tive statements,and is discussedhere with focuson the regulativeperspective.) Where combinations of components(component-level combinations) are ob-served that are not explicitly decomposedas in the case of vertical nesting, or notexplicitly identified as component-levelcombinations (e.g., using parentheses),these can be decomposed into logically-combined statements. Other than foraims, the decomposition is optional forIG Core.Extended details are provided under“General IG Extended Instructions”(Table 6), Item “Decomposition ofcomponent-level combinations”; IG CoreExamples can be found in the following. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Decomposition ofcomponent-levelcombinations(ctd.)
Operationally, combinations of compo-nents are evidenced by the presence ofmultiple attributes, objects, aims or exe-cution constraints within a single institu-tional statement. Examples of each caseare provided in the next column.Decomposition essentially entails con-structing an individual statement to cap-ture each of the unique components rep-resented in multiples within institutionalstatements, noting the relation to theoriginal statement in which multiple com-ponents are reflected. Information fromcomponent fields, other than that con-taining multiple components, is simplycarried over to all related institutionalstatements.Importantly, where decomposition actu-ally changes the meaning of the originalinstitutional statement containing multi-ple components within a particular syn-tactic field, the statement should not bedecomposed. In such cases, multiplesare typically intended to exist in coupledform. An example is provided in the nextcolumn.Note: These guidelines highlight the mo-tivation for the decomposition, and exem-plify the process. Depending on the useof annotation means and tool support,the decomposition may be signaled byannotation of component combinations,and thus occur automated without re-quiring explicit decomposition by users. Multiple Attributes Example Statement:
Certifiers and Inspectors must seek ac-creditation annually.
Decomposed as:Statement 1:
Certifiers must seek accred-itation annually.
AND
Statement 2:
Inspectors must seek ac-creditation annually.
Multiple Aims Example Statement:
Inspectors must sign and file farm inspec-tion reports following site visits.
Decomposed as:Statement 1:
Inspectors must sign farminspection reports following site visits.
AND
Statement 2:
Inspectors must file farminspections reports following site visits.
Multiple Conditions Example Statement:
Inspectors must conduct site visits in per-son twice per year.
Decomposed as:Statement 1:
Inspectors must conductsite visits in person.
AND
Statement 2:
Inspectors must conductsite visits twice per year. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Decomposition ofcomponent-levelcombinations(ctd.)
Coupled Component Example Statementnot to be decomposed:
Farmers must pay Certifier $250 for ap-plication and service fees upon entry intocontract for certification services.
The reason for foregoing decompositionin such case lies in the inseparability ofapplication and services fees, since theyare reported as a combined fee of $250.Table 5: Coding Guidance on Syntactic Elements for IG Core as Level of Expressiveness (Regulative Statements)
IG Extended enables more detailed structural analysis of institutional data than IG Core and accommodates computational application to aid ininstitutional coding and analysis. Encoding at this level is designed to be human readable, moderately computationally tractable, and moderatelycomprehensive in the detail with which syntactic properties of institutional statements are captured.Coding institutional statements on this level enforces many of the features that have been optional in IG Core and affords a fine-grained decompo-sition of statements. This includes a richer context characterisation based on predefined taxonomies, the expansion and combined attributes andaims that reconstruct atomic statements and their relationships, but also decomposes the hierarchical relationships amongst explicitly highlightedactors, object and their respective properties. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g SyntacticComponent Treatment of Syntactic Componentsby Level of Encoding Relevant Examples Complete Syntactic Classifica-tion of Examples
Attributes
Building on the IG Core coding thatidentifies attributes and their respectiveproperties, IG Extended affords a morecomprehensive decomposition of the at-tributes component into Attributes, prop-erties, along with their respective func-tional descriptors (higher-order proper-ties). In addition, we can identify re-lated attribute objects (so called to main-tain the reference to the attribute) thatcarry their own properties and parame-ters. This approach is explored in the firstexample.Within this attribute, property, parameterhierarchy, elements may be substituted byan institutional statement, in which casecomponent-level nesting applies for theencoding. We explore this approach inthe second example.Note that in all cases, named entities(e.g., United States Department of Agri-culture) are not decomposed. Example Statement:
A certified farmer whose certification issuspended by the Secretary under thissection may at any time submit a recer-tification request.
In this example, the attribute “A certifiedfarmer whose certification is suspendedby the Secretary under this section . . . ” embeds aspects central to the institu-tional configuration. Beyond the identifi-cation of properties of the attribute ( “cer-tified” ), it highlights a complex propertyin the form of a nested institutional state-ment signaled by the involvement of an-other entity. Reformulated as “Secretarysuspends certified farmer’s certification” ,we can construe this second property asa nested statement with the correspond-ing component characteristics, as shownin the coding of the example.While objects related to attributes are ob-jects, they are coded with a reference tothe attribute to ensure unambiguous as-sociation with the attributes componentin the institutional statement.
A certified ( A , prop1 ) farmer ( A ) { whose certification ( B dir ) is sus-pended [suspends ( I )] by theSecretary ( A ) under this section( C ex ) } ( A , prop2 ) may ( D ) at anytime ( C ac ) submit ( I ) a recertifica-tion ( B dir , prop ) request ( B dir ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Attributes (ctd.)
Where properties or parameters applyto multiple objects, the respective refer-ences are separated by semicolon (e.g., A a , prop1 ; A b , prop1 ). This coding isexemplified in the context of the objectcomponent.Properties can operate on an arbitrarylevel of depth. For example, properties ofproperties (second-order properties) arecoded as ( A , prop1 , prop1 ), where prop-erty identifier are unique on each level.Using a further example, we can show-case the coding of beliefs or assessmentsmore generally in the form of second-order property hierarchies that rely oncomponent-level nesting as shown before:Example Statement: Program managers who believe that acertified operation has violated the Actmay pursue revocation proceedings.
In contrast to the previous statementthat highlights complex property arrange-ments on a given level, this statementemphasises a hierarchical organisationamongst properties, where the second-order property contains a complete insti-tutional statement.
Program managers ( A ) who be-lieve ( A , prop ) { that a certifiedoperation ( A ) has violated ( I )the Act ( B dir ) } ( A , prop , prop ) may( D ) pursue ( I ) revocation proceed-ings ( B dir ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Attributes (ctd.)
Nested institutional statement = certifiedoperation has violated the Act
Similar to the previous coding, attributesare decomposed into their identifier andproperties. Here, the nested state-ment is captured in the second-orderproperty ( A , prop , prop ), since the state-ment describes the content of the belief( A , prop ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Object
Building on the differentiation into di-rect and indirect object along with theirrespective properties in IG Core, in IGExtended the coding is refined to cap-ture relationships between elements inobject specifications. Specific focus lieson the decomposition of hierarchical re-lationships between objects and proper-ties respective. In addition, embeddedobjects without direct functional relation-ship are identified, i.e., objects that arereferred to but are not explicitly acted onin the context of the institutional state-ment.As highlighted for the attributes com-ponent, elements of the object-propertyhierarchy may be substituted by aninstitutional statement, in which casecomponent-level nesting applies for theencoding.Rules: • The identification of the direct orindirect Object respectively. • The identification of properties as-sociated with the Object, both in-cluding directly functionally depen-dent (e.g., descriptors) and concep-tually independent properties. Ifan object is a named entity (e.g.,United States Department of Agri-culture), it is not decomposed. Example statement:
The Program Manager shall send a writ-ten notification of proposed suspension orrevocation of certification to certified or-ganic farmer.
Similar to the attributes coding in IG Ex-tended, this statement decomposes com-plexity embedded in the object compo-nent:Object = written notification of proposedsuspension or revocation of certification
The direct object in this statement isthe “notification” , which has the prop-erty “written” , and makes reference toanother related functionally independentobject “certification” . Functional inde-pendence here refers to the fact that thecertification does not depend on the noti-fication. The certification itself is charac-terised by further dependent objects suchas the “suspension” and “revocation” .Both objects are dependent, since theydepend on the existence of a certifica-tion. Both dependent objects share theproperty of being “proposed” . The Program Manager ( A ) shall( D ) send ( I ) a written ( B dir , prop )notification ( B dir ) of proposed( B a , , prop ; B a , , prop ) suspen-sion ( B a , ) or revocation ( B a , )of certification ( B a ) to certified( B ind , prop1 ) organic ( B ind , prop2 )farmer ( B ind ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Object (ctd.) • For each property, functional rela-tionships are identified (i.e., whichObject or property they rely on).Where such relationships exist,these properties become child prop-erties of the properties they func-tionally depend on. This processproduces an Object hierarchy thatmay or may not directly involve theObject component itself. • Where multiple children exist, im-plicit or explicit logical relationships(conjunction, disjunction, nega-tion) between different branches ofthe emerging Object hierarchy areretained. • Non-functional relationships are es-tablished between the root node ofthe Object hierarchy and the director indirect Object established in thefirst step.Where object hierarchies are specifiedoutside the object component (e.g., aspart of the context component), the sameprocess applies, even if no reference ismade to direct or indirect object. Note,however, that those are coded with lowerpriority, i.e., following object hierarchiesthat relate to direct or indirect object. Establishing these relationships, we cancode the object and respective propertiesas follows:Related functionally-independent object= certification
Functionally independent objects areidentified as individual objects withunique alphabetical index, e.g., ( B a ),( B b ), etc.Dependent objects = suspension, revoca-tion Dependent objects are coded with ref-erence to the object they depend onand unique numeric index, e.g., ( B a ,1 ),( B a ,2 ), etc.Dependent object properties = proposed Dependent object properties are iden-tified with reference to objects theyrelate to, with references to differentobjects separated by semicolon, e.g.,( B a ,1,prop ), ( B a ,2,prop ), etc.Where multiple properties exist for anobject, they are, as in IG Core coding,uniquely identified by numeric index, e.g.,( B a ,1,prop1 ), ( B a ,1,prop2 ), etc. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Object (ctd.)
Where higher-order properties exist,these are coded according as specified inthe context of the Attributes component.Object hierarchies outside the objectcomponent:Taking the example “Certifier ( A ) must( D ) send ( I ) notification ( B dir ) to in-spector ( B ind ) to produce a written re-port of the assessment ( C ex , pur )” , wefind an object hierarchy within the con-text (execution constraint) component.In this example, first-order object is thereport that is written and relates to an as-sessment, without being functionally de-pendent – as reflected in the coding. Certifier ( A ) must ( D ) send( I ) notification ( B dir ) to inspec-tor ( B ind ) to ( produce a written( B a ,1,prop1 ) report ( B a ,1 ) of theassessment ( B a ) ) ( C ex , pur ). In this example, we use parenthesesto clearly delineate the scope of theexecution constraint that containsthe external object hierarchy.
Aim
The coding of the aim is identical to IG Core. See also Item “Decomposition of component-levelcombinations” in this table (Table 6) for associated decomposition rules.
Deontic
The coding of the deontic is identical to IG Core. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Context
In IG Extended, Activation conditions( C ac ) and Execution constraints ( C ex )(as specified in IG Core) are further char-acterised in terms of ontological cate-gories, specifically capturing the natureof circumstances the conditions or con-straints describe. Those are grouped intotaxonomies comprehensively described inSection 5 (along with subcategories).The central taxonomy in this context isthe circumstantial taxonomy, an overviewof which is provided in the following,along with corresponding coding:Excerpt of Circumstance Taxonomy: • Temporal (tmp) • Spatial (loc) • State (ste) • Procedural (prc) • Method (met) • Purpose (pur) • Observed state/outcomes (eff)A fine-grained overview of subcategories(e.g., differentiation of temporal categoryinto ‘point in time’ and ‘time frame’) isprovided in Section 5, alongside furtherspecific abbreviations (e.g., tfr for timeframe). Example 1:
Upon entrance into agreement with or-ganic farmer to serve as his/her certify-ing agent, organic certifier must inspectfarmer’s operation within 60 days.
This statement is rather complex andincludes both component-level nesting,along various further constraint specifi-cations.Retracing this coding is best achieved byreformulating this statement as follows:
Organic certifier ( A ) must ( D ) inspect( I ) farmer’s operation ( B dir ) within 60days ( C ex ,tfr ) under the condition { thatthe organic farmer ( A ) enters ( I ) anagreement ( B dir ) with the organic certi-fier ( B ind ) to serve as his/her certifyingagent ( C ex ,pur ) } ( C ac ,prc ). Here the execution constraint for the ini-tial inspection ( C ex ,tfr ) is activated bythe preceding procedural activation con-dition ( C ac ,prc ), expressed as a nestedinstitutional statement, that the farmerenters an agreement, whose purposeis defined as an execution constraint( C ex ,pur ). Upon entrance ( I ) into agreement( B dir ) with organic farmer ( A ) toserve as his/her certifying agent( C ex ,pur )( C ac ,prc ), organic cer-tifier ( A ) must ( D ) inspect ( I )farmer’s operation ( B dir ) within 60days ( C ex ,tfr ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Context (ctd.)
In addition to the contextual characteri-sation, conditions or constraints can fur-ther be expressed as primitive statements(e.g., at 8am), or complex expressionsembedding complete institutional state-ments (e.g., if an organic farmer violatesconditions for certification, . . . ).Where logical relationships between con-ditions or constraints exist, these aremade explicit by annotating those as
AND , OR , and XOR relationships.Where negation exists, this is to be like-wise identified (
NOT ).Rules: • Identify involved actions (explicitand tacit) by reformulating state-ments in active terms, and expandinto (potentially multiple) state-ments. Example 2:
When rebuttal is unsuccessful or correc-tion of the noncompliance by certifiedorganic farmer is not completed withinthe prescribed time period, the ProgramManager shall send a written notificationof proposed suspension or revocation ofcertification to certified organic farmer.
This statement, similar to the previ-ous example, includes the encoding of anested institutional statement. In con-trast, it highlights the combination ofmultiple conditions statements, whereone is of implicit nature and the secondone explicit.Coding this example, we first identify thetop-level institutional statement:Top-level institutional statement = theProgram Manager ( A ) shall ( D ) send ( I )a written ( B dir , prop ) notification ( B dir )of proposed ( B a ,1,prop ; B a ,2,prop )suspension ( B a ,1 ) or revocation ( B a ,2 )of certification ( B a ) to certified organicfarmer ( B ind ). When { [certified ( A , prop1 )organic ( A , prop2 ) farmer( A )][rebuts] ( I ) unsuccess-fully ( C ex ,eff ) } ( C ac ,eff ) OR { correction ( B dir , prop ) ofthe noncompliance ( B dir ) bycertified organic farmer ( A ) is notcompleted ( I ) within the prescribedtime period ( C ex ,tfr ) } ( C ac ,tfr ),the Program Manager ( A ) shall( D ) send ( I ) a written ( B dir , prop )notification ( B dir ) of proposed( B a ,1,prop ; B a ,2,prop ) suspen-sion ( B a ,1 ) or revocation ( B a ,2 )of certification ( B a ) to certified or-ganic farmer ( B ind ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Context (ctd.) • Hint: where statements contain anaim that is linked to an object asnoun, it is indicative of a missingor implied actor specification. Suchcase is signaled by passive tense(e.g., notification is received). Theactor specification is either impliedfrom context or potentially signaledby prepositional clauses such as “ byactor ”; Example: “
Notification bycertifier is received by farmer ”. Insuch cases, the statement requiresreformulation in active terms alongwith the injection of an inferredexplicit action an explicit action,e.g., “ certifier sends notification tofarmer ”. In some cases, this mayrequire the expansion of a singlestatement into separate statementscarrying separate actions (see Ex-ample 3). • Identify logical relationships(
AND , OR , XOR ) and de-pendencies (sequence) amongstactions. • Reconstruct complete statementusing component-level nesting ofstatements where relevant. In this statement we observe principlesof the Attributes/Object-property hierar-chy (see Section 2.4), which are encodedbased on the instructions provided for ob-jects.Context clause =
When rebuttal is un-successful or correction of the noncom-pliance by certified organic farmer is notcompleted within the prescribed time pe-riod
The context clause (which contains twoactivation conditions) can be decom-posed into two statements that are logi-cally combined ( OR ). The passive aim onan object (rebuttal is unsuccessful) sig-nals an implied action, and requires re-formulation.First condition statement = When [or-ganic farmer ( A )] [rebuts ( I )] unsuccess-fully ( C ex ,eff ) The second condition (“ correction of thenoncompliance by certified organic farmeris not completed within the prescribedtime period ”) can be reformulated in ac-tive terms as “ if certified organic certi-fier has not completed correction of non-compliance within the prescribed time pe-riod ”. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Context (ctd.)
Doing so, the structure of the nestedstatement becomes overt:“ if certified ( A , prop1 ) organic( A , prop2 ) certifier ( A ) has not ( D )completed ( I ) correction ( B dir , prop )of non-compliance ( B dir ) within theprescribed time period ( C ex ,tfr ) ”Both conditions are further logicallyrelated by an inclusive disjunction(AND/OR), which is annotated explicitlyusing AND , OR , or XOR , respectively.The composition of all three statementsis shown in the example.Example 3:
When an inspection of an accredited cer-tifying agent by the Program Manager re-veals any noncompliance with the Act orregulations in this part, a written notifi-cation of noncompliance shall be sent tothe certifying agent.
Note: While highlighted here for activa-tion conditions, such logical combinationequally applies to statements containingmultiple execution constraints. { When [Program Manager( A )] reveals ( I ) any non-compliance ( B dir ) [by the ac-crediting ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind )] withthe Act or regulations in this part( C ex ,eff ) { [under the conditionthat] Program Manager ( A )[performs] ( I ) inspection ( B dir )of an accredited ( B ind , prop1 )certifying ( B ind , prop2 ) agent( B ind ) } ( C ac ) } ( C ac ), [ProgramManager ( A )] shall ( D ) [send( I )] a written ( B dir , prop1 ) noti-fication ( B dir ) of noncompliance( B dir , prop2 ) to the certifying( B ind , prop1 ) agent ( B ind ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Context (ctd.)
As highlighted before, for the purpose ofcoding the reformulation of statementsin active terms may be useful to facili-tate coding. In the first statement “
Whenan inspection of an accredited certifyingagent by the Program Manager revealsany noncompliance with the Act or regu-lations in this part ”, the acting party isthe Program Manager (actor), but theaim reveals relates to the object inspec-tion. This is indicative of a tacit ac-tion (captured as conceptual reificationin the object) on the part of the programmanager (captured in the prepositionalclause). In this case, we can decomposethe compound conditional statement intotwo statements:
Program Manager [performs] inspection (where the performance is tacit)
AND [Program Manager] reveals any non-compliance . . . (in which case the at-tribute is tacit). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Context (ctd.)
In addition, the second statement de-pends on the activation of the first action.The conditional statement can thus bedecomposed into two institutional state-ments, as follows:
When [Program Manager ( A )] re-veals ( I ) any non-compliance ( B dir ) [bythe accrediting ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind )]with the Actor regulations in this part ( C ex ,eff ) { [under the condition that] ProgramManager ( A ) [performs] ( I ) inspection( B dir ) of an accredited ( B ind , prop1 ) cer-tifying ( B ind , prop2 ) agent ( B ind ), . . . The final part of the statement (themain statement), likewise reformulatedin active terms, implies that [Pro-gram Manager] shall [send] a written( B dir , prop1 ) notification ( B dir ) of non-compliance ( B dir , prop2 ) to the certifying( B ind , prop1 ) agent ( B ind ).Structurally, in this statement we observetwo levels of component-level nestingin the form ABDIC { ABDIC { ABDIC }} ,where the main statement’s (shallsend notification; first ABDIC ) activa-tion relies on revealing potential non-compliance (second
ABDIC ), which in it-self relies on the inspection in the firstplace (last
ABDIC ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Or else
The coding of the Or else is identical to IG Core. The internal structure of nested statements isencoded according to IG Extended instructions (Table 6). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g General IG Extended Instructions for Regulative Statements
Decomposition ofcomponent-levelcombinations
In the presence of multiple Attributes, ac-tions (aims) and objects in a given state-ment, such statements are to be decom-posed into individual statements that arecombined with the corresponding logicaloperator (following the principles of hor-izontal nesting). At face value this mir-rors the approach taken in the contextof or Else statements. However, whilein the context of Or else components,combinations designate logical relation-ships amongst action alternatives, in thiscontext the purpose is to disambiguatethe relationship amongst particular ac-tors, actions and objects and to resolveincongruences between the linguistic andlogical use of conjunctives.When a disaggregation or aggregationfundamentally alters the meaning of thestatement (e.g., if a payoff associatedwith the institutional statement cannotbe unambiguously associated with an in-dividual entity), or if expressions are in-tentionally coupled (e.g., chips and fish)or proper names (e.g., Smith and Sons),the statement is not to be decomposed. Example Statement:
Certified operations or handlers mustcomply with organic farming regulations.
This can be decomposed into
Certified operations must comply with or-ganic farming regulations
AND
Certified handlers must comply with or-ganic farming regulations.
Note that the interpretation of the logicaloperator is contextual. In this example, itcarries the understanding that the spec-ified obligation applies to both certifiedoperations and certifier handlers.Naturally, this approach can lead to thedecomposition into a large number of ad-ditional statements, e.g., attribute andaction combinations - as exemplified be-low. For practical reasons, the explicitcoding can be substituted by an addi-tional annotation that reflects the needfor decomposition.
Certified ( A , prop ) operations ( A )must ( D ) comply ( I ) with organicfarming ( B dir , prop ) regulations( B dir ) AND
Certified ( A , prop ) handlers ( A )must ( D ) comply ( I ) with organicfarming ( B dir , prop ) regulations( B dir ). Recall that the minimal institu-tional statement presumes the ex-istence of context specifications,which – in absence of specific en-coding – resolves to “under all cir-cumstances” (for activation condi-tions), “without any constraints”(for execution constraints). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Decomposition ofcomponent-levelcombinations(ctd.)
Example Statement:
Certified operations or handlers must ac-cept and comply with organic farmingregulations.
Decomposed:
Certified operations must accept organicfarming regulations
AND
Certified handlers must accept organicfarming regulations
AND
Certified operations must comply with or-ganic farming regulations
AND
Certified handlers must comply with or-ganic farming regulations.
Practical considerations:If labelling is performed manually, a prac-tical consideration for such decomposi-tion is to keep track of the relationshipsof such statements, e.g., by introducingsub-identifiers. For example, assumingthe coded statement is Statement 10, thedecomposed statements could be anno-tated as 10.1, 10.2, etc.
Certified ( A , prop ) operations ( A )must ( D ) accept ( I ) organic farm-ing ( B dir , prop ) regulations ( B dir ) AND
Certified ( A , prop ) handlers ( A )must ( D ) accept ( I ) organic farm-ing ( B dir , prop ) regulations ( B dir ) AND
Certified ( A , prop ) operations ( A )must ( D ) comply ( I ) with organicfarming ( B dir , prop ) regulations( B dir ) AND
Certified ( A , prop ) handlers ( A )must ( D ) comply ( I ) with organicfarming ( B dir , prop ) regulations( B dir ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Decomposition ofcomponent-levelcombinations(ctd.)
Alternatively, annotation can be per-formed in shorthand form, or rely on tool-specific support offered by the appliedtext annotation tool that allows the in-dication of decomposition during the en-coding.In shorthand form, the decomposed state-ments can be grouped by parentheses tosignal component-level combinations asshown below.
Example: ( Operators ( A ) AND
Certifiers( A ) ) must ( D ) comply ( I ) with ( regula-tions ( B dir ) AND best practices ( B dir ) ) . In expanded form (shown on the right),parentheses are then used to signal theassociation of the decomposed state-ments. Grouping of statements usingparentheses (in bold font): ( Certified ( A , prop ) operations( A ) must ( D ) accept ( I ) organicfarming ( B dir , prop ) regulations( B dir ) AND
Certified ( A , prop ) handlers ( A )must ( D ) accept ( I ) organic farm-ing ( B dir , prop ) regulations ( B dir ) AND
Certified ( A , prop ) operations ( A )must ( D ) comply ( I ) with organicfarming ( B dir , prop ) regulations( B dir ) AND
Certified ( A , prop ) handlers ( A )must ( D ) comply ( I ) with organicfarming ( B dir , prop ) regulations( B dir ) ) Logical relation-ships amongstatement com-ponents
Where statements make tacit referenceto multiple actions, conditions or con-straints, these are likewise resolved us-ing the introductions provided in Table 7(IG Logico), Item “Logical relationshipsamong statement components”.For IG Extended this provision is recom-mended, but optional. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Regulative-constitutiveHybrids
The introduction of constitutive statements as part of IG 2.0 (see Section 4.3) provides the basis for encoding statementsthat consist of structural elements both of regulative and constitutive statements. Details are discussed in Section 4.4.Table 6: Coding Guidance on Syntactic Elements for IG Extended as Level of Expressiveness (Regulative Statements)
IG Logico is designed to support semantic analysis of institutional statements wholly relying on computational tools. Encoding at this level isdesigned to be moderately human readable, computationally tractable and comprehensive in the detail with which syntactic properties of institutionalstatements are captured.In contrast to IG Core and Extended that focus on the encoding of specific grammar components, IG Logico emphasises refinements acrossindividual components and further establishes explicit references to related statements to establish computational tractability, as well as the abilityto perform logical transformations on institutional statements. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g SyntacticComponent Treatment of Syntactic Componentsby Level of Encoding Relevant Examples Complete Syntactic Classifica-tion of Examples
Relation-centricSemantic Anno-tations
To establish relationships amongst state-ments and policies more generally, the ini-tial refinement refers to the identificationof statement references, e.g., between in-dividual statements, collections thereof,or policies more generally.To this end, an additional annotationidentifies all instances of references toother statements. Note that cross-statement references are not specific toany statement components, but applyacross complex component types, includ-ing attributes, objects and context.Rules: • Identify references in coded state-ments • Add annotation of structure(ref=value), where ref signalsthe reference nature, and valuecontains an identifier of referencedstatement, collection, or policy. Exploring this approach, we borrow thecomplex example statement previouslycoded in the context of IG Extended:Example Statement:
When an inspectionof an accredited certifying agent by theProgram Manager reveals any noncom-pliance with the Act or regulations in thispart, a written notification of noncompli-ance shall be sent to the certifying agent.
Coded form: { ( When [program manager ( A ) performs( I )] an inspection ( B dir ) of an accredited( B ind , prop1 ) certifying ( B ind , prop2 )agent ( B ind ) ) [ AND ] ( Program Man-ager ( A ) reveals ( I ) any noncompli-ance ( B dir ) with the Act or regula-tions in this part ( C ex ,eff ) ) } ( C ac ,prc ),[Program Manager ( A ) ] a written ( B dir , prop1 ) notification ( B dir ) of noncom-pliance ( B dir , prop2 ) shall ( D ) be sent[send ( I )] to the certifying agent ( B ind ). { ( When [program manager ( A )performs ( I )] an inspection ( B dir )of an accredited ( B ind , prop1 )certifying ( B ind , prop2 ) agent( B ind ) ) [ AND ] ( ProgramManager ( A ) reveals ( I ) anynoncompliance ( B dir ) withthe Act (ref=“policy”) orregulations in this part (ref=“section”) ( C ex , eff ) ) } ( C ac ,prc ),[Program Manager] ( A ) a written( B dir , prop1 ) notification ( B dir ) ofnoncompliance ( B dir , prop2 ) shall( D ) be sent [send ( I )] to the certi-fying agent ( B ind ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Relation-centricSemantic Anno-tations (ctd.)
This statement makes reference to “ theAct ” and “ regulations in this part ”, thefirst of which makes reference to thecoded policy in its entirety, whereas thesecond one focuses on a specific sectionof the policy.Both are coded by providing an additionalannotation ( ref=value ), along with thescope reference (“ value ”), i.e., an iden-tifier of relevance in the context of theencoded policy. The identifiers need to beunambiguous within the given document,including statement IDs, section headers,or documents as a whole, etc. The corre-sponding convention should be decided aspart of project-specific coding guidelines.If occurring in conjunction with exist-ing component classification, the refer-ence specification is appended (e.g.,
Act( B dir ,ref=“policy” ) ).As stated before, while, in this specificexample, cross-statement reference applyto constraints components, such refer-ences can likewise occur in other com-ponents. ( { When [Program Manager( A )] reveals ( I ) any non-compliance ( B dir ) [by the ac-crediting ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind )] withthe Act ( ref=“policy” )( C ex ,eff ) { [under the condition that] ( Program Manager ( A ) [performs( I )] inspection ( B dir ) of an ac-credited ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind ) )OR( Program Manager ( A ) [performs( I )] inspection ( B dir ) of an ac-credited ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind ) )OR( Program Manager ( A ) [performs( I )] inspection ( B dir ) of an ac-credited ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind ) ) }} ( C ac ), [Program Manager( A )] shall ( D ) [send ( I )] a written( B dir , prop1 ) notification ( B dir )of noncompliance ( B dir , prop2 ) tothe certifying ( B ind , prop1 ) agent( B ind ). )OR ( ← decomposition of “ Act ”and “ regulations in this part ”). . . I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Logical relation-ships amongstatement com-ponents
The objective of decomposition of lists,or other forms of implied conjunctions,such as multiple conditions/constraints isto make logical relationships explicit. Insuch cases enumerations are decomposedinto individual statements and combinedusing the corresponding logical operator.Note that this is similar to the expan-sion of multi-entity components in IGExtended with specific emphasis on at-tributes and objects. In this case, the re-view operates across all component typesand explicitly focuses on implied logicalrelationships.Rules: • Identify action alternatives embed-ded in lists or enumeration, or inconditions/constraints • Identify associated atomic state-ment • Establish logical operator andprecedence where needed • Expand statement via horizontalnesting to capture individual actionalternatives We use a modified example previously ex-plored under IG Extended, Context com-ponent.Example Statement:
When an inspection, review, or investiga-tion of an accredited certifying agent bythe Program Manager reveals any non-compliance with the Act or regulations inthis part, a written notification of non-compliance shall be sent to the certifyingagent.
Following the decomposition patterns forContext specification established previ-ously (IG Extended, Context), we observethat an inspection of a program managerhas to be performed, and may reveal non-compliance, and arrived at the followingcoding: { When [Program Manager ( A )] reveals( I ) any non-compliance ( B dir ) [bythe accrediting ( B ind , prop1 ) certify-ing ( B ind , prop2 ) agent ( B ind )] withthe Act or regulations in this part( C ex ,eff ) { [under the conditionthat] Program Manager ( A ) [performs( I )] inspection ( B dir ) of an accredited( B ind , prop1 ) certifying ( B ind , prop2 )agent ( B ind ) }} ( C ac ), . . . . . . ( { When [Program Manager( A )] reveals ( I ) any non-compliance ( B dir ) [by theaccrediting ( B ind , prop1 ) cer-tifying ( B ind , prop2 ) agent( B ind )] with regulations in this part( ref=“section” ) ( C ex ,eff ) { [under the condition that] ( Program Manager ( A ) [performs( I )] inspection ( B dir ) of an ac-credited ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind ) )OR( Program Manager ( A ) [performs( I )] inspection ( B dir ) of an ac-credited ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind ) )OR( Program Manager ( A ) [performs( I )] inspection ( B dir ) of an ac-credited ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind ) ) }} ( C ac ), [Program Manager( A )] shall ( D ) [send ( I )] a written( B dir , prop1 ) notification ( B dir )of noncompliance ( B dir , prop2 ) tothe certifying ( B ind , prop1 ) agent( B ind ). ) I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Logical relation-ships amongstatement com-ponents (ctd.) . . . [Program Manager ( A )] shall( D ) [send ( I )] a written ( B dir , prop1 )notification ( B dir ) of noncompli-ance ( B dir , prop2 ) to the certifying( B ind , prop1 ) agent ( B ind ). In this example, the program manager’snotification is contingent on an inspec-tion, a review, or investigation, i.e., avariation of instruments for assessment.Logically, the initial task (inspection) thatis prerequisite for further action. Theoriginal statement is { [under the condition that] ProgramManager ( A ) [performs ( I )] inspection( B dir ) of an accredited ( B ind , prop1 ) cer-tifying ( B ind , prop2 ) agent ( B ind ) } ( C ac ) Given that we now have three alternativeactions, the statement requires expansioninto three separate tasks.To achieve this, the logical relationshipbetween the tasks needs to be estab-lished. Subject to context, the coder caninterpret the relationship as either an in-clusive disjunction ( OR ) or exclusive dis-junction ( XOR ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Logical relation-ships amongstatement com-ponents (ctd.)
Since realistically (based on interpreta-tion of application context) a combina-tion of any of such tasks could equallylead to the detection of non-compliance,suggesting the combination via OR . Thisexample showcases the importance ofcoding context and interpretation, whichmakes an explicit specification necessaryfor analytical treatment of action alterna-tives.In consequence, the statement is ex-panded into OR -combined statements asfollows (Note: while not necessary inthis case, the logical combination of thestatements is signaled using surroundingparentheses): . . . [under the condition that] (( Program Manager ( A ) [performs( I )] inspection ( B dir ) of an accredited( B ind , prop1 ) certifying ( B ind , prop2 )agent ( B ind ) )OR( Program Manager ( A ) [performs ( I )] re-view ( B dir ) of an accredited ( B ind , prop1 )certifying ( B ind , prop2 ) agent ( B ind ) )OR . . . I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Logical relation-ships amongstatement com-ponents (ctd.) . . . ( Program Manager ( A ) [performs ( I )] in-vestigation ( B dir ) of an accredited( B ind , prop1 ) certifying ( B ind , prop2 )agent ( B ind ) )) This coding is embedded in the completestatement coding as shown on the right(with formatting adjustments so as tomake the discussed coding easily acces-sible).Note: While not applicable in this case,where necessary, precedence of specificcombinations has to be signaled usingparentheses (e.g., if inspection OR re-view is permitted, or as an exclusive al-ternative, the investigation, which wouldbe (simplified) represented as ( inspection OR review ) XOR investigation )) .Another aspect that requires explicit cod-ing in this example is the reference to thescope of violation, i.e., non-compliancewith the Act or regulations in this part,the logical relationship of which (here:“or”), subject to coder interpretation,has to be coded explicitly. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Logical relation-ships amongstatement com-ponents (ctd.)
In this example the relationship is charac-terized by an inclusive disjunction ( OR ),since regulations in this part are part ofthe Act. As with all other cases, thisannotation is made explicit, and conse-quently, requires decomposition of thestatement by duplication into correspond-ing statement variants: When [Program Manager ( A )] re-veals ( I ) any non-compliance ( B dir ) [bythe accrediting ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind )] with the Act( ref=“policy” )( C ex ,eff ). . . (remainder of statements) . . . OR When [Program Manager ( A )] re-veals ( I ) any non-compliance ( B dir ) [bythe accrediting ( B ind , prop1 ) certifying( B ind , prop2 ) agent ( B ind )] with reg-ulations in this part ( ref=“section” )( C ex ,eff ). . . (remainder of statements) . . . I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Logical relation-ships amongstatement com-ponents (ctd.)
Such decomposition affords a systematicassessment of the individual variants inthe context of a specific situation, butlikewise offers automation potential. Forexample, given the assumption that regu-lations in this part are a subset of the Act,we could ignore the statement variantthat assesses the compliance with regula-tions in this part. However, for the sakeof comprehensive illustration of the em-bedded institutional complexity (and sub-ject to alternative interpretations), we de-compose this example comprehensively. I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g Cross-componentSemantic Anno-tations
In addition to explicit encoding of prop-erties as specified in the underlying state-ment (e.g., explicit identification of “cer-tified” as a property of an operation), se-mantic qualities can be enriched by pro-viding additional annotations that cap-ture a differentiation of components withrespect to different ontological categoriescaptured in different taxonomies (seeSection 5), including their animacy, ac-tor roles as well as action themes of thecoded statement. In contrast to the prop-erty annotations used in IG Core and Ex-tended, the annotations introduced hereapply across all components, and explic-itly emphasize extensibility both with re-spect to additional categories within thegiven taxonomies, as well as specifica-tion of further taxonomies, e.g., based ondomain-specific or analytical necessities.Furthermore, multiple annotations of dif-ferent categories can be applied to a givencomponent at the same time, e.g., phys-ical descriptors, such as animate, can becombined with role descriptors, such asRecipient.By taxonomy, categories for such anno-tations include • Physical type: animate, inanimate • Role: Source, Recipient, Possessor,Experiencer, BeneficiaryAn extended listing of the associated tax-onomies along with their description isprovided in Section 5. Complementing the characterisation ofcontext by circumstance, entities can fur-ther be annotated by the role they playin a particular setting, as well as fur-ther properties, such as physical types.These categorizations can be extendedbeyond the specified types, apply acrossall component types, and furthermore al-low the introduction of additional tax-onomies, beyond the ones highlighted inSection 5.To annotate components with additionalcategories, the component coding is ex-tended by key-value pairs, with the keyspecifying the taxonomy, and the valuethe corresponding categorization(s). Inthe coding highlighted here, the categoryspecifications are separated from compo-nent specification by semicolon. Wheremultiple categories for a given taxonomyapply, these are separated by comma.The example on the right highlights thisapproach with respect to the physicaltype and role taxonomies specified in Sec-tion 5. { When [Program Man-ager ( A ; type=animate ; role=experiencer )] reveals( I ) any non-compliance ( B dir ; type=inanimate ) [by theaccrediting ( B ind , prop1 )certifying ( B ind , prop2 )agent ( B ind ; type=animate ; role=originator )] with theAct ( type=inanimate , ref=“policy” ) or regulationsin this part ( ref=“section” )( C ex , eff ) } ( C ac ), [ProgramManager ( A ; type=animate ; role=originator )] shall ( D ) [send( I )] a written ( B dir , prop1 ) noti-fication ( B dir ; type=inanimate )of noncompliance ( B dir , prop2 )to the certifying ( B ind , prop1 )agent ( B ind ; type=animate ; role=recipient ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g InstitutionalFunction Anno-tations
IG Logico further prescribes annotationsof statements with institutional func-tions. While other features describedabove in practice focus on the refinedcoding of attributes, objects and condi-tions, institutional actions reflect the in-stitutional function of a statement’s ac-tion.Being fined, for example, reflects theinstitutional function of “sanctioning”,adhering to regulation reflects “compli-ance”.Institutional functions as identified in thisspecification include the following set,some of which operate complementaryand are listed as comma-separated func-tion pairs: • Comply, Violate • Reward, Sanction • Monitor • Detect compliance, Detect non-compliance • DelegateThe Institutional Function Taxonomy isdescribed in Section 5. Taking the previously used statement asan example (including inferred compo-nents, but omitting any annotations), wecan identify institutional functions asso-ciated with actions. { When [Program Manager] reveals anynon-compliance [by the accrediting cer-tifying agent] with the Act or regulationsin this part } , [Program Manager] shall[send] a written notification of noncom-pliance to the certifying agent. Actions of institutional relevance include: • Reveal (non-compliance) • Send (notification)In this context revealing non-complianceabstractly corresponds to the “ detectionof a violation ”, whereas sending a notifi-cation reflects a form of “ sanctioning ”. Inconsequence, the statement can be anno-tated with these institutional functions,so as to enable inferences from a purelyinstitutional perspective without concernfor the specific operationalization of de-tecting compliance or sanctioning in aspecific scenario. { When [Program Man-ager ( A ; type=animate ; role=experiencer )] reveals( I ; function=detect viola-tion ) any non-compliance( B dir ; type=inanimate ) [bythe accrediting ( B ind , prop1 )certifying ( B ind , prop2 )agent ( B ind ; type=animate ; role=originator )] with theAct ( type=inanimate , ref=“policy” ) or regulationsin this part ( ref=“section” )( C ex ,eff ) } ( C ac ), [ProgramManager ( A ; type=animate ; role=originator )] shall ( D )[send ( I ; function=sanction )] awritten ( B dir , prop1 ) notifica-tion ( B dir ; type=inanimate )of noncompliance ( B dir , prop2 )to the certifying ( B ind , prop1 )agent ( B ind ; type=animate ; role=recipient ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g InstitutionalFunction Anno-tations (ctd.)
The annotation follows the syntacticspecification applies for other forms ofannotations, i.e., appending a key-valuepair to the component coding, where thekey is “ function ” and the value carries thecorresponding institutional function spe-cific to the annotated action.Providing an additional simplified exam-ple, we can explore the use of further an-notations:
The Program Manager may initiate re-vocation proceedings against a certifiedoperation { When the Program Manager has reasonto believe that a { certified operation hasviolated the Act } OR When a certifying agent fails to take ap-propriate action to enforce the Act } . . . . The Program Manager ( A )may ( D ) initiate ( I ; func-tion=sanction ) revocationproceedings ( B dir ) against acertified operation ( B ind ) { When the Program Manager( A ) has reason to believe( I ; function=evaluate ) thata { certified operation ( A ) has vi-olated ( I ; function=violate ) theAct ( B dir ; ref=“policy” ) } ( B dir ) OR When a certifying agent ( A )fails ( I ; function=violate ) totake appropriate action ( B dir )to enforce the Act ( C ex , pur ; ref=“policy” ) } ( C ac ). I G . C o d e b oo k V e r s i o n : . . R e g u l a t i v e S t a t e m e n t C o d i n g InstitutionalFunction Anno-tations (ctd.)
The key actions include: • Initiate (revocation proceedings),corresponding to a sanction • “Has reason to believe” reflects anevaluation on the part of the actor • “Violate” reflects a violation • “Fail to take appropriate action”likewise represents a violationIn the coding, these actions can thus beannotated with the corresponding institu-tional functions. General IG Logico Instructions for Regulative Statements
A central objective is to provide a consistent coding that reflects the most fine-granular level of encoding. While the order of encoding is looselyprescribed by the order of specification, in some cases a variation of the order may be indicated. This should be considered as part of the codingpreparation. The coding may further require iterative review, specifically with respect to annotations and logical relationships.While implicit in the multi-pass coding implied for IG Logico, a dedicated review of embedded object hierarchies (encoded as part of IG Extended)and the explication of logical relationships between component elements (e.g., specific execution constraints) is of central concern in IG Logico.Table 7: Coding Guidance on Syntactic Elements for IG Logico as Level of Expressiveness (Regulative Statements) I G . C o d e b oo k V e r s i o n : . .3 Constitutive Statement Coding Constitutive statements are treated analogous to regulative statements, offering selected syntactic cor-respondence and corresponding refinements across levels of expressiveness (visualized in Figure 10),with variations relating to the structural decomposition of selected syntax elements and the explicittreatment of statement-level decomposition for constitutive-regulative hybrids (see Section 4.4), as wellas semantic annotations on IG Logico.Figure 10: Syntax and Features of Constitutive Statements by Level of ExpressivenessMirroring the introduction of coding guidelines for regulative statements, in Table 8 we provideinstructions for constitutive statements. Given the feature overlap between regulative and constitutivestatements, we make reference to selected feature sets described in the context of regulative statementsas part of the coding guidelines. As for regulative statements, symbols are color-coded to signal theassociation with features specific to IG Core, IG Extended or IG Logico. Symbols associated with IG Corefeatures for constitutive statements are held in purple. As in the previous tables, symbols associatedwith IG Extended are held in green, and features associated with IG Logico are displayed in orange.
IG 2.0 Codebook Version: 1.0 68 . C o n s t i t u t i v e S t a t e m e n t C o d i n g IG Core enables basic, structural analysis of institutional statements. Encoding at this level is designed to be human readable and moderatelycomprehensive in the detail with which syntactic properties of institutional statements are captured. I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g SyntacticComponent Treatment of Syntactic Componentsby Level of Encoding Relevant Examples Complete Syntactic Classifica-tion of Examples
ConstitutedEntity
The encoding of the Constituted Entity,which reflects any entity created, modi-fied or otherwise introduced into the insti-tutional setting. Constituted entities canbe of physical or virtual nature, reflectconcrete or abstract concepts, typicallyincluding actors, roles, actions, and ob-jects. Constituted entities can further bedifferentiated into entity and entity prop-erty.Note that in all cases, named entities(e.g., United States Department of Agri-culture) are not decomposed into actorproperties.A heuristic to decide as to whether entitynames are decomposed relies on the insti-tutional context. If, for example, a regu-lation differentiates between organic andnon-organic farmers, the decompositioninto descriptor and associated propertiesis useful. If, however, the regulation is ex-clusively concerned with organic farmers,the decomposition is of little analyticalvalue at this level of encoding. Example statement:
There is hereby established a public FoodSecurity Advisory Board.
Entity =
Food Security Advisory Board
Entity property = public
Example statement:
No member of the Council shall be dis-qualified from holding any public office oremployment.
While reflecting structural patterns ofregulative statements, this statement pa-rameterizes members with respect torights in the context of the Council.Beyond the necessary components ( E , F and implied Context), the substantivecharacteristics that do NOT apply (seenegation applied) to the constituted en-tity are expressed as constituting proper-ties.Additional example:
Established in this Regulation subpart isthe right to appeal to a revocation or cer-tification. There is hereby ( C ex ) established( F ) a public ( E , prop ) Food Secu-rity Advisory Board ( E ).No ( NOT ) member ( E ) of theCouncil ( E , prop ) shall ( D ) be ( F )disqualified from holding any publicoffice or employment ( P ).Established ( F ) in this Regulationsubpart ( C ex ) is the right to appeal( E ) to a revocation or certification( E , prop ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g Constituted En-tity (ctd.)
Constitutive statements and Implied At-tributes: In instances in which a coderis encountering ambiguity in discerningwhether she is dealing with a constitutiveor regulative statement, one shall con-sider the wider context of the statement(e.g., implied attribute, type of surround-ing statements, etc.). A more detaileddiscussion can be found in Section 4.4.4. I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g ConstitutiveFunction
The constitutive function characterizesthe establishment, definition or introduc-tion of a constituted entity into the insti-tutional setting, and where constitutingproperties exist, functionally link consti-tuted entity and constituting properties. Example statement:
There is hereby established a public FoodSecurity Advisory Board.
Constitutive Function: [is] . . . estab-lished
In this context the constitutive functionsignals the establishment of an entity.Example:
Commissioner of Agriculture and Marketsshall be the Chairperson the Council.
Constitutive Function: [serve as]
Here the constitutive function indicates amodified position (
Chairperson ) of a spe-cific role (
Commissioner ) in a specific or-ganizational context (
Council ).While diverse in nature, the constitutingfunction can be sensibly organized alonga set of patterns discussed in the contextof IG Logico.
There is hereby ( C ex ) established( F ) a public Food Security Advi-sory Board ( E ).Commissioner of Agriculture andMarkets ( P ) shall ( D ) be ( F ) theChairperson the Council ( E ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g ConstitutingProperties
Constituting properties are optionalcomponents in constitutive institutionalstatements that capture elements func-tionally linked to the constituted entityby means of the constitutive function.Constituting properties may themselveshave properties. Example:
The Committee shall consist of a Presi-dent, Secretary, and Treasurer.
Constituting properties:
President, Sec-retary, and Treasurer
Here, the council is composed of themembers as constituting properties.Example:
A majority of the members of the Councilshall constitute a quorum.
Constituting properties: majority of themembers of the Council The Committee ( E ) shall ( D ) con-sist ( F ) of a President, Secretary,and Treasurer ( P ).A majority of the members of theCouncil ( P ) shall ( D ) constitute( F ) a quorum ( E ).Deontic The Deontic signals the extent to whichthe instruction contained in the consti-tutive statement is prescribed or signalsdiscretion. In constitutive statements theuse of the deontic can also be of conven-tional nature and relies on the contextualinterpretation based on disciplinary (e.g.,legal) traditions and stylistic conventions. Example:
A majority of the members of the Councilshall constitute a quorum.
Deontic: shall
Example:
The Council shall have an advisory com-mittee.
Deontic: shall A majority of the members of theCouncil ( P ) shall ( D ) constitute( F ) a quorum ( E ).The Council ( E ) shall ( D ) have( F ) an advisory committee ( P ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g Context
The encoding identifies the Context ofthe institutional statement. The en-coding differentiates between “Activa-tion Conditions,” which are contextualclauses that specify preconditions underwhich the statement applies, and “Ex-ecution Constraints,” which are contex-tual descriptors that qualify the consti-tuting function by augmenting the state-ment with temporal, spatial, procedural,and/or other constraining parameters. Example statement:
From 1st of January onward, Food PolicyCouncil reporting requirements apply forany communication between the Counciland Regional Council in addition to com-munal provisions.
Context clauses:
From 1st of January on-ward; in addition to communal provisions
Context encoding:Activation Condition:
From 1st of Jan-uary onwards
Execution Constraint: in addition tocommunal provisions
The activation condition signals an eventthat initiates a discretized setting inwhich the remaining statement holds.The execution constraint characterizesthe constitutive function more explicitly.
From 1st of January onward ( C ac ),Food Policy Council reporting re-quirements ( E ) apply ( F ) for anycommunication ( P ) between theCouncil and Regional Council ( P , prop ) in addition to communal pro-visions ( C ex ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g Or else
The encoding of Or else statementsidentifies consequences (e.g., payoffs) ofcompliance/non-compliance or violationwith institutional statements, which, inthe context of constitutive statements,can be of consequential as well as ex-istential nature. The encoding capturesthese consequences generally in the formof institutional statements that nest onthe leading monitored institutional state-ment,and can be expressed both in regu-lative or constitutive form.Principles of horizontal and vertical nest-ing, as described in the regulative con-text, equally apply for constitutive state-ments.The encoding of Or else statements ac-commodates both vertical and horizon-tal nesting. Vertical nesting is applica-ble when there is one payoff activity thatis specified within a distinct institutionalstatement as a consequence of an actionindicated in another institutional state-ment. Horizontal nesting is applicablewhen there are two or more payoff activi-ties that can be pursued as consequencesof an action indicated in another institu-tional statement. Example:
In student recruitment plans, diversitymust mean diversity in race, religion, sex-ual orientation and gender, or else plan isvoid.
Or else clause comprising statement: orelse plan is void
The Or else signals an existential conse-quence for the constituted entity.Naturally, the consequence can also con-sist of multiple statements that are log-ically combined (horizontal nesting), asshown below.Example:
In student recruitment plans, diversitymust mean diversity in race, religion, sex-ual orientation and gender, or else plan isvoid and to be revised within 30 days. In student recruitment plans ( C ex ),diversity ( E ) must ( D ) mean ( F )diversity in race, religion, sexual ori-entation and gender ( P ),or else ( OR ELSE ) [ plan ( E ) is ( F ) void ( P ) ] .In student recruitment plans ( C ex ),diversity ( E ) must ( D ) mean ( F )diversity in race, religion, sexual ori-entation and gender ( P ),or else ( OR ELSE ) [( plan ( E ) is ( F ) void ( P ) )AND( . . . plan is . . . ( E ) to be revised( F ) within 30 days ( C ex ) )] . In the previous example, horizon-tal nesting is signaled using paren-theses around statements (as op-posed to individual components),and vertical nesting is expressed us-ing brackets ( [ and ] ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g Or else (ctd.)
These statement combinations can signal • alternative exclusive action options– XOR s – (e.g., either suspendingXOR revoking the certification), • inclusive action options – OR s –(e.g., sanctions apply if a driver iscaught speeding AND/OR on thephone) or • co occurring action options – AND s – (e.g., fining a transgres-sion AND reporting to authorities) Note:1. Where component combinationsexist, alternatives are combined(. . . are void and to be refinedfine farmer or revoke . . . ), and aresubsequently decomposed into sep-arate logically-combined completeatomic institutional statements.2. Ambiguities with respect to the lin-guistic use of logical operators (ex-clusive and inclusive or) are to beresolved as part of this process. I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g General IG Core Instructions for Constitutive Statements
Additional an-notations forConstituted En-tity, ConstitutiveFunction, Consti-tuting Propertyand Context (Note: this isequivalent toadditional an-notations forAttribute, Objectand Contextannotationsfor regulativestatements) In addition to the identification of proper-ties embedded in the original statements,components can further be annotated us-ing additional annotation labels. Suchlabels can follow the categories listed inSection 5, or be specific to the projectobjectives.A systematic approach to labelling enti-ties is discussed under “IG Logico Instruc-tions” (Table 7), Item “Cross-componentSemantic Annotations”. This is particu-larly recommended if annotations are ofstrong relevance for the coding and of di-verse nature. Example:
The Committee shall consist of a Presi-dent, Secretary, and Treasurer.
Subject to analytical necessity, additionalannotations can for instance relate tothe identification of aspects, such as thecharacterisation of encoded objects withrespect to their animacy as either ani-mate or inanimate – signified in bracketsin the coded example. Where indicated,the annotation should be separated fromthe component specification by semicolonand have the structure “label=”, followedby the annotation.While exemplified here for constitutivestatements, this equally applies to reg-ulative statements.
The Committee ( E ; la-bel=inanimate ) shall ( D ) consist( F ) of a President, Secretary, andTreasurer ( P ; label=animate ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g Decomposition ofcomponent-levelcombinations (Note: This ap-plies to regula-tive and consti-tutive statement,and is discussedhere with focuson the regulativeperspective.) Where combinations of components(component-level combinations) are ob-served that are not explicitly decom-posed as in the case of vertical nesting,these can be decomposed into logically-combined statements. Other than forconstitutive functions, the decompositionis optional for IG Core.Operationally, combinations of compo-nents are evidenced by the presenceof multiple logically-combined tokens orclauses embedded in constituted entities,constitutive functions, constituting prop-erties or context components.Decomposition essentially entails con-structing an individual statement to cap-ture each of the unique components rep-resented in multiples within institutionalstatements, noting the relation to theoriginal statement in which multiple com-ponents are reflected. Information fromcomponent fields, other than that con-taining multiple components, is simplycarried over to all related institutionalstatements.Importantly, where decomposition actu-ally changes the meaning of the originalinstitutional statement containing multi-ple components within a particular syn-tactic field, the statement should not bedecomposed. In such cases, multiplesare typically intended to exist in coupledform. An example is provided in the nextcolumn. Details are described in “General IG Ex-tended Instructions” (Table 6), Item “De-composition of component-level combi-nations”Example (Multiple Properties):
The Committee shall consist of a Presi-dent, Secretary, and Treasurer.
While expressed in condensed form as“
The Committee ( E ) shall ( D ) consist( F ) of a ( President, Secretary, and Trea-surer ) ( P ) ” (note the parentheses),it corresponds to the following statementcomposed of three atomic statements:Statement 1: The Committee shall con-sist of a President
AND
Statement 2:
The Committee shall con-sist of a Secretary.
AND
Statement 3:
The Committee shall con-sist of a Treasurer.
Example (Multiple constitutive func-tions):
The form and function of theCouncil is hereby established.
Condensed form:
The Committee ( E ) shall ( D ) con-sist of ( F ) a ( President
AND
Sec-retary
AND
Treasurer ) ( P ). Expanded form: ( The Committee ( E ) shall ( D )consist of ( F ) a President ( P ) AND
The Committee ( E ) shall ( D ) con-sist of ( F ) a Secretary ( P ) AND
The Committee ( E ) shall ( D ) con-sist of ( F ) a Treasurer ( P ) ) .Council form ( E ) is hereby estab-lished ( F ) AND
Council function ( E ) is hereby es-tablished ( F ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g Decomposition ofcomponent-levelcombinations(ctd.)
Note: These guidelines highlight the mo-tivation for the decomposition, and ex-emplify it explicitly. Depending on theuse of annotation means and tool sup-port, the decomposition may be partiallyautomated, affording a mere annotationfor such decomposition without requiringthe user to perform statement duplica-tion.Table 8: Coding Guidance on Syntactic Elements for IG Core as Level of Expressiveness (Constitutive Statements) I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g Mirroring the progression on the regulative side, IG Extended enables more detailed structural analysis of institutional data than IG Core andaccommodates computational application to aid in institutional coding and analysis. Encoding at this level is designed to be human readable,moderately computationally tractable, and moderately comprehensive in the detail with which syntactic properties of institutional statements arecaptured.Coding institutional statements on this level enforces many of the features that have been optional in IG Core and affords a fine-grained decompo-sition of statements. This includes a richer context characterisation based on predefined taxonomies, the expansion and combined attributes andaims that reconstruct atomic statements and their relationships, but also decomposes the hierarchical relationships amongst explicitly highlightedconstituted entities, constituting properties and constitutive functions, alongside further refinements of contextual descriptors.As a central feature IG Extended makes the use of component-level combinations explicit. This specifically facilitates the decomposition of thecontext component to express institutional content at a more nuanced level. In addition, structural refinements relate to the decomposition ofrelationships and properties of constituted entities and constituting properties.For constitutive statements, IG Extended features correspond to the regulative side, with the essential difference for the application of refinementson Attributes and Objects, which, in the context of constitutive statements apply to constituted entities and constituting properties.A specific consideration is the concept of constitutive-regulative hybrids and syntactic polymorphs, both of which are of cross-cutting nature (i.e.,affecting both constitutive and regulative statements) and thus discussed in a dedicated section. Their consideration, however, applies to IGExtended.
SyntacticComponent Treatment of Syntactic Componentsby Level of Encoding Relevant Examples Complete Syntactic Classifica-tion of Examples
ConstitutedEntity
In IG Extended encoding, Constituted Entities and their properties are decomposed hierarchicallyfollowing the principles of the Attribute/Object-Property Hierarchy (Section 2.4) and is appliedanalogous to “Attributes” in IG Extended for regulative statements (Table 6). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g ConstitutingProperty
Analogous to the decomposition of ob-ject properties in the context of regulativestatements, constituting properties arelikewise decomposed following the princi-ples of the Attribute/Object-Property Hi-erarchy (as introduced in Section 2.4 andapplied in the context of “Objects” in IGExtended for regulative statements in Ta-ble 6). Example:
The Council consists of elected officialsresident in the electorate.
In this example, the individual proper-ties of the constituting property officials ,namely elected and resident in the elec-torate , are uniquely identified as proper-ties.Another feature is the richer hierarchicalstructure embedded in phrase expressingcompound property characterizations.Example:
A majority of the members of the Councilshall constitute a quorum.
In this example, the constituting propertyis captured in the phrase
A majority of themembers of the Council . While the entirephrase represents the constituting prop-erty (and is coded as such on IG Core),the embedded hierarchy, i.e., membersare a property of the Council, and themajority is a property of the members,can be explicitly captured using hierarchi-cal property annotations as shown on theright.Where properties are not functionally de-pendent on another property, they are sig-naled using unique identifiers (e.g., P a , P b ) equivalent to “Object” decomposi-tion highlighted in Table 6 and exempli-fied in the following. The Council ( E ) consists of ( F )elected ( P , prop1 ) officials ( P ) res-ident in the electorate ( P , prop2 ). ( A majority ( P , prop1 , prop1 ) ofthe members ( P , prop1 ) of theCouncil ( P ) ) ( P ) shall ( D ) consti-tute ( F ) a quorum ( E ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g ConstitutingProperty (ctd.)
Collections of functionally independententities are represented as a compoundconstituting property signaled by paren-theses. Individual compound propertiescan be uniquely identified, alongside po-tential further properties shared across allembedded entities. Example:
The Committee shall consist of a Pres-ident, Secretary, and qualified Treasurerappointed by the public.
In this example, properties specific to anentity are called out with reference to theentity ( qualified ), whereas shared proper-ties are associated with all entities ( ap-pointed by the public ). The Committee ( E ) shall ( D )consist of ( F ) a ( President( P a ) AND
Secretary ( P b ) AND qualified ( P c , prop1 ) Trea-surer ( P c ) ) ( P ) appointed by thepublic ( P , prop ).Context See “Context” in IG Extended for regulative statements (Table 6)
General IGExtended Instruc-tions
See “General IG Extended Instructions” in IG Extended for regulative statements (Table 6)
Constitutive-regulative Hy-brids
The introduction of constitutive statements as part of IG 2.0 (see Section 4.3) provides the basisfor encoding statements that consist of structural elements both of regulative and constitutivestatements. Details are discussed in Section 4.4.Table 9: Coding Guidance on Syntactic Elements for IG Extended as Level of Expressiveness (Constitutive Statements) Specific data structure patterns commonly found in institutional statements are revisited in Section 4.4.5. I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g IG Logico is designed to support semantic analysis of institutional statements wholly relying on computational tools. Encoding at this level isdesigned to be moderately human readable, computationally tractable and comprehensive in the detail with which syntactic properties of institutionalstatements are captured.In contrast to IG Core and Extended that focus on the encoding of specific grammar components, IG Logico emphasises refinements acrossindividual components and further establishes explicit references to related statements to establish computational tractability, as well as the abilityto perform logical transformations on institutional statements.While largely equivalent for regulative and constitutive statements, the only variant to the instructions provided in the context of regulativestatements is the discussion of Constitutive Function taxonomies (as opposed to Institutional Functions in the context of regulative statements)as outlined below. I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g SyntacticComponent Treatment of Syntactic Componentsby Level of Encoding Relevant Examples Complete Syntactic Classifica-tion of Examples
ConstitutiveFunction Anno-tations
Complementing the content characteriza-tion for other components, the constitu-tive function maintains the central role asa descriptor of constituted entities, andwhere constituting properties exist, linksthose to constituted entities.In an attempt to characterize the func-tion of the constitutive statement as ex-pressed in the constitutive function moregenerally, we propose a taxonomy captur-ing common relationships more generally.Doing so, we differentiate between state-ments that characterize the constitutedentity as newly introduced into the insti-tutional setting, and a commonly foundalternative, that is, the characterizationof the policy that contains the statementsitself.Entities, such as novel actors, objects,roles or action, can be • defined explicitly (“is”, “does”), • defined based on relationships,such as composition (“consistsof”), organizational embedding(“is embedded in”, “relates to”),and finally Example: Starting January 1st ( C ac ), the Connecti-cut Food Policy Council ( E ) shall ( D ) be( F ) within the Department of Agriculture( C ex ). In this example the constitutive functionsignals the constituted entity (Connecti-cut Food Policy Council) as an organiza-tional unit.Example:
The Committee shall consist of a Presi-dent, Secretary, and Treasurer.
The constitutive function signals a com-position of the constituted entity (
Com-mittee ) based on constituting properties.Example:
The purpose of this Part is to estab-lish standards for net metering in accor-dance with the requirements of Section16-107.5 of the Act.
In this example, the constitutive functionidentifies the entity as a policy and signalsthe intent underlying the policy.
Starting January 1st ( C ac ), theConnecticut Food Policy Coun-cil ( E ) shall ( D ) be ( F ; con-func=organization ) within theDepartment of Agriculture ( C ex ).The Committee ( E ) shall( D ) consist of ( F ; con-func=composition ) a ( President
AND
Secretary
AND
Trea-surer ) ( P ).The purpose of this Part ( E ) is( F ; confunc=intent ) to establishstandards for net metering in accor-dance with the requirements of Sec-tion 16-107.5 ( ref=Section/16-107.5 ) of the Act ( P ). I G . C o d e b oo k V e r s i o n : . . C o n s t i t u t i v e S t a t e m e n t C o d i n g ConstitutiveFunction Anno-tations (ctd.) • defined based on lifecycle stages(“established”, “terminated”). Afinal further form of constitutivefunctions relates to their applica-tion in the context of conferral ofrights, authority, or exertion of in-stitutional power more generally.Policies as constituted entities in institu-tional statements, in contrast, are gener-ally referred to with respect to the • lifecycle stage they are involved in(“come into force”), • relationship between and to otherstatements or policies (“amends”,“substitutes”), • intent in the form of purpose of aspecific policy, and appear as • information statements that offerinformation about the policy itself.Naturally, these characterizations are notexhaustive and can carry more specificforms. An overview of the different char-acterizations, alongside the labels used inthis context is provided in Section 5. Example: In department’s university plan, diversepopulation means diversity in religion,sexual orientation and race.
In this example, the constituted entity isdefined intensionally, that is in terms ofits underlying interpretations.
In department’s university plan( C ex ), diverse population ( E )means ( F ; confunc=definition )diversity in religion, sexual orienta-tion and race ( P ). Table 10: Coding Guidance on Syntactic Elements for IG Logico as Level of Expressiveness (Constitutive Statements) I G . C o d e b oo k V e r s i o n : . .4 Constitutive-Regulative Hybrids In addition to the specific treatment of regulative and constitutive statements as part of the codingguidelines, an aspect that demands specific attention in the coding guidelines is their combined use.While distinctively different in their function, regulative and constitutive statements of course sharestructural patterns as outlined in the context of the operational coding across varying levels of expres-siveness.However, an operational concern that links both statement types is the interleaved use in practice.In addition to the commonly found separation of constitutive and regulative statements into distinctsections (e.g., constitutive statements as part of the preamble), in regulative statements we may en-counter inline specifications of entities that are positioned in the institutional setting and are thus ofrelevance for subsequent statements. Conversely, in constitutive statements, we can potentially en-counter embedded regulative elements that regulate behaviour of the constituted entities. Linking thenested relationships of institutional statements across both types, we characterize the combined use ofconstitutive and regulative statements as constitutive-regulative hybrids (where the overall statement isof constitutive nature) or regulative-constitutive hybrids (where the leading statement is of regulativenature). Where existing, their resolution is a central feature of IG Extended (and optional for IG Core).In the following we will exemplify both variants of statement hybrids.
A typical reflection of hybrids stems from the introduction of novel entities as part of a regulativestatement, as shown in the following example (Figure 11):Figure 11: Regulative-Constitutive Hybrid ExampleAs signaled visually, this example highlights a regulative statement capturing an actor’s obligations,with the latter defined in an embedded constitutive statement, reflecting a regulative-constitutive hybrid.In this example, the constitutive statement is nested in a specific component of the regulative statement,such as the object as shown in the example in Figure 12. Note that the following figures use the samecolor-coding used in the preceding sections: Symbols associated with IG Core features for regulativestatements are displayed in blue, whereas symbols signaling constitutive statements are held in purple. In the context of this section, parentheses and logical operators are color-coded to emphasize the association with thecorresponding institutional statement type.
IG 2.0 Codebook Version: 1.0 86.4 Constitutive-Regulative Hybrids
Figure 12: Coded Regulative-Constitutive Hybrid ExampleThis interleaved representation can afford a decomposition of hybrids into individual statements byseparating the statements by syntactic components, and replication of components where overlapping.This decomposition is exemplified in Figure 13.Figure 13: Decomposed Regulative-Constitutive Hybrid Example
IG 2.0 Codebook Version: 1.0 87.4 Constitutive-Regulative Hybrids
Contrasting the embedding of constitutive statements in regulative settings, we can likewise observethe embedding of regulative statements in constitutive ones. In the following example (Figure 14),the violation of a constitutive statement can be expressed in regulative terms, following the principlesof statement-level nesting. While the leading statement is coded as a constitutive statement, theconsequences are a combination of regulative statements.Figure 14: Constitutive-Regulative Hybrid Example
In addition to the combined characterisation of constitutive and regulative hybrids, we can further ob-serve component-level nesting of constitutive statements as shown below. While in principle equallyadmissible for regulative statements, specifically the higher-order decomposition of constitutive state-ments is commonly found. Higher-order decomposition thereby implies the nesting of constitutivestatements within individual components, such as property items. Naturally, as motivated in the earlierexample, this can occur in conjunction with hybrid statements and independent of the regulative orconstitutive nature of the leading institutional statement.Figure 15: Second-order Constitutive Statement ExampleReviewing the example above (Figure 15, second-order statement in italicized bold font), we notethe reference to feed rations as a property element of livestock health care practices that is defined interms of an embedded constitutive statement that in itself captures a complex set of properties thatconstitute a feed ration.
IG 2.0 Codebook Version: 1.0 88.4.4 Institutional Statement Types & Polymorphic Institutional Statements
Another final aspect of discussion relates to the identification of constitutive and regulative statements.The function of either statement type is generally well defined based on the introduction or modifica-tion of entities or endowment of rights/authority for constitutive statements, and the specification ofoperational duties and constraints for given actors (or actor interaction) for regulative statements. Inpractice, however, statements alone may offer limited clarity as to whether they are of constitutive orregulative nature in the first place.Possible reasons for and approaches to resolution this ambiguity include the nature of the policy moregenerally, i.e., in how far the specific statement regulates central concerns of the policy, as opposed toparameterizing the scenario, or contextualizing the policy itself. Other, more specific indicators includethe positioning of the statement within the document. If located in the preamble, for example, theintended interpretation as constitutive statement is likely. Another consideration is the immediate con-text of a statement, i.e., its surrounding statements. Those may offer a clearer indication as to whetherthe statement is of configurational nature (constitutive), or bears operational weight (regulative). Afurther heuristic is of stylistic nature. Reviewing the coded document, the coder may find specificterms characteristic for constitutive or regulative statements in the context of the policy and/or field(e.g., the use of “shall” as indicative for constitutive statements, if obligations in regulative statementsare commonly expressed using more distinctive and specific deontics). Notwithstanding, a pragmaticapproach to seek support for either position, is to code the statement using both syntactic forms, andidentifying in how far the encoding affords extensive reformulation or reconstruction of a statement inorder to arrive at a sensible coding outcome.As an alternative to the focus on statement purpose, analytical objectives may drive a preferencefor either statement type in the form of a default strategy for the treatment of ambiguous statementtypes. For example, if the understanding of actor relationships across a given policy is of fundamentalconcern, but reconstruction of configurational aspects external to actorship are secondary, a potentialdefault strategy for the coding of ambiguous statements (e.g., defined as part of the project-specificguidelines) could be their interpretation as regulative.Where such ambiguity, and in consequence, flexibility exists, statements can be considered polymor-phic institutional statements , or syntactic polymorphs . This means, they can take either shape basedon the ambiguity they exhibit, but also the application they are subjected to.To substantiate this approach with an example, we can use the statement “The functions of the Board shall be: (a) give effect to the decisions and policies of the Health As-sembly; (b) act as the executive organ of the Health Assembly; (c) perform any other functions en-trusted to it by the Health Assembly.”
This statement can be read in constitutive terms, i.e., the characterization of the board in terms of itsfunctions and endowed responsibilities. The statement, however, can also be read in regulative terms, inwhich the functions of the board are expressed as obligations in operational terms. The same statementcan thus be encoded in terms of both statement types, as visualized in the following (including colorcoding to signal regulative (blue) and constitutive components (purple), respectively.)
The functions ( E ) of the Board ( A ) shall ( D / D ) be ( F ): ( (a) give ( I ) effect ( B dir ) to the decisions and policies of the Health Assembly ( B ind ); AND (b) act ( I ) as the executive organ of the Health Assembly ( C ex ); AND (c) perform ( I ) any other functions ( B dir ) entrusted to it by the Health Assembly( B dir , prop ) ) ( P ).IG 2.0 Codebook Version: 1.0 89.4.4 Institutional Statement Types & Polymorphic Institutional Statements The coding shows overlap, but the central components for regulative and constitutive statements(namely constituted entity, attributes, constitutive function, aim, as well as constituting properties)showcase the varying focal aspects of different statement types. Subject to emphasis on either theactor (i.e., ‘the Board’) or its functions (i.e., ‘the functions’) can determine the coding, or even admitboth approaches for analytical purposes.While the notion of polymorphic structure is explicit in the previous example, the coding of statementsin both terms can be more complex and may even afford reconstruction, as shown in the followingexample, in which the same statement is coded separately in constitutive and regulative terms: “No member of the Council shall be disqualified from holding any public office or employment.”
In constitutive terms the statement reflects the endowment of a right, namely the right to hold otherpublic positions in addition to a Council membership, and is encoded as follows:
No (
NOT ) member ( E ) of the Council ( E , prop ) shall ( D ) be ( F ) disqualified fromholding any public office or employment ( P ). Correspondingly, the Council member is at the center of the encoding. Depending on the analyticaluse, constitutive statements reflect an abstract conception of a right or a property. Expressing thisassurance in regulative terms – offering a more concrete characterization of involved actors and actions– requires the operationalization of behavioural constraints by introducing a conception of actorship –a tacit actor whose behaviour is constrained, alongside further structural adaptations. Reconstructingthe statement in regulative terms thus produces the following coding: [Attribute ( A )] [shall not ( D )] disqualify ( I ) member of the council ( B ) from holding anyoffice or employment ( C ex ). We can recognize that in this case, the encoding of the statement in constitutive or regulative termsmay invoke varying levels of complexity with respect to necessary adaptations of the statements, which,depending on analytical objectives, may potentially modify the semantics of a statement (in the givenexample, a concrete actorship for regulative statements is presumed, whereas the right is expressed moregenerally in constitutive terms) or afford a reformulation that may not be justifiable on methodologicalgrounds. In addition to considering a dual annotation in the first place, the potential mischaracterizationof a statement as either regulative or constitutive may sensibly be considered in inter-rater reliabilitytests, since it offers the opportunity to resolve disagreements and misconceptions early in the encodingprocess.
IG 2.0 Codebook Version: 1.0 90.4.5 Data Structure Patterns
Another aspect related to the features introduced for both regulative and constitutive statements is therecognition of data structure patterns. A pattern commonly observed is the notion of collections, suchas composition of committees, definition of practices in terms of underlying activities, delineation ofgoals or outcomes to be pursued, etc.Collections: While dominant in objects and constituting properties, respectively, those can occuracross other components (e.g., context). Central here is the identification of a collection descriptorand the corresponding elements, along with the explicit specification of logical operators that link theindividual elements.Example:
Health care practices ( E ) consist of ( F ) ( preventative measures [ AND ] appropriate nu-trition [
AND ] rest ) ( P ) .Complex elements: Where quantitative information is expressed, or listed (and thus potentially embed-ded in collections or referred to as a single item), the elements commonly follow a schematic structurespecific to individual documents (e.g., based on style or disciplinary background), but follow generalpatterns, such as variations of the following: [qualifier] [comparator] [quantity] [unit] [objectproperty] [object] .Example: significantly (qualifier) more than (comparator) 10 (quantity) tons (unit) high-quality(object property) building material (object) While those patterns can vary in extent and detail, their consideration in project-specific codingguidelines can be useful in as far as they are relevant for analytical purposes.
IG 2.0 Codebook Version: 1.0 91.5 Coding Level Configurations (Institutional Grammar Profiles)
As discussed before, the various coding levels of IG 2.0 accommodate different analytical needs aswell as complexity of the encoded documents. Commitment to a level, including all the associatedfeatures, may in some cases be too coarse-grained to accommodate analytical needs. This can include,for example, the omission of components entirely, as well as the selective considerations of features ofhigher coding levels. To capture the considered feature set, a coded dataset should be accompaniedwith the applied coding configuration.For this purpose, we specify a fine-grained configuration syntax that allows the choice of featuresacross levels of IG 2.0. The features for individual levels as specified in this document are capturedin Table 11 for regulative statements, and Table 12 for constitutive statements, along with a symbolassociation used for the ensuing specification of configurations.
Coding Level Feature Symbol
IG Core Attributes AIG Core Object BIG Core Deontic DIG Core Aim IIG Core Context CIG Core Or else OIG Extended Attributes refinements A
Ext
IG Extended Object refinements B
Ext
IG Extended Context refinements C
Ext
IG Logico Statement references RIG Logico Logical relationship annotations LIG Logico Semantic annotations SIG Logico Institutional function annotations FTable 11: IG Feature Specifications for Regulative StatementsUsing coding levels, along with – and + symbols in combination with specific features referencesas listed in the table, we can express specific coding configurations, or coding profiles, that allow theomission or inclusion of features across all levels, or the selective coding of specific components basedon lower coding levels.Abstractly specified, a configuration has the following structure (where < and > embeds the descrip-tion of the element content): < Baseline coding level > – < omitted features from baseline coding level > + < additional fea-tures from higher level > Examples:To capture the commitment to IG Core, along with the Context coding from IG Extended (e.g.,component-level nesting, use of taxonomies, is specified as the configuration
IG Core+C
Ext , wherethe +C Ext signals features from the next higher level (IG Extended).Conversely, we can specify coding on IG Core level without the consideration of Or else componentsas
IG Core–O . Where multiple components are omitted, we can specify
IG Core–IO , where featuresshould be referred to in the order as specified in the table (here:
Aim before
Or else ). Selectively cap-
IG 2.0 Codebook Version: 1.0 92.5 Coding Level Configurations (Institutional Grammar Profiles)
Coding Level Feature Symbol
IG Core Constituting Properties PIG Core Deontic DIG Core Constituted Entity EIG Core Constitutive Function FIG Core Context CIG Core Or else OIG Extended Constituting Properties refinements P
Ext
IG Extended Constituted Entity refinements E
Ext
IG Extended Context refinements C
Ext
IG Logico Statement references RIG Logico Logical relationship annotations LIG Logico Semantic annotations SIG Logico Constitutive function annotations UTable 12: IG Feature Specifications for Constitutive Statementsturing features from IG Logico in IG Core-based coding,
IG Core+R indicates the coding of statementrelationships in addition to the base IG Core coding.Finally, omission and extensions can be combined, with omissions specified first, followed by featureadditions, such as
IG Extended-BC+SF , to signal the coding of Object and Context on IG Corelevel, while considering semantic annotations and institutional functions in addition to this (reduced)IG Extended baseline. Complementing this discussion for the highest level,
IG Logico-S would implycomplete coding on IG Logico level under omission of semantic annotations.Combining both omission and extension leverages complete flexibility with respect to the compositionof features, and, where analytically useful, in principle even foregoing the inclusion of components definednecessary for institutional statements (i.e., A, I, and C component for regulative statements; F, E, andC for constitutive statements). For example, modeling the selective omission of components entirely,along with the inclusion of advanced features,
IG Core-AI+C
Ext SF signals IG Core baseline encodingunder omission of Attributes and Aim, while adding refined coding of Context (based on IG Extended),along with semantic annotations and institutional functions (from IG Logico).A common encoding level that offers the smallest possible extension to previous coding practice ofinstitutional statements based on Crawford and Ostrom’s original grammar is IG Core+C
Ext .Custom refinements: Where coders seek more fine-granular refinements (e.g., applying a subset ofthe features of a given configuration (e.g., coding objects without properties), such modifications shouldbe indicated alongside the specified configuration. Similarly, extensions (e.g., additional taxonomies, orextensions of existing ones) should likewise be documented alongside the configuration.
IG 2.0 Codebook Version: 1.0 93 Taxonomies
This section provides an overview of the taxonomies for the categorization of components, parts thereof,or annotation schemes including (but not limited to) the ones referred to from the Coding guidelinesin Section 4. The overview is largely summarizing, with essential specification of labels, but limitedconceptual elaboration, which is provided in the corresponding guidelines (Section 4). The taxonomiesfurther specify the label prefixes used to ensure unambiguous reference to the respective taxonomy/ies.Where only the circumstances taxonomy is used, the use of labels is optional.The extension of existing taxonomies and introduction of additional taxonomies (e.g., to accommo-date domain-specific characteristics or analytical necessities) is explicitly permitted. Both cases shouldbe clearly indicated and defined or referred to in conjunction with the coded dataset.
The circumstances taxonomy captures contextual characterizations with respect to temporal, spatialand various further descriptors that capture institutional context more accurately. It is a systematicextension of the descriptors of the Conditions component as highlighted in the original grammar speci-fication. Note that the listed categories include an embedded hierarchy, with more specific labels notedindented. Where possible (and analytically useful/specified in project-specific guidelines), the morespecific annotation should be used. Note that the state category is of highest abstractions and includestemporal, spatial and other categories. Where other categories do not apply, the coder can test for themore general notion of state. The suggested annotation label prefix – if applied – is ctx . • Temporal (tmp) – Conditions/Constraints associated with time - the when – Point in time (tim) – References to specific points in time ∗ Beginning (e.g., “from 1st January”) ∗ End (e.g., “until 31st January”) – Time frame (tfr) – References to time frames – Frequency (frq) – References to frequencies (e.g., “annually”) • Spatial (spt) – Conditions/Constraints associated with spatial representations – the where – Location (loc) – References to specific locations ∗ Beginning ∗ End – Direction (dir) – References to directions, inclusion of intermediary locations (e.g., “via”) – Path (pth) – References to pathways (e.g., “through the valley”) • State (ste) – References to a specific state (e.g., “during childhood”); note that state is moregeneral than temporal and spatial specification. Where possible, a more specific annotation shouldbe chosen. – Beginning – End • State transition (tra) – References to a change in state (e.g., “when traffic light switches fromred to green”)
IG 2.0 Codebook Version: 1.0 94 Taxonomies • Procedural order (prc) – Conditions/Constraints associated with explicit or implied executionorder. Operationally, this can include expressions of input into the activity identified in theinstitutional statement (e.g., “with input from . . . ”). Procedural order can further include therequired actions, de/activation of other statements or compliance with or violations of statements,respectively. • Method (met) – Conditions/Constraints associated with means or method by which an action isperformed – Means – Action as method (e.g., “by handshake”) – Instrument – Artefact as method (e.g., “by car”) • Purpose/Function (pur) – Conditions/Constraints describing the purpose or intent of an aim;generally output of action (e.g., “for the purpose of reducing pollution levels”) • Observed state/Outcome/Effect (eff) – Conditions/Constraints describing the outcome or effectof an aim by an actor involved in the action situation – a change in the environment emanatingfrom the observed actor(s); observation of compliance/non-compliance (e.g., “if an non-compliantaction is observed”; “if participant fails to meet certification requirements”). This characterizationis commonly invoked when coding discretionary actions of observers (e.g., beliefs, suspicions,evaluations).
The physical taxonomy differentiates between animate and inanimate entities, maintaining compatibilitywith annotation conventions commonly adopted in datasets coded according to the previous IG codingguidelines. The suggested annotation label prefix is type . • Animate – Living entities • Inanimate – Non-living entities, both real and mental constructs • Goal – Goal is a specialization of an inanimate mental construct
The role taxonomy serves the annotation of attributes and objects with additional labels to capturetheir role within a statement structure with respect to the action. The suggested annotation label prefixis role . • Originator/Causer/Agent – Entity from which action originates • Recipient – recipient of an artefact/sanction • Possessor – owner of an object/entity (e.g., “house owner”) • Experiencer – observer of action (e.g., “observer of non-compliance”) • Beneficiary – beneficiary of action; may not necessarily be action/artefact recipient (e.g., “welfarerecipient”) • Position – organisation or institutional role assumed by involved actor
IG 2.0 Codebook Version: 1.0 95 Taxonomies
Some types of syntactic annotations can aid the coder in discerning and capturing information thatsignals the broader function of institutional statements, as indicated by components of which they arecomprised, referred to as “institutional function.” Institutional functions facilitate the annotation ofaims in order to capture the correspondence of aims to analytical functions of relevance through specifictheoretical lenses. Exemplifying the use of the institutional grammar for the analysis from a regulatorycompliance perspective, compliance and violation behaviour is of specific concern, whereas institutionallife cycles may require the annotation of action verbs signalling the initiation of termination of insti-tutional arrangements. Note that the offered taxonomy provides examples for institutional functionsorganized by categories (alongside potential specializations), but does not claim exhaustiveness. Thesuggested annotation label prefix is function . • Compliance action – action reflecting compliance behavior – Comply – action reflecting compliance – Violate – action reflecting violation • Monitor – action reflecting the institutional function of monitoring – Detect compliance – action reflecting the detection of compliance – Detect violation – action reflecting the detection of violation • Enforce – action reflecting enforcement acts – Reward – action reflecting rewarding behaviour (regulative-incentivizing) – Sanction – action reflecting sanctioning behaviour (regulative-punitive) • Enforcement response – action reflecting responses to enforcement outcomes – Accept – action reflecting acceptance of enforcement outcome – Reject – action reflecting rejection of enforcement outcome ∗ Appeal (specialization of reject) – action reflecting appeal against enforcement outcome • Process – Life cycle – Initiate – Interrupt – Resume – Conclude • Transaction – action reflecting a request and corresponding response – Request – action reflecting a request. – Response – action reflecting a response to a request. Central difference to enforcementresponses is lack of a regulatory compliance function. ∗ Accept ∗ Reject • Decide – action reflecting a decision/discretion • Inform – action reflecting information dissemination
IG 2.0 Codebook Version: 1.0 96 Taxonomies • Declare – action reflecting change in role, position, environment, institutional fact • Assign – action reflecting the assignment of responsibilities to other actors – Delegate – action reflecting the delegation of functions/tasks, generally downwards in anorganisational structure (e.g., delegation to subordinate) – Elevate – action reflecting the elevation of functions/tasks, generally upwards in an organi-zational structure (e.g., elevation to supervisor)
While institutional function characterizations map the diverse operational expression of institutionally-relevant, the constitutive function annotations emphasize the specific role a constitutive function en-tertains with respect to the constituted entity and/or the linkage of constituted entity and constitutingproperties. Generally, the top-level distinction is the identification of the constituted entity as eitheran entity established or referenced in the context of the policy document, or the policy itself. On amore fine-grained level, the categories capture the role of the constituting function with respect tothe constituted entity (i.e., a specific entity, or the policy). The suggested annotation label prefix is confunc .The structure, alongside specific annotations, is visualized in Figure 16 and described in the following.Figure 16: Constitutive FunctionsWhere entities subject to the policy (or introduced by policy) are of concern, constitutive functionscan capture the definition of constituted entities as relevant for the parameterization of the institutionalsetting, including actors, object/artefacts, role specifications or actions. This definition is often signaledin the form of intensional definitions (e.g., explicit definition), or implied by ascription (e.g., implicitcharacterization of entity based on behavior).In addition to the definition of entities, constitutive functions can capture composition relationships(e.g,. specifying the composition of committees) and further reflect hierarchical relationships (e.g.,leadership structures, embedding positions within organizations).
IG 2.0 Codebook Version: 1.0 97 Conclusion
Further characterizations include the initiation and termination of entity lifecycles (e.g., dates oftermination), and the explicit conferral of institutional power, such as authority or rights to entities(e.g., authority to enforce, the right to vote).Complementing the perspective on entity specification as part of constitutive statements, we furtheridentify constitutive functions that characterize the policy or document itself, as opposed to focusingon the entities central to the representation of the institutional setting. Typical characterizationsinclude the policy lifecycle (e.g., date of enactment), as well as its relationship to other policy (e.g.,amending or superseding it). Further statements refer to the purpose or intent underlying a given policy.Informational statements offer supplementary information about the document, or state institutionalfacts contextualizing the policy or domain of concern.As with the preceding taxonomies, the constitutive functions taxonomy is subject to further refinementbased on ongoing empirical validation efforts.
The listing of taxonomies complements and concludes the coding guidelines for the Institutional Gram-mar 2.0. The codebook initially outlines the theoretical concepts underlying IG 2.0, including the codingon multiple levels of expressiveness (IG Core, IG Extended, IG Logico) in Section 2, the choice of whichis subject to analytical objectives. Following the discussion of essential document preparation steps(pre-coding steps) in Section 3, detailed coding guidelines are provided for regulative and constitutivestatements across all IG 2.0 levels (Section 4). This is followed by advanced concepts, such as the dis-cussion of encoding hybrid institutional statements (Section 4.4), i.e., institutional statements consistingof both regulative and constitutive components, polymorphic institutional statements (Section 4.4.4),and the discussion of structural patterns (Section 4.4.5). The listing of taxonomies in the previoussection (Section 5) concludes the substantive part of the codebook.It is important to note that the guidelines provided in this codebook are of general nature and em-phasize the operational use of IG 2.0. Doing so, they may not capture specifics potentially relevant fora given project. Instead, many aspects of the coding guidelines are of suggestive nature to inform thedevelopment of project-specific guidelines that consider application context (e.g., domain, language,types of documents, legal traditions, etc.) and analytical objectives (i.e., evaluation of encoded state-ments) more explicitly. The instructions provided here are further tool-agnostic, and open to adaptationfor arbitrary encoding means (e.g., Excel sheets, text annotation tools, etc.).For supplementary information, both including a theoretical treatment of the underlying conceptsand principles, as well as resources that support operational coding (e.g., videos, tool-specific guidance,software), please refer to https://institutionalgrammar.org/resources/ .Please further note that these guidelines will be continuously refined based on theoretical devel-opments, feedback from users as well as ongoing empirical validation efforts. To retrace subsequentchanges, please note the specific version and version history of these guidelines outlined at the beginningof this document. Irrespective of refinements, all revisions of the codebook will be retained for futurereference.
References
Brady, U., Basurto, X., Bennett, A., Carter, D. P., Hanlon, J., Heikkila, T., Lien, A., Chonaiew, S. M.,Olivier, T., Schlager, E., Siddiki, S., and Weible, C. (2018). Institutional Analysis of Rules-In-FormCoding Guidelines. Technical report, Center for Behavior, Institutions and the Environment.Crawford, S. E. S. and Ostrom, E. (1995). A Grammar of Institutions.
American Political ScienceReview , 89(3):582–600.
IG 2.0 Codebook Version: 1.0 98eferences
Crawford, S. E. S. and Ostrom, E. (2005). A Grammar of Institutions. In
Understanding InstitutionalDiversity , chapter 5, pages 137–174. Princeton University Press, Princeton (NJ).Frantz, C., Purvis, M. K., Nowostawski, M., and Savarimuthu, B. T. R. (2013). nADICO: A nestedgrammar of institutions. In
Lecture Notes in Computer Science , volume 8291 LNAI, pages 429–436.Siddiki, S., Weible, C. M., Basurto, X., and Calanni, J. (2011). Dissecting Policy Designs: An Applica-tion of the Institutional Grammar Tool.
The Policy Studies Journal , 39(1):79–103., 39(1):79–103.