• UNIT IV

• Processes: Times, Events and Situations, Classification of processes,

Procedures, Processes and Histories, Concurrent processes,
Computation, Constraint satisfaction, Change Contexts: Syntax of
contexts, Semantics of contexts, First-order reasoning in contexts,
Modal reasoning in contexts, Encapsulating objects in contexts
 5.1 Context
• Complementary Descriptions: Different contexts allow
multiple descriptions of the same object.
• Example (Figure 5.2): Tom is a baby in 1976 and an adult in
1997, showing context-based descriptions.
• Person Classification: "Person" is continuous, but "Baby" and
"Adult" are stage-specific and not coreferent.
• Predicate Logic: Tom as a baby (1976) and an adult (1997)
are distinct, non-overlapping states.
• Opaque Contexts: Tom’s descriptions as a baby and an adult
exist in separate contexts and do not contradict each other.
• Conceptual Systems: Complementarity is a human
conceptual feature, not an inherent property of reality (e.g.,
wave-particle duality).
• Stages of Life: Life stages form a continuous process but are
often categorized in discrete steps, leading to
misunderstandings.
• Discontinuities: The shift between life stages can cause
confusion, as terms like baby, child, and adult imply distinct
identities.
5.1.1 Identity
• Mathematical Identity: In math, objects are always identical to themselves (e.g., x =
x).
• Heraclitus' View: Everything is in constant flux, challenging the idea of fixed identity.
• Plato & Aristotle: Abstract entities retain identity, while physical things change over
time.
• Ship of Theseus: Questions whether an object remains the same if all its parts are
replaced.
• Rifle Identity Example: The U.S. Army ties a rifle’s identity to its stock, the only
irreplaceable part, to prevent theft or confusion.

• 2. Identity in Databases:
• Identity in systems is managed through unique serial numbers or identifiers.
• This method of tracking identity relies on conventions agreed upon by authorities or
systems.
 5.1.2. Verbs as Connectors (Nexus):

• Peirce’s View: Verbs express relationships (Thirdness) between participants

(e.g., "chase" links a cat, a mouse, and the action).
• Whitehead’s View: Verbs act as a “nexus,” connecting multiple elements in a
relationship.
• Graph Representations:
• Top Graph: Directly shows the action of chasing.
• Middle Graph: Represents "chase" as a separate concept linking cat and mouse.
• Bottom Graph: Uses a context box to depict "chase" as a nexus, illustrating connections.
• Formal Representation: The three graphs correspond to different logical forms
of the "chase" action.
• Logical Variation: They show how abstract concepts and actions can be
modeled using different relations and formulas.
Identity: Depends on context and conventions.
Verbs as Connectors: Show relationships between entities.
Representation: Relationships can be modeled in different ways.
.
• Chase Formula: Defines a cat (x) as the agent, a mouse (y) as the
theme, and the chase (z) as the action linking them.

• Moving Information: Shifting information from a nested graph moves

relationships (like "Has") outside, linking elements directly.

• Conceptual Relations & "Has": Every relationship (e.g., "agent,"

"theme") can be defined using the "Has" relation.

• Example: The "Agnt" relation links an act to an animate being,

indicating that the act has an agent.
 5.1.3. Defining Verbs with
Purpose:
• Defining Verbs with Purpose:
• Chase Definition: Involves an animate being (e.g., a cat) rapidly following a mobile
entity (e.g., a mouse) with the intent to catch it.
• Purpose: Links the agent (cat), action (chasing), and intended outcome (catching).
• Context Box: Separates the act of chasing from the goal (catching), as they occur in
different situations.
• Rules and Inferences:
• Time Sequence: Chasing occurs before catching.
• Contingency: Catching is not guaranteed; the chase may fail.
• Success or Failure: The chase succeeds if the cat catches the mouse; otherwise, it fails.
Complexity of Verbs: Verbs like "chase" involve an agent, a theme, and a purpose.
Logical Structure: They can be broken down into logical relations and rules.
Interaction Rules: Define how the agent, theme, and purpose interact within different
contexts.
• Axioms & Chasing: If successful, the agent catches the theme; failure
is possible.
• Word Senses: Words have multiple meanings.
• Example: ChaseAway differs from chase—the goal is to drive away,
not catch.
• Indexicals in Context: ChaseAway means the agent forces the theme
to leave a specific place (#here).
• Context Dependence: Terms like #here and #now depend on the
surrounding situation.
• Representation Challenge: Basic predicate calculus lacks syntax for
contexts but can be extended with special symbols.
• Key Idea: Word meanings vary by context and must be clearly
defined.
• Contextual Elements: Place and time need special handling.
• Formal Logic: Predicate calculus requires extensions to represent
context-dependent meanings
5.2 SYNTAX OF CONTEXT:

• 1. Meaning of Context:
• The word context has different meanings depending on the area of study
(linguistics, philosophy, artificial intelligence).
• Two Major Senses:
• Basic Meaning: The surrounding text or discourse that gives meaning to a
word or phrase.
• Derived Meaning: The situation, environment, or background related to the
subject of interest.
• 2. Ambiguity of Context:
• Context can refer to the text itself, what the text is about, the information it
provides, or its potential uses. The confusion comes from which of these aspects
is the focus.
• 3. Functions of Context:
• Syntax: Organizing or grouping parts of text.
• Semantics: Describing or referring to real or hypothetical situations.
• Pragmatics: The purpose or reason for using the text or referring to a
situation.
• Context in Computing & Logic:
• LISP: Uses the quote operator to treat code as data.
• Logic: Quoting blocks standard inference rules, enabling new
interpretations.
• Purpose: Context sets boundaries for interpretation.
5.2.1 PEIRCE'S CoNTEXTS
• Peirce’s Notation for Contexts:
• Graphical Notation: Ovals mark propositions within a context.
• Function: Separates levels of meaning, with outer statements
commenting on inner ones.
• Peirce’s Theory of Contexts:
• Represented logical statements using graphs.
• Focused on semantics (meaning) and pragmatics (purpose).
• Studied how information moves in and out of contexts.
• Three Basic Primitives in Existential Graphs (EGs):
• Existential Quantifier: Bars show variables.
• Conjunction: Two graphs in the same context mean "AND" (e.g., x∧y).
• Negation: An oval without lines represents negation (e.g., ¬p).
• Example – “If a farmer owns a donkey, then he beats it”:
• EGs: Two ovals for negations, identity lines link variables.
• CGs: Context as a concept, coreference links, explicit negation ("-").
• Translation into Logical Formulas:
• EG formula: ¬(∃x)(∃y)(farmer(x)∧donkey(y)∧owns(x,y)∧¬beats(x,y)
• CG formula: ¬(∃x:Farmer)(∃y:Donkey)(owns(x,y)∧¬beats(x,y))
• Key Difference Between EGs and CGs:
• EGs: Use lines of identity for existential quantifiers and connections.
• CGs: Separate functions—concepts (e.g., [Farmer], [Donkey]) represent variables,
and arrows show relations.

• Scroll and Implication:

• Scroll: Represents implication (if-then) in Peirce’s system.
• Symbol (~): Used for implication in formulas, requiring variables to be rewritten as
universal quantifiers.

• Additional Contexts – Shading and Tinctures:

• Shading/Colors: Represent different contexts (e.g., possibility, impossibility).
• Purpose: Distinguishes logical relations like If-Then and Possibility.
Discourse Representation Theory
(DRT)
• Hans Kamp's DRT: Developed Discourse Representation Structures
(DRSs) to map language to logic.
• EG & DRS Equivalence: Structurally identical, sharing primitives,
scoping rules, and logic translation.
• Example: Farmer-Donkey Sentence → Both DRS & EG map to the
same predicate calculus formula.
• PSN (Gary Hendrix): Uses overlapping contexts, unlike nested EG, CG,
and DRS structures.
 REsoLVING INDEXICALS
• Peirce’s Indexicals: Defined context-dependent references (e.g., he, here, now).
• Conceptual Graphs (CGs): Use # for general indexicals and qualifiers like #he, #it.
• Example (Farmer-Donkey Sentence):
• Left CG: Uses indexicals (e.g., he = #he, it = #it*).
• Right CG: Resolves coreference labels (he → ?x (farmer), it → ?y (donkey)).
• Predicate Calculus Limitation:
• Lacks indexical notation and explicit context structure.
• Translation only possible after resolving indexicals.
• Verb Representation:
• Own = state (experiencer (Expr) & theme (Thme)).
• Beat = action (agent (Agnt) & patient (Ptnt)).
• Patient is more affected than a theme in an action.
 Transformation Rules:
• Convert logical forms, e.g., "every" → implication.
• Example: Every farmer who owns a donkey beats it transforms into Figures 5.7, 5.9,
5.10 CG forms.
• Defined Quantifier (∀):
• Ranges over subtypes (e.g., farmers who own donkeys).
• Expands into primitive CG form.
• Negation of Possibility Rule (-Psbl):
• Defined as not possible if false that it is possible.
• Indexical Resolution:
• Requires context search for pronouns (you, him) and implicit objects (drink → liquid).
• Research Area:
• Some quantifier transformations are well known, but indexical resolution remains an active
research topic.
 CONVERSATIONAL IMPUCATUREs.
• Conversational Implicatures & Indexical Resolution
• Paul Grice's Observations:
• When no referent for an indexical exists, listeners assume missing information.
• Conversational implicatures help interpret unspoken background details.
• Grice’s Implicatures: Listeners assume missing info when referents are unclear.
• Indexical Resolution:
• You → Hypothetical person (x).
• He → Most likely horse (y).
• Liquid → Assumed water (z).
• Final CG: If x (person), y (horse), and z (water) exist, then x can lead y to z, but x
can’t make y drink z.
• AI & Logic: Resolving indexicals needs context + reasoning rules like
nonmonotonic logic.
 5.3 Semantics of Contexts
• William James' View (1897): Space-time has no intrinsic meaning; humans
create separate serial orders to make sense of it.
• Context Definition: A package of information about a bounded chunk of
the world.
• Situation Semantics (Barwise & Perry, 1983):
• Situations are manageable chunks of space-time, real or imaginary.
• Meaning flows from situations → statements → interpretation → action.
• Symbol grounding: Symbols gain meaning through triadic relationships (speaker →
concept → listener).
• Example (Figure 5.12):
• Situation Representation:
• Images (sound/picture) link to situations via image (Imag) relations.
• Description (Dscr) relation: Links a situation to a true proposition.
• Statement (Stmt) relations: Express a proposition in different formats
(English, CG, KIF).
• Predicate Calculus Representation:
• Maps a situation (s) to:
• Images (x = sound, y = picture) stored in files.
• Statements (p) in multiple formats:
• English: "A plumber is carrying a pipe."
• CG: [Plumber]—(Agnt)—[Carry]—(Thme)—[Pipe]
• KIF: (exists (?x plumber) (?y carry))
• summary:
• Contexts organize meaning by relating language to real-world situations.
• Situations structure meaning, grounding abstract symbols in reality.
• Multimedia representation (images, CGs, predicate logic) allows flexible
information retrieval and display.
McCARthy CONTEXTS.
• Predicate Calculus Representation
• Situation (s) Mapping:
• Images:
• x = sound (stored in file)
• y = picture (stored in file)
• Statements (p) in Different Formats:
• English: "A plumber is carrying a pipe."
• Conceptual Graph (CG): [Plumber]—(Agnt)—[Carry]—(Thme)—[Pipe]
• Knowledge Interchange Format (KIF): (exists (?x plumber) (?y carry))
Meaningful Situations

• Barwise & Perry (1983): Situations are bounded space-time regions.

• James' Critique: Arbitrary regions have disjointed events with no
meaning.
• Examples:
• College Lecture:
• Time shift (+30 min): Overlaps two lectures → unnatural.
• Space shift (half a room): Mixes two classes → worse.
• Sun-based coordinates: Moves into deep space.
• Complex Cases:
• Fiction (Sherlock Holmes): No real space-time region.
• U.S. Legal History: Intertwined with other events.
 Meaning-Preserving Translations
Propositions: Abstract meanings, expressed through statements in different
languages.
• Peirce (1905): Meaning = translation of a proposition into another form.
• Translation Criteria:
• Invertibility: Can map back without losing meaning.
• Proof Preservation: Sentences remain provably equivalent.
• Vocabulary Preservation: Key words remain or use synonyms.
• Structure Preservation: Logical form (∃, ∧, ¬) stays intact.
• Challenges:
• Natural Language: Ambiguous, harder to define strict equivalence.
• Computational Constraints: Efficient translation must be polynomial-time
computable.
Examples of Meaning-Preserving
Translations
• Meaning-Preserving Translations
• Two Logic Notations:
• L1L_1 (Peano): A,v,−,⇒,∃,∀A, v, -,
• L2L_2 (Peirce): X,+,−,−<,L,IIX, +, -, -<, L, II
• Examples:
• p∧q⇒pXq
• (∃x)p⇒′T..xp
• Preservation Criteria:
• Invertible: Can map back accurately.
• Proof-Preserving: Equivalent formulas remain provable.
• Vocabulary & Structure-Preserving: Core meaning and logic remain
unchanged.
• Refining Translations:
• Sorting (f1f_1) – Orders conjunctions.
• Renaming (f2f_2) – Standardizes variables.
• Deleting Duplicates (f3f_3) – Removes redundancy.
• Replacing Synonyms (f4f_4) – Maps terms to definitions.
• Example (Leibniz's Theorem):
• Original: ((p⇒r)∧(q⇒s))⇒((p∧q)⇒(r∧s))
• Peirce Normal Form: Simplifies 864 equivalent forms into one.
• Key Idea:
• Efficient transformations preserve meaning and simplify logic without
recursion.
 Meaning in Natural Languages
• Chomsky: Transformations affect quantifiers & pronouns; later theories
use constraints.
• AI Approaches:
• Schank: Conceptual dependency (11 actions).
• Shapiro: Semantic networks for logic-based meaning.
• Implicit Meaning:
• Eased concerns → Inflation not rising.
• Heightened concerns → Inflation rising.
• Needs background knowledge (Cyc Project).
• Key Idea:
• Meaning depends on syntax, semantics, pragmatics, and knowledge.
Tinctured Existential Graphs (Peirce, 1906)

• Purpose: Used colors (tinctures) to distinguish different contexts in logic.

• Three Context Types:
• Actual Context:
• Represents truth about the world.
• White (argent) for general truth; other metals for specific truth.
• Modal Context:
• Represents possibilities relative to actual events.
• Colors: Blue (logical/subjective possibility), Red (objective), Green (interrogative), Purple (freedom).
• Intentional Context:
• Describes an agent’s intention about reality.
• Furs: Gray (necessitated), Yellow (intent), Brown (commanded), Orange (compelled).
• Key Idea:
• Tinctures are meta-level markers, guiding how logic applies to discourse, not affecting
logical operators like ∧, ¬, ∃.
 Classifying Contexts

• Tinctured Contexts:
• White (Argent): Standard FOL truth.
• Other Metals: Specialized FOL (e.g., physical, mathematical truths).
• Modal Contexts:
• Dark Blue: Logical possibility.
• Light Blue: Belief-based (epistemic logic).
• Red: Physical possibility.
• Green: Query mode (Prolog/SQL).
• Purple: Freedom/obligation (deontic logic).
• Intentional Contexts:
• Hope(a, p) = Fear(a, ¬p).
• Ambivalence(a, p) → Contradiction (p & ¬p).
• Key Idea: Peirce’s system classifies contexts but remains incomplete.
 5.4 First-Order Reasoning in
Contexts
• 1. Contexts in Logic
• Syntax: Contexts enclose propositions.
• Semantics: An agent asserts, believes, or assumes the proposition.
• Pragmatics: Contexts separate statements from meta-statements (e.g.,
beliefs about statements).
• 2. Soundness & Completeness
• Soundness: If provable, it is true.
• Completeness: If true, it is provable.
• Gödel’s Theorem: Higher-order logic is incomplete—some true
statements are not provable.
• 3. Import-Export Rules
• Move information in/out of contexts while preserving truth.
• Three Contexts:
• Actual: Physical or mathematical truth.
• Modal: Possible vs. necessary truths.
• Intentional: Agents’ beliefs about possible worlds.
• 4. Peirce’s Rules of Inference
• Five Rules:
• Erasure: Remove info in positive contexts.
• Insertion: Add info in negative contexts.
• Iteration: Copy a proposition within a context.
• Deiteration: Remove redundant copies.
• Double Negation: Insert/remove ¬¬ freely.
• Supports First-Order Logic (FOL) more efficiently than Principia Mathematica.
6.Peirce’s Proof & Algebraic Notation
• Existential Graphs (EGs): Peirce’s 7-step proof for Leibniz’s theorem was far
simpler than Principia Mathematica's 43 steps.
• Algebraic Notation: Logical operators affect negation depth:
• Implication (⇒): ¬(p∧¬q)→ p +1, q +2 depth.
• Disjunction (∨): ¬(¬p∧¬q)→ +2 depth.
• Universal (∀x p): ¬(∃x¬p)→ +2 depth.

7.Cut-and-Paste Theorem: Any outer proof can be moved into a nested

positive context. P→Q, R→(P→Q), without changing its validity.
• Generalized Modus Ponens (GMP): Works for nested contexts.
• Hypergeneralized Modus Ponens (HMP): Moves partial implications across
contexts.
• With HMP, we can infer R, because we move P→Qinto the appropriate
context of (P→Q)→toR
• 8. Generalized Modus Ponens (GMP)
• Extends modus ponens to work within nested contexts.
• Example: If p ⇒ q and p is a specialization of q, then q holds in the nested
context.
• 9. Exporting Information from Contexts
• Hypergeneralized Modus Ponens (HMP): Moves partial info across
context boundaries.
• Qualified Export (QE): Exports statements as conditional (if-then)
propositions.
5.5 Modal Reasoning in Contexts
• 1. Peirce’s and Kripke’s Modal Logic
• Leibniz: Possible worlds define necessity (true in all worlds) and
possibility (true in some world).
• Peirce’s Approach:
• Graphical Notation: Used sheets of assertion for worlds.
• Algebraic Notation: Necessity as a universal quantifier over states of affairs.
• Kripke’s Worlds:

• Possible worlds (K) → One real world w0w_0w0​.

• Accessibility Relation (R) → Determines which worlds are connected.
• Evaluation Function (φ) → Determines truth value in a world.
• Modal Definitions:
• ◊p=(∃w:World)(R(w0,w)∧φ(p,w)=T)→ Possible.
• □p=(∀w:World)(R(w0,w)⇒φ(p,w)=T)→ Necessary.
• This states that p is possible if there exists at least one accessible world w
where p is true. Interpretation: "There is some possible world where p holds.“
• Example:
• "It is possible that I win the lottery" means there is some possible world
where I do.
Example: Sherlock Holmes' world may be accessible
from ours, but cartoon worlds are not.
• 3. Criticism of Possible Worlds
• Quine (1948): Criticized vague ontological status.
• Kripke’s model is useful but does not explain meaning.
• 4. Hintikka’s Model Sets
• Alternative to Kripke: Modal logic based on sets of propositions instead of
worlds.
• Similar to McCarthy’s contexts.
• 5. Dunn’s Laws and Facts
• Replaced worlds with "laws and facts" pairs ⟨M,L⟩
• Accessibility defined by law inclusion: R(w1,w2)=L1⊆M2
• Restated modal logic without assuming infinite possible worlds.
• 6. Completing the Pushout (Category Theory)
• Mathematical framework linking possible worlds to contexts.
• Pushout: Moves theories from worlds → contexts to make them
computable.
• 7. Situations as Pullbacks
• Reverse of pushout: Starts with contexts → situations.
• Allows real-world reasoning from logical descriptions.
• 8. Legislating Modalities
• Modalities arise from agents (lawgivers) legislating truths.
• Five Kinds of Possibility:
• Logical: Provable tautologies.
• Subjective: What an agent knows.
• Objective: Laws of nature.
• Interrogative: What is questioned.
• Freedom: What is permissible.
• 9. Exporting Modal Information
• Conditions must match (same modal, temporal, intentional status).
• Two key metalevel rules:
• Description Rule: If context describes xxx, any proposition about xxx can be
imported/exported.
• Communication Rule: Information from a trusted source can be exported.
• 10. Stratified Metalevels (Tarski’s Theory)
• Levels of language:
• L0→ Describes objects.
• L1​→ Describes truth of statements in L0L_0L0​.
• L2→ Describes proofs and models of L1L_1L1​.
• Prevents paradoxes like "This sentence is false."
• 11. First-Order Logic in Contexts
• Nested contexts preserve first-order reasoning.
• Modal logic can be replaced with metalevel reasoning.
• Contexts act as a "firewall" to prevent inconsistencies.
5.6 Encapsulating Objects in
Contexts
• 1. Object-Oriented Features in Logic
• Logic supports object-oriented concepts:
• Object class definitions vs. instances.
• Methods (procedures) vs. method executions (processes).
• Time-sequenced execution of methods.
• Message passing to trigger methods.
• 2. Encapsulating Object Classes
• Without contexts, object properties and behaviors mix with others.
• Encapsulation: Each class has a context with:
• Universal properties (e.g., every car has a model, engine, body, four wheels).
• Methods (e.g., StartEngine, Accelerate, TurnLeft).
• 3. Object Instances
• Surrogates (OIDs): Unique identifiers for instances.
• Example: A car instance PCXX999 at 9:51:02 GMT has:
• Model Mustang, engine S6F901T, body RJ88107.
• Engine not running, wheels not rotating.
• 4. Object Methods (Procedures vs. Processes)
• Procedures (in class) → Apply to all objects.
• Processes (in instance) → Execution of procedures.
• Zooming into StartEngine method:
• If-condition: Engine off, ignition turned on.
• Then-condition: After 5 sec, engine runs at 750 rpm.
• 5. Message Passing
• Messages trigger methods inside an object.
• Example: A message tells car PCXX999 to start the engine.
• Indexicals (#) → Used for unknown internal references (e.g., this in
Java).
• 6. Executing a Procedure (StartEngine Example)
• Step-by-step logical inference triggers execution:
• Export message contents to car’s context.
• Copy procedure definition from the class.
• Replace indexicals (e.g., #now → 9:51:02 GMT).
• Apply logical rules: Infer state changes over time.
• 7. Intermediate and Final State
• At 9:51:07 GMT:
• Engine now running at 750 rpm.
• Inference rules applied to move results into the car’s main context.
• 8. Using Multiple Notations
• CGs (Conceptual Graphs) can integrate other models:
• Finite-state machines.
• Petri nets.
• Predicate logic.
• Even specialized notations (e.g., music scores).
summary
• Encapsulation of objects in contexts keeps logic structured.
• Objects = Static properties + Dynamic processes.
• Message passing triggers actions using logical inference.
• Multiple formalisms (CGs, FSMs, Petri nets) can be combined.