logic chapt 11

Foundations of Fuzzy Logic and Fuzzy Set Theory

  • Shaggy Logic and Human Thought:

    • Martin Gardner observed: "Shaggy logic underlies both ordinary language and most human thinking."
    • Ordinary human discourse and intuitive reasoning are fundamentally grounded in imprecise, non-binary structures rather than rigid two-valued systems.
  • Fuzzy Logic versus "Fuzzy Math":

    • Public figures and authorities are frequently criticized for "fuzzy math" when their data are dubious, unverified, or arrived at through suspect methodology.
    • In colloquial usage, "fuzzy math" is a disparaging label questioning the accuracy of data or the soundness of computational methods.
    • Fuzzy logic is not "fuzzy" in the colloquial sense; it is a rigorous, precise logic designed to model fuzzy concepts, rather than being an imprecise or defective system of logic itself.
    • Fuzzification systematically accounts for the imprecision of empirical data, the natural vagueness of human language, and the inherent uncertainty present within complex systems.
  • Boolean Logic versus Shades of Gray:

    • Standard two-valued Boolean logic is restricted to systems characterized by exactly two operational states: true or false, off or on, black or white.
    • Fuzzy logic serves as an extension of many-valued logic, permitting propositions to occupy intermediate states representing "shades of gray."
    • Truth values are assigned across a continuous interval rather than a binary dichotomy:
    • A completely true proposition is assigned a truth degree of 11.
    • A completely false proposition is assigned a truth degree of 00.
    • Propositions that are neither 100%100\% true nor 100%100\% false receive a real number within the closed interval [0,1][0, 1], corresponding precisely to their degree of truth.
    • The closer a proposition is to absolute truth, the closer its truth degree approaches 11; conversely, the closer it is to absolute falsehood, the closer its truth degree approaches 00.
    • Relative truth values allow certain assertions to be judged truer than others: for example, the propositions "Albert Einstein is (or was) smart" and "Bart Simpson is smart" may both be regarded as true, but one possesses a higher degree of truth than the other.
  • Crisp, Fuzzy, and Vague Concepts:

    • Concepts within formal deductive logic are "crisp," functioning like the discrete on/off switches of an electrical circuit.
    • In ordinary human cognition, the majority of concepts are fuzzy rather than crisp, and these concepts are manipulated effortlessly in daily reasoning.
    • When a concept exhibits extreme fuzziness, it is formally designated as vague.
    • Martin Gardner demonstrated this distinction through contrasting statements:
    • "Bob will be back in a few minutes" represents a fuzzy proposition.
    • "Bob will be back sometime" represents a vague proposition.
  • Origins of Fuzzy Set Theory:

    • Fuzzy logic parallels and emerges from fuzzy set theory, just as classical Boolean logic corresponds to traditional crisp set theory and categorical class membership.
    • The formal framework of fuzzy set theory was pioneered in the United States by Lotfi Zadeh in 19651965.
    • Zadeh introduced fuzzy set theory as a rigorous mathematical apparatus to model and quantify the imprecision encountered in real-world phenomena.
    • In classical set theory, an element either belongs or does not belong to a given set; in fuzzy set theory, elements are defined by a continuous degree of membership.
    • Consequently, propositions in fuzzy logic are evaluated based on their degree of truth, allowing intermediate truth values between absolute truth and absolute falsehood.

Fuzzy Sets, Membership Degrees, and Truth Values

  • Quantifying Membership within a Universe of Discourse:

    • Natural language predicates such as "tall" or "smart" do not have objective binary boundaries: there is no single height threshold that definitively separates "tall" individuals from "not tall" individuals.
    • Let the universe of discourse be the set of "tall people": an individual Mary is assigned a numerical membership degree between 00 and 11 based on her physical height:
    • Complete membership is denoted by a value of 11, while absolute non-membership is denoted by 00.
    • If Mary is 7 feet7\,\text{feet} tall, she is unequivocally a member of the set of tall people; her membership degree is 11, the proposition "Mary is tall" is completely true, and ordinary Boolean logic governs the statement.
    • If Mary is 4 feet4\,\text{feet} tall, she is unequivocally not a member of the set; her membership degree is 00, the proposition "Mary is tall" is completely false, and Boolean rules again apply.
    • If Mary is 5′ 5′′5'\,5'' tall (1.651 meters1.651\,\text{meters}), the proposition "Mary is tall" holds a partial truth degree, and Mary possesses an intermediate degree of membership in the set of tall people, quantified as a value strictly between 00 and 11.
  • Valuation of Compound Propositions:

    • Compound propositions in fuzzy logic are evaluated by analyzing the specific truth degrees of their constituent atomic propositions.
    • The logical connectives (such as conjunction, disjunction, and negation) are computed using symbolic operators adapted from standard logic to operate across continuous values between 00 and 11.
  • Interpretations of Truth Degrees:

    • There is no single universal interpretation of what a truth degree represents; the semantic meaning depends directly on the specific domain of application.
    • When modeling natural language concepts such as tallness or smartness, a truth degree frequently quantifies a degree of similarity to an idealized prototypical standard.

Probability versus Degree of Membership

  • Historical Anticipation by Leibniz:

    • As early as 17041704, Gottfried Leibniz proposed using mathematical probabilities to quantify degrees of truth.
    • Leibniz noted that Aristotle had previously maintained that valid inferences could be derived from the probable, specifically in his Topics.
    • Leibniz wrote: "Opinion, based on the probable, also deserves perhaps the name knowledge; otherwise nearly all historical knowledge and many other kinds will fall. But without quarreling over names, I hold that the investigation of degrees of probability would be very important, that we are still lacking in it, and that this lack is a great defect of our Logic. For when we cannot decide a question absolutely, we might still determine the degree of likelihood from the data, and can consequently judge reasonably which side is the most likely. I have more than once said that we should have a new kind of Logic which would treat of degrees of probability, since Aristotle in his Topics has done nothing less than that."
  • Conceptual Distinction between Membership and Probability:

    • While probabilities can be mathematically applied to evaluate degrees of membership, the two concepts are fundamentally distinct.
    • Probability deals with the likelihood of outcomes in the presence of incomplete information or stochastic events.
    • Fuzzy logic deals with degrees of truth based on the similarity of an entity to an imprecisely defined concept.
  • The James Bezdek Thought Experiment:

    • Fuzzy expert James Bezdek illustrated the divergence between fuzzy membership and statistical probability through the thought experiment of two bottles of liquid offered to an individual who has spent a week in a blazing hot desert without water:
    • Bottle A is assigned a degree of membership of 0.900.90 in the class of potable liquids.
    • Bottle B is assigned an empirical probability of 0.900.90 of being potable.
    • Interpretation of Bottle A:
    • The value of 0.900.90 reflects Bottle A's high degree of physical similarity to an ideally pure, potable beverage (such as pure mountain spring water).
    • It guarantees that the liquid is predominantly pure and contains only minimal, non-lethal impurities (such as minor mineral sediment or flat taste).
    • Interpretation of Bottle B:
    • The value of 0.900.90 signifies that the liquid is entirely potable in 90%90\% of instances, but carries a 10%10\% chance of being entirely non-potable or deadly poison.
    • For instance, Bottle B could be a bottle chosen at random from a crate containing 1010 bottles, where 99 bottles contain pure spring water and 11 bottle contains pure poison.
    • Decision Outcome:
    • A rational individual will invariably choose Bottle A, because Bottle A guarantees safe hydration with minor imperfections, whereas Bottle B entails a 10%10\% probability of immediate death.

Practical Applications and Theoretical Limitations of Fuzzy Logic

  • Industrial and Commercial Applications:

    • Fuzzy Expert Systems: Represent the most widespread implementation of fuzzy logic, modeling human commonsense reasoning and handling vague qualitative information in:
    • Medical and system diagnosis
    • Digital image processing
    • Pattern recognition
    • Financial modeling and economic systems
    • Complex data analysis
    • Fuzzy Controllers: Specialized devices designed to execute automated physical adjustments in mechanical, chemical, and electrical systems:
    • Conventional examples include domestic thermostats and automotive cruise control systems.
    • Japanese manufacturing extensively incorporates fuzzy controllers into commercial and consumer hardware, including refrigerators, washing machines, television camcorders, air conditioning units, palm-top computers, automobiles, cameras, industrial robots, and high-speed bullet trains.
  • Theoretical and Practical Difficulties in Everyday Reasoning:

    • Incomparability of Natural Language: Not all statements in natural language can be systematically compared on a unified scale of truth degrees.
    • Context Sensitivity: The exact truth degree assigned to an identical proposition fluctuates dramatically based on surrounding environmental and social contexts:
    • For example, the statement "The dress is red" may be assigned a truth degree of 70%70\% on a redness scale when viewed against the neutral background of an all-white room.
    • The same dress may receive a truth degree of only 20%20\% when placed on a display rack surrounded by deeply vibrant, purely red garments.
    • Determination of Truth Degrees: If empirical propositions exist along an infinite continuum of gray rather than binary true/false values, there is no universally objective mechanism for assigning exact numerical values.
    • Breakdown of the Law of Noncontradiction: In classical logic, the law of noncontradiction states that a proposition cannot be both true and false simultaneously: ¬(P∧¬P)\neg (P \wedge \neg P).
  • Under fuzzy logic, this foundational law is violated: an individual can be simultaneously tall (to a degree of 0.60.6) and not tall (to a degree of 0.40.4).

  • This creates substantial theoretical tension when attempting to separate counterintuitive valid assertions from genuine logical fallacies.

Classification and Anatomy of Fallacies

  • Aristotle's Foundations in On Sophistical Refutations:

    • Aristotle opened On Sophistical Refutations with the foundational observation: "There exist both reasoning and refutation which appear to be genuine but are not really so."
    • Prompted by the deceptive rhetorical tactics of the Sophists, Aristotle sought to systematically catalog invalid argumentative patterns to instruct thinkers on how to detect, analyze, and refute deceptive reasoning.
    • Aristotle divided fallacies into two overarching taxonomic classes:
    • Fallacies dependent on language (linguistic fallacies).
    • Fallacies independent of language (substantive or genuine fallacies).
    • An early translator of Aristotle noted:
    • "Some of these fallacies would hardly deceive the most simple minds; others, which Aristotle seems to have been the first person to expose and define, are capable not only of deceiving the innocent but also of escaping the notice of arguers who are employing them."
  • Fallacies Dependent on Language:

    • These arguments are fallacious purely because the linguistic phrasing generates a deceptive semblance of logical structure where none exists.
    • They fundamentally depend upon ambiguity, grammatical vagueness, and equivocation (the exploitation of words with double meanings or terms that shift definitions across clauses).
    • The Ham Sandwich Syllogism: Demonstrates linguistic equivocation through a series syllogism:
    • Premise 1: A ham sandwich is better than nothing.
    • Premise 2: Nothing is better than eternal happiness.
    • Conclusion: Therefore, a ham sandwich is better than eternal happiness.
    • Structural Defect: The term "nothing" is used equivocally; in Premise 1, "nothing" signifies a quantitative value of zero or absence of a meal, whereas in Premise 2, "nothing" functions as a universal negation meaning "it is not the case that any existing thing is better than eternal happiness."
  • Fallacies Independent of Language:

    • Designated as genuine logical fallacies because their invalidity does not depend upon lexical ambiguity, but rather violates the structural laws of formal deduction.
    • In syllogistic systems, these are classified as non sequitur arguments (inferences where the conclusion fails to follow logically from the premises).
    • Across categorical logic, 256256 distinct combinations of standard categorical propositions (AA, EE, II, and OO) can be structured across the four classic figures into three-line syllogisms, yet only a minuscule fraction constitute logically valid moods.
    • Standard named formal syllogistic fallacies include:
    • Fallacy of the undistributed middle
    • Fallacy of illicit major
    • Fallacy of illicit minor
    • Fallacy of drawing a positive conclusion with a negative major premise
    • Fallacy of drawing a positive conclusion with a negative minor premise
    • Fallacy of drawing a negative conclusion from positive premises

Common Informal and Formal Fallacies

  • Petitio Principii (Begging the Question):

    • The fallacy of assuming the very premise or proposition that the argument is intended to demonstrate.
    • Occurs when an arguer attempts to substantiate a non-self-evident claim by subtly smuggling the conclusion into the underlying premises, typically disguising the circularity through alternative phraseology.
  • Argumentum Ad Hominem:

    • An informal fallacy wherein an arguer bypasses the substantive logical merits, data, and structure of an opposing claim to instead launch an attack against the opponent's personal character, background, or circumstances.
  • Appeals to Force, Multitude, and Unqualified Authority:

    • Argumentum ad Baculum (Appeal to Force): Coercively bases the acceptance of a conclusion upon physical, psychological, or political threats rather than evidence.
    • Argumentum ad Populum (Appeal to the Multitude): Validates a conclusion based purely on its widespread popularity, cultural prevalence, or emotional resonance with the masses.
    • Argumentum ad Verecundiam (Appeal to Authority): Defends an assertion by citing an authoritative figure who lacks legitimate expertise, credentials, or domain relevance in the specific subject under dispute.
  • Ignoratio Elenchi (Argument from Ignorance) and Red Herrings:

    • In its traditional definition, ignoratio elenchi denotes an ignorance of the specific proposition that needs to be established against an interlocutor.
    • Modern formulations often manifest as an illicit shift in the burden of proof, maintaining that:
    • A proposition must be true because it has not yet been proven false.
    • A proposition must be false because it has not yet been proven true.
    • Frequently overlaps with the "red herring" fallacy, in which an irrelevant topic is deliberately introduced into discourse to derail attention from the core thesis under examination.
  • The Fallacy of Conversion:

    • Identified and documented by Aristotle as early as 350 B.C.350\,\text{B.C.}, who observed:
    • "The refutation connected with the consequent is due to the idea that consequence is convertible. For whenever, if A is, B necessarily is, men also fancy that, if B is, A necessarily is."
    • John Stuart Mill observed in 18741874 that the illicit conversion of a universal affirmative proposition ("All AA are BB, therefore all BB are AA") and the illicit conversion of conditional propositions are among the most persistent errors in human discourse.
    • Explanations for why individuals repeatedly fall victim to illicit conditional conversion include:
    • The cognitive strain of processing abstract symbols in experimental settings.
    • Intrinsic processing difficulties associated with linguistic negation.
    • The unwarranted insertion of implicit auxiliary assumptions.
    • The common cognitive tendency to misinterpret a unidirectional conditional statement ("If PP, then QQ") as a bidirectional biconditional statement ("PP if and only if QQ").

Psychological, Linguistic, and Cognitive Dimensions of Logical Error

  • Peirce's Psychological Perspective:

    • Charles Sanders Peirce argued that "faulty thinking" in formal logic might properly belong to the investigative domain of experimental psychology rather than pure normative logic.
  • Mary Henle's Deductive Framework (1962):

    • Mary Henle demonstrated that deductive errors made by ordinary individuals do not prove that human beings are fundamentally irrational or incapable of logic.
    • Henle noted that formal logical structures (such as valid syllogisms) represent normative ideals—meaning "how we ought to think"—rather than descriptive accounts of real-time cognition.
    • Humans generally possess intact internal inferential systems capable of executing valid deductive transitions from one proposition to another; however, the "effective premises" mentally utilized by the reasoner frequently diverge from the premises originally supplied by the investigator.
    • Henle identified four specific cognitive processes that disrupt deductive performance:
    • Failure to Accept the Logical Task: The inability or refusal to isolate formal logical validity from factual accuracy. Reasoners systematically conflate whether an argument is formally structurally sound with whether the conclusion matches empirical truth.
    • Restatement of Premises or Conclusions: Unconscious alterations made to the wording of an argument that warp the intended logical structure.
    • Omission of Premises: Systematically dropping key premises during internal processing.
    • Introduction of Outside Knowledge: Smuggling external common sense, personal beliefs, and background empirical knowledge into the deductive sequence as unstated auxiliary premises.
  • Experimental Testing of Deductive Resistance:

    • Later studies introduced direct incentives—including monetary payments for correct logical answers—to induce subjects to attend strictly to the formal validity of deductive tasks.
    • Even when financial rewards were introduced, experimental subjects continued to fail the tasks at identical rates, demonstrating that subjects do not willfully refuse the logical task, but rather lack the cognitive clarity to grasp what formal logical validity entails.
  • Everyday Common Sense versus Formal Validity:

    • In colloquial dialogues, reasoners default to practical logic and unstated probabilistic assumptions, as demonstrated in two test scenarios posed to a university colleague:
    • Scenario 1: "All chairpersons must report to the meeting. You are not a chairperson. Do you have to attend?" The colleague answers: "No." (Fails the formal test of validity by invalidly denying the antecedent, relying on the unstated pragmatic assumption that non-chairpersons are excluded).
    • Scenario 2: "All math professors work hard. You are not a math professor. Do you work hard?" The colleague answers: "Yes." (The reasoner draws directly upon self-knowledge and personal work habits, entirely disregarding the logical independence of the conclusion from the premises).
    • Impact of Emotional Salience: Henle established that emotional involvement and strong personal convictions severely undermine logical objectivity: "The more personally relevant the material employed, the more difficult it will be to accept the logical task."
  • Form, Cognitive Load, and Memory Limitations:

    • Reasoning difficulty is dictated by the structural form of an inference and its required cognitive processing steps.
    • Elementary deduction schemes like modus ponens are readily understood even by young children, whereas other valid deductive forms confound adults.
    • Human working memory and available internal computational space are limited; when deductive proofs require retaining, updating, and retrieving multiple steps simultaneously, cognitive overload occurs.
  • Verification Bias versus Falsification:

    • In evaluating deductive arguments, human reasoners possess a strong cognitive verification bias: they actively seek scenarios where premises combine to affirm or verify the conclusion.
    • In contrast, rigorous formal logic demands falsification: the reasoner must systematically test memory and imagination to construct a counterexample wherein the premises are simultaneously true while the conclusion is false.
    • The presence of a single valid counterexample establishes that the inference is formally invalid; the total absence of any counterexample proves the inference is valid.
    • The Acorn and Football Example:
    • Syllogism 1: "All football players are strong. This man is strong. Is he a football player?"
      • Result: Solvers readily produce counterexamples from memory (e.g., weightlifters, construction workers, or personal acquaintances who are exceptionally strong but do not play football), prompting the correct formal response: "Cannot be determined."
    • Syllogism 2: "All oak trees have acorns. This tree has acorns. Is it an oak?"
      • Result: Solvers struggle to formulate a counterexample, because standard botanical reality and lexical definitions specify an acorn strictly as the seed or fruit of an oak tree.
      • Error: Rather than recognizing that the argument formally commits the fallacy of affirming the consequent (warranting only the conclusion "Cannot be determined"), most reasoners insist that the tree must be an oak, allowing semantic knowledge to override formal structural rules.
  • Aversion to Indeterminate Conclusions:

    • Psychological experiments demonstrate that individuals harbor a profound aversion to "can't tell," "cannot be determined," or "no conclusion is possible" outcomes.
    • Subjects strongly desire definitive, closed conclusions; unless explicitly trained, prompted, and encouraged to utilize "can't tell" responses, they will systematically commit the formal fallacies of affirming the consequent and denying the antecedent.
  • Four Core Sources of Deductive Error:

    • In their synthesis of critical thinking and reasoning research, Phares O'Daffer and Bruce Thornquist isolated four primary drivers of deductive failure:
    • Adding to, altering, or ignoring items from the given premises.
    • Allowing empirical factual content to supersede the logical inference pattern (traditional patterns of conversational discourse overriding formal logic).
    • Linguistic processing barriers, including the quantity and placement of negations, sentence length, lexical complexity, and severe cognitive overload.
    • Inability to