Skip to main content

There is no universally agreed-upon classification of logical fallacies, and many existing categorizations contain overlapping concepts. After reviewing relevant literature and prior studies, we have identified the following 15 types of logical fallacies for labelling our data.

  • False Dilemma The presentation of an issue as having only two possible outcomes, either right or wrong, without recognising that additional alternatives may exist.
  • Equivocation The misleading use of a word or phrase that has multiple meanings, creating ambiguity and leading to confusion in interpretation or reasoning.
  • False Premise The establishment of an argument based on an unfounded, non-existent, or unreasonable assumption, leading to flawed reasoning or invalid conclusions.
  • False Analogy The assumption that if A and B share certain characteristics, then B must also possess other attributes of A, despite lacking a valid basis for this inference.
  • Wrong Direction The incorrect attribution of causality by reversing the cause-and-effect relationship, assuming the effect is the cause and the cause is the effect.
  • Fallacy of Composition The mistaken assumption that what is true for a part of something must also be true for the whole, disregarding the possible differences between individual components and the entire entity.
  • Begging the Question The use of a statement as both the premise and the conclusion, assuming the truth of what is to be proven instead of providing independent support.
  • False Cause The incorrect assumption that a causal relationship exists between two events solely because one follows the other, failing to account for coincidence or other influencing factors.
  • Inverse Error The mistaken reasoning that if A implies B, then not A must imply not B, overlooking the possibility that B may still occur due to other factors.
  • Improper Transposition The incorrect inference that if A implies B, then B must also imply A, failing to recognise that implication is not necessarily reversible.
  • Improper Distribution or Addition The erroneous reasoning that individual effects can be directly summed or distributed across a group without considering their actual impact or interaction.
  • Contextomy The act of selectively quoting or altering a statement, advertisement, or published material in a way that distorts its original meaning, often misrepresenting the intent of the original source.
  • Nominal Fallacy The mistaken interpretation of a metaphorical or figurative expression as a literal statement, leading to a misunderstanding of its intended meaning.
  • Accident Fallacy The misapplication of a general rule to a specific case where exceptions should be considered, treating the rule as absolute without regard for context or relevant circumstances.
  • Self Contradiction Te statement that directly negates its own truth rather than circularly assuming its conclusion as a premise.

The distribution of fallacies in the shitty-advice dataset is: