SOFISMA. — In its primitive usage, *σόφισμα* (from *σοφία*, wisdom) is any invention or discovery of thought (*inventum*, *scium*). After dialectic, through plausible devices (cf. the examples cited by Aristotle in *De philosophicis elenchis*, ch. 4 ff.), was made to serve negative purposes by the Sophists (v. SOFISTI), the term *s.* also took on a pejorative sense and was used to denote an illogical (or false) argument that appears correct: “when it *seems* to prove, it is a *sophisma*, not a demonstration” (*ἐν δὲ σχί-νεται, σόφισμα ἔστει, οὐκ ἀπὸ δεῖξις*; *Top.*, VIII, 11, 162 a 14-15; cf. *De soph. el.*, 1, 165 a 21; and Sextus Empiricus, *Pyrrh. hypoth.*, II, 22, 229 ff.: “sophisms harm the truth by an appearance of plausibility”).
The further meaning assumed by the term, as an intentionally captious argument — whereby *s.* is sometimes distinguished from the simple *paralogism* or illogical argument — is also of Sophistic origin: this sense is not explicitly found in Aristotle, who equates *sophisma* and *paralogism* (*De soph. el.*, 1, 164 a 21), but is nevertheless contained in the description of *sophisma* as “a contentious argument” (*σόφισμα δὲ συλλογισμός ἐριστικός*; *Top.*, VIII, 11, 162 a 16-17; cf. 12, 162 b 3-5; cf. *Rhet.*, 1, 1355 b 17: “the sophist is such by deliberate choice,” *κατὰ τὴν προαίρεσιν*), and in the assertion that the sophist aims to profit from an apparent science (*De soph. el.*, 1, 165 a 22). The aforementioned meaning is, however, expressed by the Latin term *fallacia* (from *fallere*, to deceive), used in the Middle Ages as a synonym for *s.* (though already Seneca, following Cicero, rendered *sophisma* as *cavillatio*; cf. *Ep.*, 11, and Cicero, *Acad.*, IV, 24, 75: “sophismata appellantur fallaces conclusiunculae”; cf. also St. Augustine, *De doctr. christ.*, II, ch. 34, n. 48). There is finally a third meaning — though debatable — indicated by Lalande (*Vocab. techn. et crit. de la philos.*, 5th ed., Paris 1947, pp. 980-90), according to which *s.* is an argument that, from true premises or premises accepted as such, leads to impossible and paradoxical conclusions and yet appears irrefutable because logically correct: such, for example, according to some, are the arguments of Zeno or Kantian antinomies. The existence of such *sophisms*, according to Lalande, would show that the laws of logic “n'ont pas un champ d'application illimité” (*loc. cit.*).
A classification of the various forms of *s.* and a detailed exposition of methods to counter them is given by Aristotle in *De soph. el.*, ch. 4 ff. Aristotle distinguishes forms of *s.* deriving from expression or language (“secundum dictionem”) and equivocation in terms; ambiguity of phrasing; variation in accent; transition from the divided sense of a proposition to its composite sense and vice versa; similarity of expression (cf. ch. 4, 165 b 24 ff.); and forms extraneous to language (“extra dictionem”): attribution to the subject of what belongs only to one of its determinations, *fallacia accidentis*; taking in an absolute sense what is true only in a particular sense, *dictum simpliciter et dictum secundum quid*; ignorance or confusion of the issue, *ignorantia elenchi*; flaw in inference, *fallacia consequentis*; assuming as a premise what is to be demonstrated, *petitio principii*; confusion of what is not a cause with the cause, *non causa pro causa*; complex question demanding contradictory answers, posed as a single question: *sophisma plurium interrogationum ad modum unius* (cf. ch. 5, 166 b 28 ff.). Aristotle’s classification was drawn from the practice of ancient Sophistry. Among further forms of *s.* not listed in Aristotle’s logic, and from which modern scientific methodology has not always been exempt, the following should be noted: generalization based on insufficient induction (*sophisma inductionis*) or mere analogy (*sophisma analogiae*), and deductions from a hypothesis presented as certain truths (*sophisma hypothesis*).