{
  "id": "verisimilitude",
  "title": "Verisimilitude",
  "lang": "en",
  "source_lang": "en",
  "summary": "Verisimilitude is truthlikeness: the extent to which a statement or theory approximates reality, or, in the arts, the appearance of plausibility.",
  "category": "philosophy",
  "subcategory": "metaphysics-epistemology",
  "type": "concept",
  "aliases": [
    "Truthlikeness",
    "Truth-likeness"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/verisimilitude/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/verisimilitude/",
      "json": "https://aiwiki.page/zh/verisimilitude.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Verisimilitude means likeness to truth. In philosophy of science, it is the degree to which a statement, model, or theory approximates the truth, even if it is not wholly true. The concept makes it possible to ask whether one false theory describes reality better than another. In aesthetics and literary criticism, the term usually refers instead to the appearance of truth or plausibility: a representation’s capacity to seem credible within its setting and conventions. These meanings must be distinguished, because something can appear convincing without accurately representing reality."
    },
    {
      "heading": "Truth, truthlikeness, and probability",
      "text": "In its scientific and philosophical use, verisimilitude concerns the relationship between a representation and the world. It is closely associated with truth, but it does not simply replace a true-or-false distinction with degrees of truth.\n\nUnder a classical interpretation, a factual proposition is either true or false. Nevertheless, two false propositions can differ in how closely they describe the facts. If a rod is exactly 100 centimetres long, the claims that it is 99 centimetres long and that it is 20 centimetres long are both false when interpreted as exact statements. On an ordinary measure of numerical error, the first is closer to the truth. A theory of verisimilitude seeks to extend such comparisons to more complex descriptions and scientific theories.\n\nVerisimilitude also differs from probability. A probability assigned to a hypothesis can represent uncertainty about whether that hypothesis is true. Truthlikeness concerns how accurately it describes reality, whether or not anyone knows that reality. A hypothesis assigned a low probability can happen to be true; a hypothesis known to be false can still be a useful approximation.\n\nThree questions therefore need to be separated:\n\n- Truth: Is the claim correct?\n- Truthlikeness: How closely does it represent the relevant facts?\n- Evidential support: How strongly does the available evidence support accepting it?\n\nThe third question belongs to epistemology. An account of objective truthlikeness does not automatically provide a reliable method for recognizing which theory is closest to the truth."
    },
    {
      "heading": "Popper’s account and scientific progress",
      "text": "Karl Popper made verisimilitude central to his account of scientific progress. His falsificationism rejected the idea that successful tests establish the truth of a universal scientific theory. Theories remain conjectural, and even well-supported theories may eventually prove false. Nevertheless, Popper argued that science can aim at truth and sometimes advance toward it.\n\nTruthlikeness was intended to explain how this is possible. If successive theories are false, their shared falsity need not make them equally unsuccessful descriptions of the world. A later theory might explain more phenomena correctly, correct some errors of its predecessor, and introduce fewer or less serious errors of its own.\n\nPopper developed his influential account in Conjectures and Refutations (1963). His approach combined two considerations:\n\n- The amount of truth a theory contains.\n- The amount of falsity it contains.\n\nThis was intended to avoid identifying progress either with mere truth or with sheer informativeness. A trivial truth says little about the world. A bold theory may say much more, but greater content is not beneficial if it mainly introduces errors.\n\nThe proposal is particularly important for scientific realism, which treats scientific theories as attempts to describe a reality that exists independently of their acceptance. Verisimilitude offers a way to articulate the realist claim that a theory may be approximately correct without being literally true in every respect."
    },
    {
      "heading": "True content, false content, and the 1974 criticism",
      "text": "Popper’s qualitative proposal used a theory’s logical consequences. Its truth content consists of its true consequences; its falsity content consists of its false consequences.\n\nIn simplified form, theory \\(B\\) would be more truthlike than theory \\(A\\) if:\n\n- \\(B\\) had at least all of \\(A\\)’s truth content;\n- \\(B\\) had no more falsity content than \\(A\\);\n- at least one comparison was a strict improvement.\n\nThe attraction is clear: a better theory should gain truth without gaining error, or lose error without losing truth.\n\nIn 1974, Pavel Tichý and David Miller independently demonstrated a fundamental defect in this consequence-based proposal. Under its intended logical framework, the conditions do not deliver the desired comparisons between false theories. The true and false consequences of a theory cannot vary independently in the way the definition requires.\n\nAn important source of difficulty is that false theories have many true consequences. From a false statement \\(p\\), one can derive \\(p \\lor q\\), which is true whenever \\(q\\) is true. Such logical relationships make it difficult to treat true content and false content as independently adjustable collections.\n\nThe criticism did not show that every conception of truthlikeness is incoherent. It showed that Popper’s original formal account failed to capture the comparisons it was designed to explain. Subsequent work therefore explored other ways to represent accuracy, information, and error."
    },
    {
      "heading": "Similarity and distance approaches",
      "text": "An influential alternative treats truthlikeness through similarity to the actual state of the world. Rather than counting a theory’s true and false consequences, it asks how far the possibilities permitted by that theory lie from actuality. Pavel Tichý, Graham Oddie, and Ilkka Niiniluoto are among the philosophers associated with the development of such approaches.\n\nComparing complete descriptions\n\nSuppose a simplified domain contains three features, each represented by either 0 or 1. The actual state is\n\n\\[w^{*}=(1,1,1).\\]\n\nTwo complete descriptions are\n\n\\[A=(1,1,0), \\qquad B=(0,0,1).\\]\n\nIf distance is defined as the number of mismatching features, then\n\n\\[d(A,w^{*})=1,\\qquad d(B,w^{*})=2.\\]\n\nOn this measure, \\(A\\) is more truthlike than \\(B\\). Both descriptions are false, but \\(A\\) gets more of the specified features right.\n\nThis illustration also reveals an important assumption: the three features have been given equal weight. A different weighting could change the comparison.\n\nComparing incomplete theories\n\nScientific theories rarely provide complete descriptions of every relevant fact. They commonly permit many possible worlds or models. A distance-based account must therefore assess a set of possibilities rather than a single description.\n\nDifferent proposals consider, among other things:\n\n- the distance of the nearest permitted possibility from actuality;\n- the distance of the most remote permitted possibility;\n- an average of distances;\n- combinations that balance closeness with informativeness.\n\nThe choice matters. Measuring only the nearest possibility can make every true theory appear maximally successful, since a true theory permits the actual state. Yet a very weak true theory may permit numerous inaccurate alternatives and convey little information.\n\nThis produces a central tension: an account should reward accuracy without making uninformative claims automatically superior to informative approximations."
    },
    {
      "heading": "Measurement, language, and relevance",
      "text": "There is no universally accepted measure of verisimilitude suitable for every scientific domain. Several difficulties recur.\n\nChoice of features. A comparison depends on which properties are represented. Two models may agree closely about observable quantities while differing greatly about underlying mechanisms.\n\nChoice of scale. Numerical error can be measured in absolute, relative, squared, or other terms. These measures need not rank approximations identically.\n\nLanguage dependence. A ranking based on how many elementary claims are correct may change when the same subject matter is expressed through different predicates. Logical equivalence alone does not supply a unique similarity metric.\n\nImportance of errors. Counting mistakes treats them as interchangeable. In practice, a small numerical discrepancy and a mistaken account of the system’s structure can have very different significance.\n\nScope. A model may be highly accurate within a restricted domain and seriously inaccurate outside it. Comparisons must specify the circumstances and questions to which they apply.\n\nConnections with statistics help clarify some of these issues. A loss function specifies how prediction errors are penalized. Such a function can make a comparison precise, but selecting it already involves decisions about which discrepancies matter. Predictive loss is therefore one operational measure of performance, not automatically a complete measure of a theory’s truthlikeness."
    },
    {
      "heading": "Approximation, idealization, and scientific use",
      "text": "Verisimilitude provides a vocabulary for discussing progress among theories that contain false assumptions or idealizations. Classical mechanics, for example, remains highly accurate for many macroscopic applications at speeds much lower than the speed of light, despite not being an unrestricted description of physical reality.\n\nHowever, several achievements should not be equated:\n\n- accurate predictions;\n- an adequate scientific explanation;\n- correct identification of causal mechanisms;\n- usefulness for a particular calculation;\n- overall closeness to the truth.\n\nA model may predict well without correctly representing its subject’s underlying structure. Conversely, a theory may identify an important mechanism while providing only rough quantitative predictions.\n\nIdealization introduces a further complication. Scientific models deliberately omit factors, simplify geometries, or consider limiting cases. Their departures from literal truth can help isolate relevant dependencies. Assessing their truthlikeness requires attention to what they are intended to represent, rather than treating every omission as an equally serious error.\n\nThere is also a distinction between actual and estimated truthlikeness. Actual truthlikeness depends on how the world is. Estimated truthlikeness depends on available evidence, background assumptions, and methods of inference. Since the full truth is generally unknown, scientists cannot ordinarily compare theories directly against a complete description of actuality. Experimental success may justify preferring one theory without guaranteeing that it is closer to the truth in every respect."
    },
    {
      "heading": "Verisimilitude in literature and the arts",
      "text": "In literature, theatre, and film, verisimilitude generally means credible representation rather than measurable proximity to factual truth. Fictional events can be verisimilar even though they never occurred.\n\nA historical novel may achieve this effect through convincing social practices, material details, speech, and motives. A fantasy narrative may achieve it through consistent rules and intelligible consequences, despite containing impossible events. Plausibility is therefore not identical to everyday realism.\n\nA major classical antecedent is Aristotle’s Poetics. Aristotle distinguishes the poet’s concern with what could happen according to probability or necessity from the historian’s concern with what has happened. He also allows that a convincing impossibility may be preferable in poetry to an unconvincing possibility. Dramatic credibility consequently depends on the organization of the work, not simply on whether its events are factually possible.\n\nArtistic verisimilitude can involve:\n\n- Internal consistency: the work respects its established conditions.\n- Psychological intelligibility: actions are credible in relation to characters’ motives.\n- Cultural plausibility: details fit the represented setting.\n- Generic expectations: events make sense within the conventions of a particular kind of narrative.\n\nThese expectations vary across audiences and historical periods. A credible representation may reproduce familiar conventions rather than accurately depict reality. Conversely, a factually accurate event may initially seem implausible because it conflicts with audience expectations. This is the decisive difference between the term’s two principal uses: scientific truthlikeness concerns correspondence with reality, whereas artistic verisimilitude can be achieved through the convincing appearance of such correspondence."
    }
  ],
  "links": [
    "philosophy-of-science",
    "aesthetics",
    "truth",
    "proposition",
    "probability",
    "evidence",
    "epistemology",
    "karl-popper",
    "falsificationism",
    "logical-consequence",
    "possible-worlds",
    "statistics",
    "loss-function",
    "classical-mechanics",
    "speed-of-light",
    "scientific-explanation",
    "literature",
    "theatre",
    "film",
    "aristotle",
    "narrative"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}