Evidence

Evidence for a proposition is what supports the proposition. It is usually understood as an indication that the supported proposition is true. What role evidence plays and how it is conceived varies from field to field.

For other uses, see Evidence (disambiguation).

In epistemology, evidence is what justifies beliefs or what makes it rational to hold a certain doxastic attitude. For example, a perceptual experience of a tree may act as evidence that justifies the belief that there is a tree. In this role, evidence is usually understood as a private mental state. Important topics in this field include the questions of what the nature of these mental states is, for example, whether they have to be propositional, and whether misleading mental states can still qualify as evidence.

In phenomenology, evidence is understood in a similar sense. Here, however, it is limited to intuitive knowledge that provides immediate access to truth and is therefore indubitable. In this role, it is supposed to provide ultimate justifications for basic philosophical principles and thus turn philosophy into a rigorous science. However, it is highly controversial whether evidence can meet these requirements.

In philosophy of science, evidence is understood as that which confirms or disconfirms scientific hypotheses. Measurements of Mercury's "anomalous" orbit, for example, are seen as evidence that confirms Einstein's theory of general relativity. In order to play the role of neutral arbiter between competing theories, it is important that scientific evidence is public and uncontroversial, like observable physical objects or events, so that the proponents of the different theories can agree on what the evidence is. This is ensured by following the scientific method and tends to lead to an emerging scientific consensus through the gradual accumulation of evidence. Two issues for the scientific conception of evidence are the problem of underdetermination, i.e. that the available evidence may support competing theories equally well, and theory-ladenness, i.e. that what some scientists consider the evidence to be may already involve various theoretical assumptions not shared by other scientists. It is often held that there are two kinds of evidence: intellectual evidence or what is self-evident and empirical evidence or evidence accessible through the senses.

Other fields, including the sciences and the law, tend to emphasize more the public nature of evidence (for example, scientists tend to focus on how the data used during statistical inference are generated).^[1]

In order for something to act as evidence for a hypothesis, it has to stand in the right relation to it. In philosophy, this is referred to as the "evidential relation" and there are competing theories about what this relation has to be like. Probabilistic approaches hold that something counts as evidence if it increases the probability of the supported hypothesis. According to hypothetico-deductivism, evidence consists in observational consequences of the hypothesis. The positive-instance approach states that an observation sentence is evidence for a universal hypothesis if the sentence describes a positive instance of this hypothesis. The evidential relation can occur in various degrees of strength. These degrees range from direct proof of the truth of a hypothesis to weak evidence that is merely consistent with the hypothesis but does not rule out other, competing hypotheses, as in circumstantial evidence. In law, rules of evidence govern the types of evidence that are admissible in a legal proceeding. Types of legal evidence include testimony, documentary evidence, and physical evidence.^[2] The parts of a legal case that are not in controversy are known, in general, as the "facts of the case." Beyond any facts that are undisputed, a judge or jury is usually tasked with being a trier of fact for the other issues of a case. Evidence and rules are used to decide questions of fact that are disputed, some of which may be determined by the legal burden of proof relevant to the case. Evidence in certain cases (e.g. capital crimes) must be more compelling than in other situations (e.g. minor civil disputes), which drastically affects the quality and quantity of evidence necessary to decide a case.

Nature of evidence[edit]

Notion[edit]

Understood in its broadest sense, evidence for a proposition is what supports this proposition. Traditionally, the term is sometimes understood in a narrower sense: as the intuitive knowledge of facts that are considered indubitable.^[3]^[4]^[5] In this sense, only the singular form is used. This meaning is found especially in phenomenology, in which evidence is elevated to one of the basic principles of philosophy, giving philosophy the ultimate justifications that are supposed to turn it into a rigorous science.^[6]^[4]^[7] In a more modern usage, the plural form is also used. In academic discourse, evidence plays a central role in epistemology and in the philosophy of science. Reference to evidence is made in many different fields, like in science, in the legal system, in history, in journalism and in everyday discourse.^[8]^[9]^[10] A variety of different attempts have been made to conceptualize the nature of evidence. These attempts often proceed by starting with intuitions from one field or in relation to one theoretical role played by evidence and go on to generalize these intuitions, leading to a universal definition of evidence.^[8]^[9]^[11]

One important intuition is that evidence is what justifies beliefs. This line of thought is usually followed in epistemology and tends to explain evidence in terms of private mental states, for example, as experiences, other beliefs or knowledge. This is closely related to the idea that how rational someone is, is determined by how they respond to evidence.^[8]^[9]^[12]^[13]^[14] Another intuition, which is more dominant in the philosophy of science, focuses on evidence as that which confirms scientific hypotheses and arbitrates between competing theories.^[15] On this view, it is essential that evidence is public so that different scientists can share the same evidence. This leaves publicly observable phenomena like physical objects and events as the best candidates for evidence, unlike private mental states.^[8]^[9]^[14] One problem with these approaches is that the resulting definitions of evidence, both within a field and between fields, vary a lot and are incompatible with each other. For example, it is not clear what a bloody knife and a perceptual experience have in common when both are treated as evidence in different disciplines. This suggests that there is no unitary concept corresponding to the different theoretical roles ascribed to evidence, i.e. that we do not always mean the same thing when we talk of evidence.^[8]^[9]^[11]

Characteristics[edit]

On the other hand, Aristotle, phenomenologists, and numerous scholars accept that there could be several degrees of evidence.^[16] For instance, while the outcome of a complex equation may become more or less evident to a mathematician after hours of deduction, yet with little doubts about it, a simpler formula would appear more evident to them.

Riofrio has detected some characteristics that are present in evident arguments and proofs. The more they are evident, the more these characteristics will be present. There are six intrinsic characteristics of evidence:^[17]

In addition, four subjective or external characteristics can be detected over those things that are more or less evident:

These ten characteristics of what is evident allowed Riofrio to formulate a test of evidence to detect the level of certainty or evidence that one argument or proof could have.^[17]

Nature of the evidential relation[edit]

Philosophers in the 20th century started to investigate the "evidential relation", the relation between evidence and the proposition supported by it.^[1] The issue of the nature of the evidential relation concerns the question of what this relation has to be like in order for one thing to justify a belief or to confirm a hypothesis.^[15] Important theories in this field include the probabilistic approach, hypothetico-deductivism and the positive-instance approach.^[8]^[41]

Probabilistic approaches, also referred to as Bayesian confirmation theory, explain the evidential relation in terms of probabilities. They hold that all that is necessary is that the existence of the evidence increases the likelihood that the hypothesis is true. This can be expressed mathematically as $P(H\mid E)>P(H)$ .^[42]^[43] In words: a piece of evidence (E) confirms a hypothesis (H) if the conditional probability of this hypothesis relative to the evidence is higher than the unconditional probability of the hypothesis by itself.^[44] Smoke (E), for example, is evidence that there is a fire (H), because the two usually occur together, which is why the likelihood of fire given that there is smoke is higher than the likelihood of fire by itself. On this view, evidence is akin to an indicator or a symptom of the truth of the hypothesis.^[11] Against this approach, it has been argued that it is too liberal because it allows accidental generalizations as evidence. Finding a nickel in one's pocket, for example, raises the probability of the hypothesis that "All the coins in my pockets are nickels". But, according to Alvin Goldman, it should not be considered evidence for this hypothesis since there is no lawful connection between this one nickel and the other coins in the pocket.^[9]

Hypothetico-deductivism is a non-probabilistic approach that characterizes the evidential relations in terms of deductive consequences of the hypothesis. According to this view, "evidence for a hypothesis is a true observational consequence of that hypothesis".^[8]^[15]^[45]^[46] One problem with the characterization so far is that hypotheses usually contain relatively little information and therefore have few if any deductive observational consequences. So the hypothesis by itself that there is a fire does not entail that smoke is observed. Instead, various auxiliary assumptions have to be included about the location of the smoke, the fire, the observer, the lighting conditions, the laws of chemistry, etc. In this way, the evidential relation becomes a three-place relation between evidence, hypothesis and auxiliary assumptions.^[15]^[47] This means that whether a thing is evidence for a hypothesis depends on the auxiliary assumptions one holds. This approach fits well with various scientific practices. For example, it is often the case that experimental scientists try to find evidence that would confirm or disconfirm a proposed theory. The hypothetico-deductive approach can be used to predict what should be observed in an experiment if the theory was true.^[47] It thereby explains the evidential relation between the experiment and the theory.^[15] One problem with this approach is that it cannot distinguish between relevant and certain irrelevant cases. So if smoke is evidence for the hypothesis "there is fire", then it is also evidence for conjunctions including this hypothesis, for example, "there is fire and Socrates was wise", despite the fact that Socrates's wisdom is irrelevant here.^[8]

According to the positive-instance approach, an observation sentence is evidence for a universal hypothesis if the sentence describes a positive instance of this hypothesis.^[41]^[48]^[49] For example, the observation that "this swan is white" is an instance of the universal hypothesis that "all swans are white". This approach can be given a precise formulation in first-order logic: a proposition is evidence for a hypothesis if it entails the "development of the hypothesis".^[8]^[15] Intuitively, the development of the hypothesis is what the hypothesis states if it was restricted to only the individuals mentioned in the evidence. In the case above, we have the hypothesis " $\forall x(swan(x)\rightarrow white(x))$ " (all swans are white) which, when restricted to the domain "{a}", containing only the one individual mentioned in the evidence, entails the evidence, i.e. " $swan(a)\land white(a)$ " (this swan is white).^[8]^[15] One important shortcoming of this approach is that it requires that the hypothesis and the evidence are formulated in the same vocabulary, i.e. use the same predicates, like " $swan$ " or " $white$ " above. But many scientific theories posit theoretical objects, like electrons or strings in physics, that are not directly observable and therefore cannot show up in the evidence as conceived here.^[8]^[15]