In respect of evidence and proof, the mainstream tradition of Anglo-American Evidence Law is the probability theory, which is also known as atomistic, represented by Bayesianism. According to atomism, the fact-finder would firstly evaluate the probative value of each evidence in an atomistic way, and then integrate these values based on an algorithm. Atomism consists of two dimensions: in the first dimension, atomism evaluates a particular evidence in a space relatively empty of matter, considering each evidence as an island and evaluating each evidence in a relatively isolated way. In the second dimension, atomism involves separating evidence into smaller units – evaluating evidences one by one, other than as a whole. It can be said that, within the evidence related Anglo-American discourse traditions, mainstream opinions related to evidence organization and evaluation are still “atomistic” to date, that is to say, restructuring of and criticizing on arguments related to evidence mainly refer to logical analysis on the relationship among individual propositions based on evidence. This atomistic way of analysis in the Anglo-American Evidence Law may trace back to empiricism and rationalism tradition of UK, the former of which include philosophers such as Bacon, Locke, Bentham, and Mill, while the latter include evidence specialists such as Gilbert, Bentham, Steven and Wigmore. Especially the “rationalist tradition” claimed by Twining constitutes the direct intellectual basis of atomistic method: epistemology is knowability, other than skepticism; more of a correspondence theory of truth than coherence; as a kind of social value, exploration on truth has obtained a sublime yet not necessarily overwhelming priority.However, there are some unsurmountable inherent difficulties within this atomistic method and rationalist tradition behind it, therefore, it is confronted with challenge from an alternative method, which is holistic. According to holism, a large amount of evidences and their probative value should be evaluated as a whole, or the meaning or value of any specific evidence atoms would depend on the role it plays when related to other existing evidences of the expounder, which is to say that, the determination of admissibility and probative value of evidence are contextualized. It is not only atomism method represented by probability theory that holistic approach wants to challenge, but even the mainstream tradition of Anglo-American Evidence Law, if the approach may achieve success. In the opinion of holists, apart from problems such as complexity, difficulty to calculate and paradox, the biggest deficiency of probability theory is failure to reflect the actual process of decision-making by the fact-finder. Under this circumstance, some holistic approaches against traditional atomistic method were put forward, represented by story model, theory of anchored narrative, the relative plausibility theory, Mc Cormick’s theory of narrative coherence, Jackson’s semiotic model of fact-construction, Amaya’s theory of optimal coherence.Story model is a theory put forward by Pennington and Hastie through psychological experiments. According to this theory, a core process of decision-making by a juror is building “narrative structure” to organize and explain evidences. Pursuant to story model theory, the juror would accept the story that explain the evidence best out of alternatives, which generally needs to meet certainty principles such as coverage, coherence and uniqueness. The anchored narratives theory was put forward by Wagenaar and his partners, the major doctrine of which is: fact-finder draws conclusion based on two judgments: one is judgment on the plausibility of narrative by the prosecution, while the other is on whether the narrative by the prosecution is correctly anchored on some common-sense beliefs widely recognized. Combing these two judgments, the fact-finder will form an evaluation on the credibility of the narrative by the prosecution, which will be the basis of the final decision. Both story model theory and anchored narrative theory are psychological models in respect of legal decision-making, reflecting the actual process of decision-making by the fact-finder. In these two models, narratives or stories play an important role. Although stories are necessary for organizing and evaluating evidence, they are also dangerous due to narrative fallacy, prejudice to the defendant and undefined standards of story selection.Out of dissatisfaction with traditional Bayesian model, Allen has put forward “the relative plausibility theory”. The main argument of the relative plausibility theory is that, fact-finding involves the determination of the comparative plausibility of stories set forth in the trial by both parties. While in the determination, the plausibility of stories is defined by coherence, integrity, uniqueness, economy and probability of stories. After the relative plausibility theory, Allen has cooperated with Pardo to further develop and deepen his holism theory with inference to the best explanation from scientific philosophy. Therefore, the holism theory of Allen is to connect story model theory with abduction. According to Allen, inference to the best explanation can explain both the macroscopic and microscopic structures of the trial. Although Allen claimed his researches to be empirical, he hadn’t fully carried through the stance with obvious traces of normative in his theories. One of the major problem of relative plausibility is its genetic relationship with free proof, while free proof regarding competency of evidence is inconsistent with some concepts of modern Evidence Law. In respect of actual problems, Allen supported a kind of veritisim in the perspective of naturalized epistemology, and was against explaining the truth in the perspective of probability.Mc Cormick has developed an influential theory of legal reasoning, in which the narrative coherence is a key concept. In Mc Cormick’s opinion, the nature of narrative coherence is a property of set of fact propositions, while the set is understood as a whole. What makes a story coherent is its explicability based on a singular set of explanatory principles, which are about a causality relationship and motivation type. According to Mc Cormick, narrative coherence plays an important role in the justification of factual statements. Specifically, in respect of facts and evidences that cannot be proved through direct observation, narrative coherence is a test for the truth or probability of them. Meanwhile,Mc Cormick suggested that, when an hypothesis of dispute fact forms the best coherence with evidence and a set of descriptive principles, the fact-finder will be justified and accept the hypothesis. Thus, on the one hand, Mc Cormick deems correspondence theory as the definition of “truth” and coherence theory as the criterion of “truth”; on the other hand, he deems coherence as a test of justification.But in fact, narrative coherence can only as a test of truth in the process of discovery,playing a preliminary test role, rather than a test of justification.Compared to Mc Cormick, Jackson has put forward a semiotic model of fact-construction. In his opinion, the major premise of syllogism implies a narrative form and the minor premise should be constructed as a narrative form, while judicial judgment involves the comparison of these two narratives, to see whether they are coherent or fit. Another distinctive feature of Jackson’s narrative coherence lies in that he has distinguished the semantics and pragmatics of trial. He thought that, two narrativizations are needed for fact-construction: narrativization of the semantics of the story being told and the narrativization of the pragmatics about those stories. In other words, the judgment relates to two stories: the story in the trial and the story of the trial. In respect of the truth, Jackson’s opinion is most vibrant and special, supporting a pragmatics notion of truth, instead of coherence theory of truth. When it comes to differentiating non-epistemic “truth” from epistemic “truth”, correspondence theory of truth and meaning constructivism can be combined together.Coherence-based reasoning is another holism method, the typical theory of which is optimal coherence theory put forward by Amaya. In her model, she thinks that, a conviction regarding disputed legal and factual problem can only be justified when it is in optimal coherence, that is to say, an epistemic responsible legal decision-maker may accept it being justified based on its coherence under similar conditions. Whether the legal decision-maker is epistemic responsible, depends on whether they have performed some epistemic duties, and whether they have carried out inquiry and deliberation in a way with epistemic virtue. Amaya thinks that, compared to other theories, optimal coherence puts forward a distinctive concept of coherence and specific inferential model, making optimal coherence practical to a certain extent. Moreover, in the perspective of naturalized epistemology, the theory has not only the plausibility of psychology, but also the attraction of normativity. In respect of truth, Amaya insists on correspondence theory of truth, while coherence theory is more of justification, both of which can be combined together despite some difficulties.However, because Amaya confused intention an extention of coherence, she did not solve the vagueness of coherence.Meanwhile, optimal coherence theory did not overcome the riddle of induction, it is insufficient to be a theory of justification.It is indicated at the end of this paper that, bayesism and probability theory only play a limited role in judicial practice, but it does not mean that the atomism approach in general useless. Atomism and holism are much an artificial division, not a relationship of being diametrically opposed and mutually substituted,but a relationship of being complemented. In this relationship, atomism emphasizes on argumentation, while holism pays more attention to the holistic overview, therefore, a hybrid model combining these two has been developed. In addition, there are researches from psychology and cognitive science showing that, the epistemic process including fact-finding is a dual-process, as a result, Bayesian model and narrative model are both necessary for the fact-finder. In respect of the judicial concept of truth, we should first distinguish the epistemic “truth” from non-epistemic “truth”, and then understand it from the perspective of systemic, namely these two “truths” are inseparable for understanding the judicial concept of truth, and can be combined together. Hereinto, non-epistemic “truth”, namely the correspondence theory of truth, provides the definition and meaning of “truth”; while epistemic “truth”, including coherence theory, constructivism, etc., provides the test criterion, justification project or descriptive method for “truth”. |