Non-Bayesian Accounts of Evidence: Howson’s Counterexample Countered

There is a debate in Bayesian confirmation theory between subjective and non-subjective accounts of evidence. Colin Howson has provided a counterexample to our non-subjective account of evidence: the counterexample refers to a case in which there is strong evidence for a hypothesis, but the hypothesis is highly implausible. In this article, we contend that, by supposing that strong evidence for a hypothesis makes the hypothesis more believable, Howson conflates the distinction between confirmation and evidence. We demonstrate that Howson’s counterexample fails for a different pair of hypotheses.     

旧证据问题：一种动态的消解方案

旧证据问题是指贝叶斯确证理论在讨论旧证据确证新理论（假说）的问题上将出现不合理的结果,它的消解是贝叶斯确证理论辩护的一个非常重要的环节。本文倡导一种基于Levi理论的动态方案来消解这一问题。在这种方案下,知识汇集的动态性使得旧证据在新理论引入前后的置信概率可能是不相同的,在此基础上我们证明,经新理论扩充后而得到的知识汇集无论是一致扩充还是不一致扩充,均存在一个旧证据E,它对该新理论具有确证作用。     

“The Ravens Paradox” is a misnomer

I argue that the standard Bayesian solution to the ravens paradox— generally accepted as the most successful solution to the paradox—is insufficiently general. I give an instance of the paradox which is not solved by the standard Bayesian solution. I defend a new, more general solution, which is compatible with the Bayesian account of confirmation. As a solution to the paradox, I argue that the ravens hypothesis ought not to be held equivalent to its contrapositive; more interestingly, I argue that how we formally represent hypotheses ought to vary with the context of inquiry. This explains why the paradox is compelling, while dealing with standard objections to holding hypotheses inequivalent to their contrapositives.     

On Ratio Measures of Confirmation

There are different Bayesian measures to calculate the degree of confirmation of a hypothesis H in respect of a particular piece of evidence E. Zalabardo (Analysis 69:630–635, 2009) is a recent attempt to defend the likelihood-ratio measure (LR) against the probability-ratio measure (PR). The main disagreement between LR and PR concerns their sensitivity to prior probabilities. Zalabardo invokes intuitive plausibility as the appropriate criterion for choosing between them. Furthermore, he claims that it favours the ordering of pairs evidence/hypothesis generated by LR. We will argue, however, that the intuitive non-numerical example provided by Zalabardo does not show that prior probabilities do not affect the degree of confirmation. On account of this, we conclude that there is no compelling reason to endorse LR qua measure of degree of confirmation. On the other side, we should not forget some technicalities which still benefit PR.     

The Role of Coherence of Evidence in the Non-Dynamic Model of Confirmation

This paper examines the role of coherence of evidence in what I call the non-dynamic model of confirmation. It appears that other things being equal, a higher degree of coherence among pieces of evidence raises to a higher degree the probability of the proposition they support. I argue against this view on the basis of three related observations. First, we should be able to assess the impact of coherence on any hypothesis of interest the evidence supports. Second, the impact of coherence among the pieces of evidence can be different on different hypotheses of interest they support. Third, when we assess the impact of coherence on a hypothesis of interest, other conditions that should be held equal for a fair assessment include the degrees of individual support which the propositions directly supported by the respective pieces of evidence provide for the hypothesis. Once we take these points into consideration, the impression that coherence of evidence plays a positive role in confirmation dissipates. In some cases it can be shown that other things being equal, a higher degree of coherence among the pieces of evidence reduces the degree of confirmation for the hypothesis they support.     

A prognostic utility bias in judgments of similarity between past and present instances: How available information is deemed useful for prediction

A variety of theorists have proposed that individuals desire to be confident in their inferences and predictions. One way this can be accomplished is by overestimating the extent to which available information provides a logic basis for such conclusions. Thus, when one's goal is to make an accurate prediction, if information about a prior known instance has potential implications for the prediction, one may overestimate the similarity of the prior known instance to the current instance. In this way, the perceived utility of the information as a basis for prediction can be inflated. To test this idea, three studies were conducted. In each study, subjects were asked to judge the similarity of past instances to an upcoming instance for which an outcome prediction was required. Judging a past instance as similar to the upcoming instance should only provide a firm basis for outcome prediction if outcome information regarding the past instance is known. Therefore, it was expected that when subjects are asked to make an outcome prediction, they would judge a past instance more similar to the upcoming instance if outcome information about the past instance is known than if it is not. Support for this hypothesis was obtained along with evidence concerning the conditions under which the eflect does and does not occur.     

The rank hypothesis and lexical decision: a reply to Adelman and Brown (2008)

J. S. Adelman and G. D. A. Brown (2008) provided an extensive analysis of the form of word frequency and contextual diversity effects on lexical decision time. In this reply, the current authors suggest that their analysis provides a valuable tool for the evaluation of models of lexical access and that the results they report are broadly supportive of the rank hypothesis suggested by W. S. Murray and K. I. Forster (2004)--more supportive, in fact, than the originally reported data. However, Adelman and Brown's conclusion that the results of these analyses can be taken as evidence against rank (and thereby serial models of lexical access) and for instance models is rejected. It is shown that at least one instance model makes the wrong predictions and that Adelman and Brown's conclusions rest on the assumption that lexical decision time presents a pure measure of the time involved in lexical access. Results from eye tracking are reported, which also support a rank account, as do results from analyses that show that a log frequency account is clearly inadequate. Finally, it is demonstrated that, unlike other models, the rank account continues to make accurate predictions regarding the form of both reaction time and error rate effects.     

The Semantic Processing of Motion Verbs: Coercion or Underspecification?

Underspecification and coercion are two prominent interpretive mechanisms to account for meaning variability beyond compositionality. While there is plentiful evidence that natural language meaning constitution exploits both mechanisms, it is an open issue whether a concrete phenomenon of meaning variability is an instance of underspecification or coercion. This paper argues that this theoretical dispute can be settled experimentally. The test case are standard motion verbs (e.g. walk, ride) in combination with ±telic directional phrases, for which both underspecifaction and coercion analyses have been proposed in the literature. A self-paced reading study which incorporates motion verbs, directional phrases and durative/completive temporal adverbials (1) aims at determining the aspectual value of such verbs, and (2) compares the hypotheses of the Underspecification and Coercion Accounts. The results of the reading time experiment (flanked by a corpus study and a completion study) indicate that motion verbs are aspectually underspecified. They combine with ±telic directional phrases with equal ease. The combination with a mismatching temporal adverbial is an instance of coercion, causing additional processing costs.     

Showing that the race model inequality is not violated

When participants are asked to respond in the same way to stimuli from different sources (e.g., auditory and visual), responses are often observed to be substantially faster when both stimuli are presented simultaneously (redundancy gain). Different models account for this effect, the two most important being race models and coactivation models. Redundancy gains consistent with the race model have an upper limit, however, which is given by the well-known race model inequality (Miller, 1982). A number of statistical tests have been proposed for testing the race model inequality in single participants and groups of participants. All of these tests use the race model as the null hypothesis, and rejection of the null hypothesis is considered evidence in favor of coactivation. We introduce a statistical test in which the race model prediction is the alternative hypothesis. This test controls the Type I error if a theory predicts that the race model prediction holds in a given experimental condition.     

Colors without circles?

Realists about color, be they dispositionalists or physicalists, agree on the truth of the following claim: (R) x is red iff x is disposed to look red under standard conditions. The disagreement is only about whether to identify the colors with the relevant dispositions, or with their categorical bases. This is a question about the representational content of color experience: What kind of properties do color experiences ascribe to objects? It has been argued (for instance by Boghossian and Velleman, 1991) that truths like (R) cannot be used in an account of the colors as they would result in ‚circular’, and therefore empty, contents. It has also been argued (for instance by Harman, 1996) that switching to an account of color in terms of a functional account of color sensations would result in a circular, and therefore empty, account. In this paper, I defend a realist account of color in terms of a (non-reductive) functional account of color sensations. Such an account of sensations has been suggested by Pagin (2000), and it can be applied to color sensations without the resulting account of the colors themselves being circular or empty. I argue that the so-called transparency of experience does not provide any argument against such an account. I also argue that on such an account, the issue of physicalism vs. dispositionalism boils down to the question of the modal profile of the color concepts.     

The phonological loop and the irrelevant speech effect: Some comments on Neath (2000)

Neath (2000) presents a useful overview of the evidence to be explained by any model of the effects of irrelevant speech on immediate serial memory and proposes a model accompanied by computational simulation. While his review is in general accurate, it is limited in its explanation of the crucial characteristics of the disrupting sounds. It also neglects strategic issues, particularly the tendency for subjects to switch strategy as list length increases. As a result, his model fails to account for the absence of an interaction between irrelevant speech and acoustic similarity for lists of span length. Points of issue between Neath’s feature hypothesis and the phonological loop interpretation are outlined, and the contribution of his computational simulation is discussed.     

Second-order probability affects hypothesis confirmation

Bayesian confirmation measures give numerical expression to the impact of evidence E on a hypothesis H. All measures proposed to date are formal—that is, functions of the probabilities Pr(E∧H), Pr(E∧¬H), Pr(¬E∧H), Pr(¬E∧¬H), and nothing more. Experiments reported in Tentori, Crupi, and Osherson (2007) suggest that human confirmation judgment is not formal, but this earlier work leaves open the possibility that formality holds relative to a given semantic domain. The present study discredits even this weaker version of formality by demonstrating the role in confirmation judgments of a probability distribution defined over the possible values of Pr(E∧H), Pr(E∧¬H), Pr(¬E∧H), and Pr(¬E∧¬H)—that is, a second-order probability. Specifically, when for each of the latter quantities a pointwise value is fixed with a maximal second-order probability, evidence impact is rated in accordance with formal and normatively credible confirmation measures; otherwise evidence impact is systematically judged as more moderate.     

Confirmation and justification. A commentary on Shogenji’s measure

So far no known measure of confirmation of a hypothesis by evidence has satisfied a minimal requirement concerning thresholds of acceptance. In contrast, Shogenji’s new measure of justification (Shogenji, Synthese, this number 2009) does the trick. As we show, it is ordinally equivalent to the most general measure which satisfies this requirement. We further demonstrate that this general measure resolves the problem of the irrelevant conjunction. Finally, we spell out some implications of the general measure for the Conjunction Effect; in particular we give an example in which the effect occurs in a larger domain, according to Shogenji justification, than Carnap’s measure of confirmation would have led one to expect.     

Determinants of social comparison choices

Two experiments were conducted to clarify previous, inconsistent findings regarding various hypothesized determinants of social comparison choices. A bogus personality trait scale was administered to subjects, and they were either informed of the range of scores in their group (R conditions) or not informed (NR conditions). Experiment 1 closely replicated support for a range-seeking hypothesis and for Festinger's theory of social comparison processes (1954), provided only minimal support for a positive instance hypothesis, and failed to obtain support for a desirability hypothesis. Experiment 2 was run to complement Experiment 1 in providing an unequivocal test of the positive instance and desirability hypotheses. The results suggested a multifactor model to account for comparison choices, the factors being whether or not a score was (a) an uncertain element (i.e., the highest score), (b) a positive instance, and/or (c) desirable.     

Hempel’s logic of confirmation

This paper presents a new analysis of C.G. Hempel’s conditions of adequacy for any relation of confirmation [Hempel C. G. (1945). Aspects of scientific explanation and other essays in the philosophy of science. New York: The Free Press, pp. 3–51.], differing from the one Carnap gave in §87 of his [1962. Logical foundations of probability (2nd ed.). Chicago: University of Chicago Press.]. Hempel, it is argued, felt the need for two concepts of confirmation: one aiming at true hypotheses and another aiming at informative hypotheses. However, he also realized that these two concepts are conflicting, and he gave up the concept of confirmation aiming at informative hypotheses. I then show that one can have Hempel’s cake and eat it too. There is a logic that takes into account both of these two conflicting aspects. According to this logic, a sentence H is an acceptable hypothesis for evidence E if and only if H is both sufficiently plausible given E and sufficiently informative about E. Finally, the logic sheds new light on Carnap’s analysis.     

Five reasons to doubt the existence of a geometric module

It is frequently claimed that the human mind is organized in a modular fashion, a hypothesis linked historically, though not inevitably, to the claim that many aspects of the human mind are innately specified. A specific instance of this line of thought is the proposal of an innately specified geometric module for human reorientation. From a massive modularity position, the reorientation module would be one of a large number that organized the mind. From the core knowledge position, the reorientation module is one of five innate and encapsulated modules that can later be supplemented by use of human language. In this paper, we marshall five lines of evidence that cast doubt on the geometric module hypothesis, unfolded in a series of reasons: (1) Language does not play a necessary role in the integration of feature and geometric cues, although it can be helpful. (2) A model of reorientation requires flexibility to explain variable phenomena. (3) Experience matters over short and long periods. (4) Features are used for true reorientation. (5) The nature of geometric information is not as yet clearly specified. In the final section, we review recent theoretical approaches to the known reorientation phenomena.     

Identifying Hypothesis Confirmation Behaviors in a Simulated Murder Investigation: Implications for Practice

Investigators have been shown to be prone to accessing information that confirms their preferred hypothesis. This tendency has been termed hypothesis confirmation. Hypothesis confirmation behaviour was explored in two studies using a computer administered simulation of a murder investigation. In Study 1, hypothesis confirmation behaviour did not differentiate successful and unsuccessful participants. However, unsuccessful participants stored more confirmatory information than non‐confirmatory information. Successful participants did not show this pattern. In Study 2, unsuccessful participants tended to settle early on a suspect and then acquired information that supported their decision. Successful participants tended to consider a number of suspects for longer and accessed a greater diversity of information. Taken together, these findings are highly suggestive of the role of information acquisition as an important component of hypothesis confirmatory behaviour, which is associated with unsuccessful information processing in the context of a simulated murder investigation. Implications for practice are presented. Copyright © 2012 John Wiley & Sons, Ltd.     

Modeling Customer Loyalty from an Integrative Perspective of Self-Determination Theory and Expectation–Confirmation Theory

Purpose  This research conceptualizes and tests an integrative model of customer loyalty by linking two important theories: expectation–confirmation theory and self-determination theory. Design/Methodology/Approach  The model is examined using data obtained from 207 part-time students who have encountered the service of a skincare and beauty salon in Taiwan. These students work as full-time professionals in a variety of industries during the daytime and are financially independent for their daily consumption. Findings  The empirical results of this study indicate that loyalty is positively influenced by both intrinsic regulation and identified regulation, while introjected regulation and external regulation are insignificantly related to loyalty. Affected positively by service expectation and service confirmation, satisfaction has positive influences on all four dimensions of self-determined motivation—namely, intrinsic regulation, identified regulation, introjected regulation, and external regulation. Implications  The findings of this study show that the proposed model helps to learn about loyalty formation and its mediating mechanism in service contexts. Intrinsic regulation and identified regulation may be applied as two potential checkpoints for management to learn the actual status of customer loyalty based on a constant service quality offered by the service provider. Originality/Value  This study is one of the earliest to integrate expectation–confirmation theory and self-determination theory to explore loyalty. Besides, this study transplants the traditional application of self-determination theory from educational service to commercial service in general so that efficient strategies can be made for boosting loyalty.     

Does the evidence confirm theism more than naturalism?

Conclusion Schlesinger's argument for the confirmation of Theism over Naturalism fails. Principle E on which it is based is unacceptable. But even if Principle E were acceptable the argument would not succeed, for other evidence, not considered by Schlesinger, pulls in the other direction cancelling out the confirmatory effect of the evidence he does cite. Further, granted Principle E, it is not clear that theism as it is usually understood makes the evidence he cites more probable than naturalism; consequently he is not justified in concluding that the evidence he cites confirms theism more than naturalism. Finally, Schlesinger fails to take into account the possibility that the initial credibility of theism may be less than naturalism since theism may be incoherent and naturalism may not be; he also fails to take into account the total available evidence relevant to the confirmation of theism and naturalism.