Deductive rationality in validating and testing conditional inferences

We asked people to validate conditional inferences (e.g., "A, therefore C" with "if A then C"). People are more likely to look for falsifications ("A and not-C") versus confirmations ("A and C") given a forced choice. Second, falsification rates are lower for logically valid versus invalid inferences. Logically valid inferences are inferences that follow necessarily. Experiment 1 (N = 96) shows that emphasising this logicality constraint increases falsification rates in the validation task and corroborates that validation-by-falsification increases logically correct inference evaluations. Experiment 2 (N = 41) corroborates the other way round that people who are more likely to make logically correct evaluations, show higher falsification performance in the validation task. The results support mental-models theory and suggest alternative theories similarly need to specify how people would go about looking for counterexamples. We proffer such a specification for two alternatives to the model theory. (PsycINFO Database Record (c) 2008 APA, all rights reserved).     

Deductive and inductive conditional inferences: Two modes of reasoning

A number of single- and dual-process theories provide competing explanations as to how reasoners evaluate conditional arguments. Some of these theories are typically linked to different instructions—namely deductive and inductive instructions. To assess whether responses under both instructions can be explained by a single process, or if they reflect two modes of conditional reasoning, we re-analysed four experiments that used both deductive and inductive instructions for conditional inference tasks. Our re-analysis provided evidence consistent with a single process. In two new experiments we established a double dissociation of deductive and inductive instructions when validity and plausibility of conditional problems were pitted against each other. This indicates that at least two processes contribute to conditional reasoning. We conclude that single-process theories of conditional reasoning cannot explain the observed results. Theories that postulate at least two processes are needed to account for our findings.     

Hypothetical and certain inferences from conditional arguments read in texts

In 3 experiments, affirmative and hypothetical probes were presented after narrative texts containing conditional arguments. According to the data, readers represented modus ponens deductions as certain, except when it was only a weakly necessary cause of a given effect. They represented any logically invalid inferences resulting from affirming the conditional consequent as hypothetical, except when it was the effect of strongly sufficient cause. Accordingly, readers must be processing conditional syntax as an asymmetric constraint. However, the underlying causal knowledge can be sufficient either to discredit or warrant the inferences. Thus according to the theory of natural logic, readers can draw formal deductions and be convinced of their necessity. This provides further evidence that readers can represent their inferences as hypothetical (N. Campion, 2004).     

Causal and conditional inferences: A comment on Cummins (1995)

Cummins (1995) offers an analysis of causal and truthfunctional sufficiency and necessity to predict and explain the effects on conditional inferences of two pragmatic factors: alternative causes and disabling conditions. However, the justification of these predictions is inconsistent. This note offers a modified analysis which puts her predictions on a sounder base: it is proposed that alternative causes and disabling conditions affect judgments of argument validity under three different models for the causal conditional.     

Processing of logically valid and logically invalid conditional inferences in discourse comprehension

Two competing theories of processing of conditionals (if-then) were tested. Syntactic theories posit that people only draw inferences conforming to the logically valid modus ponens (MP) schema. Mental models theories predict that people draw MP and invalid affirming-the-consequent (AC) inferences. Three experiments tested these predictions. Participants read short stories that conformed to either the MP or AC form but without conclusions, and they completed either priming or recognition tasks. Results indicate that both MP and AC inferences occur during discourse processing: MP and AC premise forms prime their respective conclusions, participants erroneously judged that they had read the conclusions to MP and AC arguments, and AC inferences did not stem from a biconditional interpretation of conditionals. Findings support mental models theories.     

Priming effect in affirmative complex conditional connectives

We report the results of an experiment that examined the mental representations underlying the comprehension of complex conditional connectives (such as “A, on condition that B”) and the conditional if (such as “A, if B”). The mental representations during the comprehension stage were analysed using a “priming methodology”. The experiment showed that participants read the possibility “A and not-B” faster when it was primed by “A, if B” than when it was primed by “A, on condition that B”. The finding suggests that people understand the sentence “A, on condition that B” as biconditional; other possibilities (“A and B”, “not-A and B”, “not-A and not-B”) were primed equally by both connectives. We discuss the implications of this for current theories of reasoning.     

Stimulus control during conditional discrimination

Pigeons were used to assess stimulus control during the development of a conditional discrimination. The training consisted of three stages. In Stage 1, key pecks were reinforced in the presence of a white line tilted 40 degrees to the right of vertical on a green background and non-reinforced when the same line appeared on a red background. In Stage 2, key pecks were reinforced when a white vertical line appeared on a red background and were non-reinforced in the presence of a 40 degrees slanted line on a red background. In Stage 3, key pecks were reinforced in the presence of the green background regardless of the line tilt, but were differentially reinforced in the presence of the red background (as in Stage 2). Generalization tests were conducted after each stage of training and consisted of five white lines on backgrounds that were green, red, or dark. The effects of the differential reinforcement contingencies on control by line orientation were restricted to the condition in which the red light appeared and resulted in behavioral control that could be characterized as: if red, pay closer attention to line tilt than if not red.     

Attention and generalization during a conditional discrimination

A conditional discrimination was established and analyzed, using four pigeons. The discrimination was among four compound stimuli projected on the response key—a white circle or triangle on a red or green background—during two conditions of illumination in the chamber, no illumination or flashing illumination. The two lighting conditions indicated whether the stimuli on the key containing triangles or those containing red would be the occasion for reinforcement. After the discrimination formed, generalization to intermediate and extreme values of the conditional stimulus and the attention of the birds to separate aspects of the stimulus on the key under each of the conditional stimuli were studied. All subjects generalized across values of the conditional stimulus, the lighting of the chamber. But subjects differed in the manner in which they treated the compound stimuli: two tended to attend to one or the other aspect of the stimulus on the key depending on the conditional stimulus, and two offered no evidence of such selective attention. Thus, the differential control of responding by the conditional stimuli cannot be attributed to a shift in attention between the figure and ground aspects of the compound stimuli.     

Behavioral interactions and stimulus control during conditional discriminations

Two pigeons were exposed to factorial combinations of two values of line tilt and two frequencies of houselight flashes. During each of four baseline stages, key pecking in the presence of all four combinations was reinforced according to a variable-interval schedule. The baseline phases were followed by four different conditional discrimination training procedures in which reinforcement availability for pecking in the presence of the line tilts depended upon the houselight frequency. The subjects acquired each conditional discrimination. Behavioral contrast occurred during the acquisition and abolition of the discriminations. Generalization tests, given after each conditional discrimination, revealed that both the line tilt and houselight frequency dimensions controlled pecking only after conditional discriminations in which reinforcement availability depended upon the value of both dimensions.     

Working memory and inferences: evidence from eye fixations during reading

Eye fixations during reading were monitored to examine the relationship between individual differences in working memory capacity-as assessed by the reading span task-and inferences about predictable events. Context sentences predicting likely events, or non-predicting control sentences, were presented. They were followed by continuation sentences in which a target word represented an event to be inferred (inferential word) or an unlikely event (non-predictable word). A main effect of reading span showed that high working memory capacity was related to shorter gaze durations across sentence regions. More specific findings involved an interaction between context, target, and reading span on late processing measures and regions. Thus, for high- but not for low-span readers, the predicting condition, relative to the control condition, facilitated reanalysis of the continuation sentence that represented the inference concept. This effect was revealed by a reduction in regression-path reading time in the last region of the sentence, involving less time reading that region and fewer regressions from it. These results indicate that working memory facilitates elaborative inferences during reading, but that this occurs at late text-integration processes, rather than at early lexical-access processes.     

Automatic goal inferences

The social psychological literature on automatic social inferences has focused on one construct that helps explaining human behavior—traits (e.g., Gilbert, Pelham, & Krull, 1988; Trope, 1986; Winter & Uleman, 1984). The dispositional roots of behavior, however, go beyond relatively stable constructs such as traits to include more transient causes such as one’s intentions and goals. Evidence from young infants and adult chimpanzees, knowledge acquired in the text-comprehension literature and hypotheses derived from the Automatic Causal Inferences framework (Hassin, Bargh, & Uleman, 2002), seems to converge: they all suggest that perceivers may automatically infer goals from behaviors. This paper reports four studies that examine this hypothesis. The first two use surprise cued-recall, and look at goal inferences when the road to goal achievement seems straightforward and when it seems blocked. Studies 3 and 4 use on-line methodologies—probe recognition task and lexical decision—to examine whether these inferences are made at encoding.     

Intuitions for inferences

In this paper, I explore a question about deductive reasoning: why am I in a position to immediately infer some deductive consequences of what I know, but not others? I show why the question cannot be answered in the most natural ways of answering it, in particular in Descartes’s way of answering it. I then go on to introduce a new approach to answering the question, an approach inspired by Hume’s view of inductive reasoning.     

Spontaneous causal inferences

Three studies examine the hypothesis that people spontaneously (i.e., unintentionally and without awareness of doing so) infer causes (the Spontaneous Causal Inference, or SCI, hypothesis). Using a cued-recall paradigm, Study 1 examines whether SCIs occur and Study 2 allows for a comparison between implicitly inferred and explicitly mentioned causes. Study 3 examines whether SCIs can be fully explained in terms of spreading activation to general, abstract schemes. It is suggested that STIs (e.g., Winter & Uleman, 1984), and spontaneous predicting inferences (e.g., [McKoon and Ratcliff, 1986a] and [McKoon and Ratcliff, 1986b]), may be better understood in their relation to SCIs.     

Associative change in the representations acquired during conditional discriminations: further analysis of the nature of conditional learning

Three experiments with rats investigated how the associative strengths of the representations that underlie conditional learning change when they are conditioned in compound. The results of each experiment suggest that the representation whose associative strength is most discrepant from the asymptote supported by the outcome of the trial undergoes the greatest change in associative strength. These results parallel those from simple Pavlovian conditioning (e.g., R. A. Rescorla, 2000). are inconsistent with unique-cue and configural accounts of conditional learning, and support a connectionist analysis of learning in which a "winner-takes-all" rule applies to the hidden units that can be activated and acquire associative strength at a given point in time.     

Spontaneous goal inferences are often inferred faster than spontaneous trait inferences

We present four experiments in which participants were exposed to texts depicting behaviors that afforded inferences about actors' traits and goals. Results from a false recognition task with varying response deadlines revealed heightened activation of goal inferences already within a 350 ms response deadline. In contrast, trait inferences were made only when there was no response deadline, and when the behavior also implied a goal. These results indicate that spontaneous inferences on goals are often encoded more strongly in memory and are reactivated much more quickly in comparison with spontaneous trait inferences. Moreover, spontaneous trait inferences are often facilitated when an inference is first made on the goal of the behavior. These findings are discussed in light of recent developmental and neuroscientific evidence on social inferences, and current theories on impression formation.     

词语型类别属性归纳中分类与属性推理过程的时间特征

先前的研究中没有清晰分离类别属性归纳任务中分类过程与属性推理过程的ERP特征。本研究采用了事件相关电位技术(ERPs), 通过在词语型类别属性归纳任务的结论呈现阶段分开呈现结论类别和结论属性, 探讨了类别属性归纳中分类过程和属性推理过程的ERP特征。实验结果显示：在结论类别呈现阶段, 结论无关类别比结论关联类别诱发了更大的N400, 表明N400与类别属性归纳中的语义分类过程密切相关。在结论属性呈现阶段, 无论是在类别关联还是在类别无关条件下, 不一致属性比一致属性均诱发了更大的前部N2成分, 反映了属性知觉上的失匹配或规则的违背。随后, 在300~400 ms的时间窗口发现, 一致属性比不一致属性在前部脑区诱发了更大的类似于P3a的正偏转, 反映了对一致属性的更多关注。同时, 实验还发现, 当结论类别与前提类别关联时, 在300~400 ms的时间窗口, 一致属性比不一致属性在后部脑区也诱发了更大的类似于P3b的正偏转, 反映了推理预期的满意程度; 随后, 在400~600 ms的时间窗口, 不一致属性比一致属性在后部脑区诱发了更大的正偏转, 这与推理规则的违反相关。