Online research in philosophy

This paper examines the nature of model-based reasoning in the interplay between theory and experiment in the context of biomedical engineering research laboratories, where problem solving involves using physical models. These "model systems" are sites of experimentation where in vitro models are used to screen, control, and simulate specific aspects of in vivo phenomena. As with all models, simulation devices are idealized representations, but they are also systems themselves, possessing engineering constraints. Drawing on research in contemporary cognitive science that construes cognition as occurring in a complex distributed system comprising people and artifacts, I argue that reasoning with model systems is a constraint satisfaction process involving co-construction, manipulation, and revision of mental and physical models.

This article reports on a study of children's deductive reasoning in solving novel relational problems. Detailed protocols were obtained from 264 children (aged 9- 12 years) who verbalised their thinking as they solved the problems. The study included the development of a three-phase theory based on Johnson-Laird and Byrne's mental models perspective, but with some distinct modifications. These include a focus on the relational complexity entailed in model construction and in premise integration, and the advancement of four reasoning principles that are applied throughout problem solution (in contrast to Johnson-Laird's falsification processes as the hallmark of deductive reasoning). The reported case studies and the results of statistical analyses supported predictions arising from the proposed theory, including the key role of the reasoning principles. The results also showed that problem difficulty is a function of relational complexity, not of the number of models to be constructed, as argued by Johnson-Laird and Byrne.

We report two experiments on temporal reasoning with problems, such as: John has cleaned the house. John is taking a shower. John is going to read the paper. Mary always does the dishes when John cleans the house. Mary always drinks her coffee when John reads the paper. What for Mary is the relation between doing the dishes and drinking coffee? The experiments showed that problems such as this one, which require one mental model, elicited correct answers more often than did those requiring multiple models (e.g. with the second premise modified to "John has taken a shower", so that the order between the events in the first two premises is not fixed). These multiple-model problems, in turn, elicited more correct answers than did multiple-model problems with no valid answers (e.g. with the second premise modified to "John has taken a shower", and the fifth premise modified to "Mary always drinks her coffee when John takes a shower"). One-model problems were also solved more quickly than multiple-model problems, which were solved more quickly than problems with no valid answers. These results corroborated the predictions of the mental model theory of reasoning.

Current theories of reasoning such as mental models or mental logic assume a universal cognitive mechanism that underlies human reasoning performance. However, there is evidence that this is not the case, for example, the work of Ford (1995), who found that some people adopted predominantly spatial and some verbal strategies in a syllogistic reasoning task. Using written and think-aloud protocols, the present study confirmed the existence of these individual differences. However, in sharp contrast to Ford, the present study found few differences in reasoning performance between the two groups, in terms of accuracy or type of conclusion generated. Hence, reasoners present an outward appearance of ubiquity, despite underlying differences in reasoning processes. These findings have implications for theoretical accounts of reasoning, and for attempts to model reasoning data. Any comprehensive account needs to account for strategic differences and how these may develop in logically untrained individuals.

Practical reasoners are resource-bounded—in particular they require time to derive consequences of their knowledge. Building on the Timed Reasoning Logics (TRL) framework introduced in [1], we show how to represent the time required by an agent to reach a given conclusion. TRL allows us to model the kinds of rule application and conflict resolution strategies commonly found in rule-based agents, and we show how the choice of strategy can influence the information an agent can take into account when making decisions at a particular point in time. We prove general completeness and decidability results for TRL, and analyse the impact of communication in an example system consisting of two agents which use different conflict resolution strategies.

We report the results of three experiments designed to assess the role of suppositions in human reasoning. Theories of reasoning based on formal rules propose that the ability to make suppositions is central to deductive reasoning. Our first experiment compared two types of problem that could be solved by a suppositional strategy. Our results showed no difference in difficulty between problems requiring affirmative or negative suppositions and very low logical solution rates throughout. Further analysis of the error data showed a pattern of responses, which suggested that participants reason from a superficial representation of the premises in these arguments and this drives their choice of conclusion. Our second experiment employed a different set of suppositional problems but with extremely similar proofs in terms of the rules applied and number of inferential steps required. As predicted by our interpretation of reasoning strategies employed in Experiment 1, logical performance was very much higher on these problems. Our third experiment showed that problems that could be solved by constructing an initial representation of the premises were easier than problems in which this representation was not sufficient. This effect was independent of the suppositional structure of the problems. We discuss the implications of this research for theories of reasoning based on mental models and inference rules.

Recently Lafont [6] showed the finite model property for the multiplicative additive fragment of linear logic (MALL) and for affine logic (LLW), i.e., linear logic with weakening. In this paper, we shall prove the finite model property for intuitionistic versions of those, i.e. intuitionistic MALL (which we call IMALL), and intuitionistic LLW (which we call ILLW). In addition, we shall show the finite model property for contractive linear logic (LLC), i.e., linear logic with contraction, and for its intuitionistic version (ILLC). The finite model property for related substructural logics also follow by our method. In particular, we shall show that the property holds for all of FL and GL - -systems except FL c and GL - c of Ono [11], that will settle the open problems stated in Ono [12].

Three experiments are reported in which the relationships between task format, item type, and strategy usage were investigated for a two-dimensional relational inference task. Contrary to past findings with linear syllogisms, it was found that parallel presentation (presenting problem statements simultaneously) did not result in any increased use of deduction rule processes compared with serial presentation (presenting problem statements individually). Instead, the results suggested that mental models were used by the majority of subjects, and that multiple models were more likely to be constructed with parallel presentation. It is proposed that, in general, multiple model construction will be more frequent for deduction tasks where the cognitive load is relatively low. Hence, contrary to suggestions by Polk and Newell (1995), reasoning in this way appears to be prevalent and highly robust - where supported by task format - even where the use of this strategy is disadvantageous.

This paper reports three studies of temporal reasoning. A problem of the following sort, where the letters denote common everyday events: A happens before B. C happens before B. D happens while B. E happens while C. What is the relation between D and EEfficacylls for at least two alternative models to be constructed in order to give the right answer for the right reason (D happens after E). However, the first premise is irrelevant to this answer, and so if reasoners were to ignore it, then they would need to construct only one model. Experiment 1 showed that one-model problems were answered faster and more accurately than multiple-model problems. When the question preceded the premises in the statement of the multiple-model problems there was a slight tendency for the latencies of response to speed up in the predicted way. Experiment 2 modified the procedure, in part by using practice problems with many irrelevant premises, so that reasoners might grasp the advantage of ignoring them. Its results showed that when the premises preceded the question, the multiple-model problems were significantly harder than one-model problems. But when the question was presented first, the difference was significantly reduced in line with the theory's prediction. Experiment 3 used only problems with valid conclusions (i.e., one-model problems and multiple-model problems), and so the construction of multiple models was never necessary. However, there was still a significant difference between one-model problems and multiple-model problems.

According to the mental models theory, reasoning performance is primarily influenced by the number of models of a problem that can be constructed. This study investigates whether the content of the model may also influence performance. Linear reasoning problems were devised that either described a believable (script-consistent) or an unbelievable (script-inconsistent) order of actions. The results of two experiments showed that conclusions were inferred more slowly and less accurately on the basis of an unbelievable model than on a believable one. Experiment 2 revealed that script knowledge facilitated as well as impeded reasoning performance. Conclusion evaluation was faster and more accurately for script-consistent problems than for neutral problems, whereas model construction and conclusion evaluation occurred respectively more slowly and less accurately for script-inconsistent problems than for neutral problems. These results show that the content of the model is a noteworthy factor influencing reasoning performance.