Working memory limits are best defined in terms of the complexity of the relations that can be processed in parallel. Complexity is defined as the number of related dimensions or sources of variation. A unary relation has one argument and one source of variation; its argument can be instantiated in only one way at a time. A binary relation has two arguments, two sources of variation, and two instantiations, and so on. Dimensionality is related to the number of chunks, because (...) both attributes on dimensions and chunks are independent units of information of arbitrary size. Studies of working memory limits suggest that there is a soft limit corresponding to the parallel processing of one quaternary relation. More complex concepts are processed by or In segmentation, tasks are broken into components that do not exceed processing capacity and can be processed serially. In conceptual chunking, representations are to reduce their dimensionality and hence their processing load, but at the cost of making some relational information inaccessible. Neural net models of relational representations show that relations with more arguments have a higher computational cost that coincides with experimental findings on higher processing loads in humans. Relational complexity is related to processing load in reasoning and sentence comprehension and can distinguish between the capacities of higher species. The complexity of relations processed by children increases with age. Implications for neural net models and theories of cognition and cognitive development are discussed. (shrink)

The core issue of our target article concerns how relational complexity should be assessed. We propose that assessments must be based on actual cognitive processes used in performing each step of a task. Complexity comparisons are important for the orderly interpretation of research findings. The links between relational complexity theory and several other formulations, as well as its implications for neural functioning, connectionist models, the roles of knowledge, and individual and developmental differences, are considered.

An application of the Method of Analysis of Relational Complexity (MARC) to suppositional reasoning in the knight-knave task is outlined. The task requires testing suppositions derived from statements made by individuals who either always tell the truth or always lie. Relational complexity (RC) is defined as the number of unique entities that need to be processed in parallel to arrive at a solution. A selection of five ternary and five quaternary items were presented to 53 psychology students using a pencil (...) and paper format. A computer-administered version was presented to 50 students. As predicted, quaternary problems were associated with higher error rates and longer response times than ternary problems. The computer-administered form was more difficult than the pencil and paper version of the test. These differences are discussed in terms of RC theory and alternative processing accounts. Together, they indicate that the relational complexity metric is a useful and parsimonious way to quantify complexity of reasoning tasks. (shrink)

Current conceptions of the nature of human reasoning make it no longer tenable to assess children's inference by reference to the norms of logical inference. Alternatively, the complexity of the mental models employed in children's inferences can be analysed. This approach is applied to transitive inference, class inclusion, categorical induction, theory of mind, oddity, categorical syllogisms, analogy, and reasoning deficits. It is argued that a coherent account of children's reasoning emerges in that there is correspondence between tasks at the same (...) level of complexity across different domains, and that the inferences of younger children, while impressive and important, are consistently simpler than those of older children. (shrink)

We argue that if a different definition of sentence complexity is adopted and processing capacity is assessed in a way that is consistent with that definition, then the Caplan & Waters distinction between interpretive versus postinterpretive processing is unnecessary insofar that it applies to the thematic role assignment in relative-clause sentences.

Bayesian rationality is an important contribution to syllogistic inference, but it has limitations. The claim that confidence in a conclusion is a function of informativeness of the max-premise is anomalous because this is the least probable premise. A more plausible account is that confidence is inversely related to complexity. Bayesian rationality should be supplemented with principles based on cognitive complexity.

Cowan's review shows that a short-term memory limit of four items is consistent with a wide range of phenomena in the field. However, he does not explain that limit, whereas an existing theory does offer an explanation for capacity limitations. Furthermore, processing capacity limits cannot be reduced to storage limits as Cowan claims.

The distinction between uninformed learning (type-1) and learning based on recoding using prior information (type-2) helps to clarify some long-standing psychological problems, including misunderstanding of mathematics by children, the need for active construction of concepts in cognitive development, and the difficulty of configural learning tasks. However, an alternative to recoding some type-2 tasks is to represent the input as separate dimensions, which are processed jointly. This preserves the original structure, but is subject to processing capacity limitations.

We propose that the missing link from nonhuman to human cognition lies with our ability to form, modify, and re-form dynamic bindings between internal representations of world-states. This capacity goes beyond dynamic feature binding in perception and involves a new conception of working memory. We propose two tests for structured knowledge that might alleviate the impasse in empirical research in nonhuman animal cognition.