In September 2018 I wrote two blog posts about Baron’s search-inference framework (here) as described in chapter 1 of Baron’s textbook (1994 2nd edition) on human decision-making entitled Thinking and Deciding and about thinking and rationality (here), which corresponds to chapters 2 and 3 in the same book. I felt that I should read the rest of the book and decided that I would make abstracts of its chapters in the process. Here follows the rest of Part I of the book which reviews basic concepts for the theory of thinking and Baron’s reevaluation of the traditional theory of logic. Part I is entitled ‘Thinking in General’, so I used the same title for this blog post. I recommend everybody to read the original, which is both enlightening and highly readable. I must warn though that there is a fourth edition, which is much revised. In the next two blog posts I will cover Parts II and III.
Problem solving Common ways of problem solving include trial-and-error as well as insight. Trial and error may be blind to or guided by insight. Insight is characterized by the sudden recognition of a solution. This may be triggered by either a ‘suggestion from below’ (the solution is present, e.g. Archimedes’ Eureka moment) or ‘a suggestion from above’ (an imagined solution). The distinction between the two can be difficult. Assumptions and habits in problem solving can be useful but may also lead to functional fixedness or mechanization, respectively, which prevent us from expanding the search for possibilities. Computer-based problem solving by Simon et al. has not only led to ideas about problem space and stepwise evaluation, but also to a simple classification of solution methods, including trial and error, hill climbing, means-ends analysis, and working back from sub-goals. The appropriateness of each method depends on the knowledge available, which depends on the problem and on the learning ability of the problem solving agent. Expert problem solvers have learned to recognize problems and are able to work forward directly from the givens. Heuristics or heuristic methods such as those described by Polya (1945) generate suggestions by ‘tickling memory’ to facilitate more extensive search for useful possibilities and evidence. In general, beliefs about the efficacy of thinking are important. (Ch. 4 of Baron, J. 1994 Thinking and deciding 2nd ed.)
Basic learning processes This chapter concerns the basic mechanisms underlying all learning and discusses the nature of memory as well as the role of learning strategies. Learning is essential to thinking and deciding. The limitations of short-term, primary or working memory are pointed out. Decision making or design judgment (i.e. the consideration of possibilities, evidence and goals) requires its interplay with secondary memory. Chunking can expand the limitations of primary memory, but requires an additional effort so it will not work In novel situations. Encoding and association in memory follow certain principles. Understanding is useful in retrieving interconnected information from our memory and so is classification (dogs, cats), which may also be instrumental in language development. Practice is of limited use for memory skills. Memorization strategies for lists of items and other simple material may involve rehearsal, association, and categorization. (Ch. 5 of Baron, J. 1994 Thinking and deciding 2nd ed.)
Learning of complex material This chapter concerns the learning of complex material of the sort that requires understanding rather than memorizing. Piaget’s theory of learning stages with general restructuring from one stage to another seems false. Instead there is rather a restructuring in specific domains. This is called learning by understanding, more typically of mature theories instead of naïve ideas. Learning by understanding is not the same as learning by induction nor is it the same as learning by algebra. Complex knowledge such as a language or geometry is organized so learning it likely follows the laws of transfer. Learning by understanding requires seeing the structure. Understanding is expanded by Perkins (1986) to involve knowing: (1) the structure of what we what to understand; (2) the (sub)purposes of the structure; and (3) the arguments about why the structure serves the purpose. A design is anything whose structure serves a purpose. Heuristics are designs. We can think of knowledge-to-learn as design. This can be fitted to the search-inference framework. Interesting is that learning with understanding resists the learning of falsehoods. Learning good habits is another topic. It is often associated with rewards. This may be counterproductive, as some of our personal goals involve the pursuit of rewards, while others do not. As a general rule educators should use a minimum extrinsic reward approach. (Ch. 6 of Baron, J. 1994 Thinking and deciding 2nd ed.)
Intelligence and creativity This chapter discusses the role of thinking in intelligence and creativity. Baron defines intelligence as those general mental abilities that help people achieve their goals, whatever those goals may be, in any real environment. Intelligence is different from more special abilities, which have to do with expertise. Certain abilities that constitute intelligence are fixed capacities (mental speed, good memory), while others are not since they are learned dispositions (search thoroughly). Children can be trained to become more reflective. Schooling seems to be generally beneficial, not just in the basic skills (reading arithmetic etc.), but also in developing more reflective thinking. Many doubt the possibility of a cross-culturally valid conception of intelligence. Creativity plays a role by looking for new goals or new solutions for old goals. The unconscious is often important. Some incubation may be needed. Other explanations for novel ideas are equally plausible. Thorough search for possibilities and goals, as well as self-critical search for evidence are crucial to success. There are neither good IQ tests nor good tests for creativity. What makes for novelty is largely the desire to create. (Ch. 7 of Baron, J. 1994 Thinking and deciding 2nd ed.)
The teaching of thinking Schools, families, institutions of society, business managers, journalists, politicians, and of course individuals all have a role to play in the encouragement of good thinking. We need cultures of inquiry, intellectual challenge and good thinking so that students will want to think well. Good thinking habits must be instilled, but so must the arguments for using them. The wrong beliefs about bad thinking must be exposed, including the unhelpful belief that changing one’s mind is a sign of weakness, or the belief that use of intuition alone is the best way to make decisions, or the belief that experts will sort it out, or the belief that one’s thinking is not effective. Standards and beliefs are useless unless a student has the goal of discovering the truth and making good decisions. This requires a strong commitment to “intellectual honesty” and a tolerance of ambiguity. The goal of being a good citizen is also important. Transfer of learning is achieved by using general thinking methods, including statistical methods and methodological or logical reasoning. The tutorial method and thinking assignments are good methods for teaching good thinking in general or in specific cases. These methods may be integrated in existing lessons. By way of background, some of the past ideas in the educational movement for critical and reflective thinking are reviewed, from John Dewey (1933) to de Bono and Paul (1984). (Ch. 8 of Baron, J. 1994 Thinking and deciding 2nd ed.)
Formal logic Nowadays, formal logic is no longer considered the normative model of thinking, which it was taken to be in the past. Logic is a normative model of inference, arrived at by reflection about arguments. It is concerned with the rules for drawing conclusions from evidence with certainty. Logicians have developed several systems of formal logic: propositional logic, categorical logic, predicate logic, modal logic. No sharp boundary separates modern logic from semantics, the part of modern linguistics that deals with meaning. Some propositional logic is self-evident and is also called natural logic. Categorical logic is much more difficult. An alternative approach is to form a mental model and derive a tentative conclusion by examining the model. Several types of logical errors in hypothesis testing or failures or biases in hypothesis elimination are discussed, including: (1) poor thinking; (2) resistance to correction; (3) rationalization; (4) content effect; and (5) effects of prior belief. Formal logic is not a complete theory of thinking. Because logic covers only inference, it cannot help us understand errors that result from insufficient search. Another major limitation is that it deals only with conclusive arguments. Few of the rational conclusions in daily life fall in this category. (Ch. 9 of Baron, J. 1994 Thinking and deciding 2nd ed.)
Logic and everyday reasoning An important argument for logic as a design is that it does not depend on personal, subjective judgments. However, because it has no concept of “weight” of evidence or “strength” of possibilities, logic provides no way to compare two competing arguments that are both valid. It also provides no way to distinguish reasonable from unreasonable conclusions. More generally, there are too few cases in which we can accept premises with certainty or in which the rules of logic can lead us to conclusions that we would have reached anyway. Attempts have been made to make logic more relevant to everyday reasoning. One way is to attempt to enumerate fallacies of reasoning in logical terms. Many of these were pointed out by Aristotle and include non sequitur arguments such as the argument from ignorance or the appeal to multitude. Others include the strawman fallacy and begging the question. The traditional list of fallacies does not seem to be of much use nowadays. Most logical errors committed are often best characterized as the more or less wilful ignorance of evidence that goes against the reasoner’s conclusion. In Baron’s view many of the so-called fallacies in real life are cases of overweighing the evidence. Good arguments use relevant, but mostly not decisive, evidence. Poor arguments typically ignore some relevant possibility, goal, or piece of evidence. Standard rules of discussion in persuasion dialogues are explained as is the Toulmin argumentation scheme. (Ch. 10 of Baron, J. 1994 Thinking and deciding 2nd ed.)