Reasoning, Realism and Experience, The Case of Posterior Judgements Essays

Reasoning, Realism and Experience, The Case of Posterior Judgements Essays Reasoning, Realism and Experience, The Case of Posterior Judgements Essay Reasoning, Realism and Experience, The Case of Posterior Judgements Essay Reasoning is a key area in cognitive research, involving the use of logical thinking to find results or conclusions. The topic consists of two main areas: deductive and inductive reasoning. Realism is another area of focus, this is the theory that every statement is true or false regardless of whether this can be verified. One final area is experience, an alternative focusing on knowledge acquired through the senses rather than abstract reasoning. The nature of this study is to investigate the relationship between these three areas and posterior judgements. Such judgements are formed through assessing the likelihood of an event by updating a prior probability in light of new evidence. Reasoning is one of the oldest research topics in cognitive psychology. For Eysenck and Keane (2000) [1], a key question regarding human nature and reasoning is Are human beings rational? . Many philosophers believe that the laws of logic are the laws of thought and therefore reply yes to such a question. The psychology of reasoning has progressively developed since early research. Researchers have often drawn distinctions between two models of reasoning, deductive and inductive reasoning. Both relate to types of decisions made about particular instances or premises. A premise is formed when a number of propositions are related together by a logical operator. Eysenck (1993) [3], states that inductive reasoning is a form of reasoning whereby a generalised conclusion is drawn from specific information, therefore the conclusion cannot be shown to be necessarily true. He defines deductive reasoning as a form of reasoning in which definite conclusions follow on the basis that certain statements or premises are assumed to be true. In order to develop a greater understanding of these distinctions they need top be researched further. According to Johnson-Laird and Byrne (1991) [1] deductive reasoning is a central intellectual ability. This ability is vital for various human functions such as: formulating plans; determining consequences; interpreting and formulating instructions; pursuing arguments and solving problems. For Eysenck and Keane (2000) [1], a world without deduction would be a world without science, technology, laws, social conventions and cultures. Deductive reasoning makes use of logical systems to characterise the abstract structure of reasoning problems. One particular logical system used is the propositional calculus, a logic where propositions are manipulated using a small set of logical operators, for example, if. then. Eysenck and Keane (2000) [1], change their earlier question slightly to investigate deductive research, it becomes Are humans logical? . In simple terms, do people conform to logical interpretations such as if. hen, and if so, will they make valid inferences and reject invalid inferences provided by the propositional calculus. In Problem-Solving research, Newell and Simon (1972) [1] devised the problem-space theory. This takes the notion of an idealised problem space to characterise abstract structures of problems independently of any psychological proposals (Eysenck and Keane, 2000 [1]). Some logics have been used in a similar way in reasoning research. Such logics are devised to characterise the abstract structure of reasoning problems and to determine categories of responses (i. . correct or incorrect responses). Logical systems are similar to mathematical systems in that symbols are used to represent things, for example, the length of a car is represented by L1 and the length of a bus is represented by L2. Mathematical operators can then be used to manipulate the two variables and produce a new statement. In the case of reasoning, logical symbols are used in place of sentences and logical operators such as: not, or and if. then, if and only if, are used to manipulate the situation. Although logical operators use common words, it is essential to remember that they all have very different meanings. Eysenck and Keane (2000) [1], demonstrate the use of logical operators in more detail. Using the propositional calculus, they choose the letter P to represent the sentence If it is raining, and Q to represent Alicia gets wet. A logical operator is then applied to relate the two together creating: If P then Q, therefore If it is raining, then Alicia gets wet. Truth tables are used to determine conclusions from such logical statements as provided by Eysenck and Keane (2000) [1]. In logical systems such as the one they provided, only one of two truth values are possible, these being true and false. P can only be true or false because in the statement it is raining, therefore it is either raining or it isnt. The truth tables lay out the possibilities for a proposition and consequently explain how logical act on that proposition. The tables make it possible to define valid and invalid inferences. If someone concludes that if P then Q and P as a valid inference, this is called a modus ponens. If they conclude if P then Q and not Q as a valid inference it is known as a modus tollens. Many people make a modus ponens, however not many people are willing to state a modus tollens. For Eysenck and Keane (2000) [1], the importance of the logical analysis presented here is that it allows us to characterise the abstract structure of reasoning problems and gives us a criterion for determining whether certain conclusions are valid or invalid, correct or in error. In 1993, Eysenck [3], stated that affirmation of the consequent and denial of the antecedent are important matters of focus. A demonstration of affirmation of the consequent is : Premises: If it is raining, then Alicia gets wet. Alicia gets wet. Conclusion: Therefore, it is raining. A demonstration of denial of the antecedent is: Premises: If it is raining, then Alicia gets wet. It is not raining Conclusion: Alicia does not get wet. Evans, 1989 [3], stated that most people regard these conclusions as being valid. They are in fact, invalid. In the first example, it does not need to be raining for Alicia to get wet she may have been swimming or taken a shower. This is also applicable for the second example. Therefore, it is evident that deductive reasoning is prone to error when it comes to affirmation on the consequent and denial of the antecedent. Evans (1989 [3]) found that few errors are made with modus ponens but that error rates for modus tollens often exceed thirty per cent. There is no clear definition as to why such errors are made with modus tollens. It is thought to be partly due to a lack of practice in thinking about what is not the case (Eysenck 1993 [3]). The key element in research on deductive reasoning is whether or not people think rationally and logically. Henle (1962) [3], stated that consistent errors may be a result of people misunderstanding the question, even if they apply logical thinking to it. She also claimed that some errors were due to the subjects failure to accept the logical task. Braine, Reiser and Rumain, 1984 [3], developed Henles theory further. According to their natural deduction theory, most errors found in deductive reasoning occur due to a failure of comprehension. For Braine et al. (1984), people have a mental rule corresponding to modus ponens. As a result, premises based on modus ponens are easier to handle and therefore pose no comprehension problems. Deductive reasoning research covers a wide variety of tasks, any adequate theory of deduction should be able to explain the phenomena rising from such research (Eysenck and Keane 2000 [1]). Two main theories meet such a challenge, the Abstract-Rule theory and Mental Models theory. The Abstract-Rule theory assumes people reason validly by applying abstract, content-free rules of inference. It suggests people adopt a mental logic in order to make conclusions from statements or premises. Evidence from conditional reasoning shows that people are not completely rational, invalid inferences are often made in place of valid inferences. The Abstract-Rule theory proposes that humans use sets of comprehensive rules and apply them to any area of knowledge. The theory was used by Braine et al. (1984 [3]) in demonstrating that people only make invalid inferences due to a lack of understanding for the logical task. A representative case is that of Braine and OBriens (1991 [3]) Abstract-Rule theory. This theory states that deductive reasoning is mediated by basic abstract rules. The premises or arguments are encoded into abstract rules and inferences are then created. It predicts that people are natural logicians who are slightly fallible at the edges (Eysenck and Keane 2000 [1]). Most abstract rule theories have a reasoning rule corresponding to the modus ponens and the modus tollens is a harder inference to make due to the fact that no single rule can be applied to it. For Eysenck and Keane (2000) [1] people still apply their logically valid rules but because the input to the rules is erroneous, the output is often erroneous too. The Mental Models theory assumes that people reason by manipulating mental models of a set of premises, in a similar manner to semantic methods of proof in logic. This theoretical approach on deductive reasoning was proposed by Johnson-Laird in 1983 [1]. In simple form, the model is a representation of the state of affairs described in the premises of a problem and it may be in the form of imagery (Eysenck 1993 [3]). Such a representation depends on the interpretation of the premises. The Mental Model differs to the Abstract-Rule model because it creates a central role for comprehension in reasoning. Humans develop models through their comprehension of linguistic description, their description is therefore reliable on these models. Eysenck and Keane demonstrate a construction of a mental model (1990 [3]). Premises: The lamp is on the right of the pad. The book is on the left of the pad. The clock is in front of the book. The vase is in front of the lamp. Conclusion: The clock is to the left of the vase. Johnson-Laird (1983 [3]) believed people construct such a model in a simplified form, using the information contained in the premises: Book Pad Lamp Clock Vase It is often the case that people use more than one model in consistence with the premises. A second model often constructed differs slightly from the one above: Lamp Pad Book Vase Clock Johnson-Laird (1983 [3]) states that someone who constructed only the first mental model would mistakenly conclude the clock to be on the left of the vase. It would be evident to someone who constructed both models that the clock is not necessarily to the left of the vase. Eysenck (1993 [3]) summarises Johnson-Lairds Mental Model theory in the following points. Firstly, comprehension of the premises of a problem leads to construction of one or more mental models. Secondly, the model or models constructed are used to produce novel conclusions not specified by the premises directly. He stated that there is a check to decide whether there are any additional models to invalidate conclusions. Finally, the above three processes all depend on the processing, resources of working memory. It can therefore be affected by limited capacity. Research shows that deduction has received vas amounts of attention. It is evident that people construct mental models or constructions to try and resolve a situation by making the correct inference. Posterior judgements involve incorporating new evidence to update previous judgements. For Lance Rips (1994 [5]), one reason deduction has played a role in cognitive psychology is that it has been difficult for psychologists to envision what purpose deduction serves. Logical operators are combined with variables or premises to provide a model enabling cognitive processes such as problem solving or categorisation. Rips (1994 [5]) states that categorisation is of importance due to the fact that beliefs about category membership are not deducible from evidence available to us. It is more often the case that evidence provides an inductive warrant for categorising, as in more judgemental situations. It is clear that as people receive more information and evidence they build constructions to represent problems which may or may not lead to their predictions and previous judgements changing. The alternative aspect to deductive reasoning is that of inductive reasoning. Eysenck (1993 [3]) states that much of the research on inductive reasoning has been concerned with concept learning. Bourne, 1966 [3], described a concept as existing whenever two or more distinguishable objects or events have been grouped or classified together and set apart from other objects on the basis of some common feature or property characteristic of each. Bruner, Goodnow and Austin, 1956 [3], conducted a well known piece of research on concept learning. They used stimuli consisting of rectangular cards picturing various shapes. The cards varied in four dimensions as follows: the number of borders around the edges, the number of shapes in the centre of the cards, the shapes themselves and the colour of the shapes. Bruner et al. (1956 [3]) used typically conjunctive concepts in their experiment, it involved a number of features being presented together to produce a positive card, for example, three black circles. Many of their studies employed a selection paradigm. The subjects were offered all cards and selected one at a time, the concept was not revealed to them. After each selection they were told whether they had chosen a positive or negative instance of the concept. They could volunteer hypothesis to the experimenter about the concept of the experiment. Subjects appeared to use limited strategies, one being conservative focusing. This is focusing on a first positive instance and then choosing a following card that differs in only one attribute. If this card is also positive then the attribute changed is clearly irrelevant to the concept. Yet, if the second card chosen is a negative instance, then the attribute which varied is part of the concept. Another strategy used is successive scanning. This strategy is used to begin with a specific hypothesis which subjects attempt to test by selecting cards that will provide useful information. Bruner et al (1956 [3]) discovered that focusing was generally more successful than scanning because fewer cards needed to be selected before the concept was identified. Wason (1960 [1,3]) devised an interesting approach to concept learning resembling the work of Bruner et al. (1956 [3]) and their selection paradigm. His task involved four cards lying on a table, each card had a letter on one side and a number on the other. The subject is informed of a rule applicable to all four cards. For example, if there is an R on one side of the card then there is a 2 on the other side of it. Wasons task was to select only the cards that would not need to be turned over to determine the if rule is correct or not. The findings of this task were taken as evidence to confirm a persons tendency to confirm hypotheses in reasoning situations, although it is considered valuable, there is still a lot of controversy about its utility as a tool to examine human reasoning. Inductive reasoning shows a generalised conclusion is drawn from specific information, the conclusion cannot be proven as true. In the case of posterior judgements, inductive reasoning would have not have much effect on previous probabilities to the same extent as deductive reasoning. Specific judgements are made and no concrete conclusion is produced.