Domainspecific Rule Theories

Relative to the other theories we have just reviewed, the domain-specific rule theories have considerably less coverage. Recall, that most of these theories espouse a dual-process view of reasoning; namely, that some reasoning component handles deductive competence which is supplemented by domain-specific rules (Cosmides' theory is one exception to this view). Thus, in this section, we will not discuss how these theories explain the basic conditional inference patterns and context effects (because they are silent on these) but rather deal solely with the selection-task effects they have been specifically designed to explain. Of course, on parsimony grounds, these theories are at a disadvantage relative to a single-process account that can explain these and other phenomena.

The selection task and domain-specific rules

Domain-specific-rule theories are a family of theories that have arisen to explain the various thematic or deontic effects in the selection task. These theories differ in the content of the domainspecific rules they posit. Some theories maintain that the content is very specific prior experience used by analogy (see Griggs, 1983; and Chapter 15), but most of these theories hinge on the use of schemata that are specific to classes of situations (e.g., permission situations or various contractual situations). Even though these schematic rules are relatively abstract, they are more specific than the content-free, abstract rules we met earlier. We consider two variants of domain-specific-rule theories.

Pragmatic reasoning schemata: Domain-specific rules for permissions and obligations

Cheng and Holyoak (1985; Cheng, Holyoak, Nisbett, & Oliver, 1986) call their domain-specific rules pragmatic reasoning schemata because they are sensitive to the pragmatics of the situation (see Chapter 9 on schemata). Permission situations are one class of such situations that occur regularly in everyday life; for example, to gain permission to enter university you must satisfy the precondition of achieving a certain exam result. The schemata for such situations are abstract in that they potentially apply to a wide range of content domains, but unlike abstract rules they are constrained by particular inferential goals and event relationships of certain broad types (see Holland et al., 1986). Pragmatic reasoning theory proposes that:

• People have specific rules that deal with particular types of situation; for instance, permission schemata and obligation schemata.

• Rules in the permission schemata take the form "If an action is to be taken, then a precondition must be satisfied"; if you are asked to test a rule that elicits a permission schema then the appropriate rule from this schema is applied (its logic is like that of the conditional in Table 16.1).

• Obligation schemata have also been elaborated for situations in which one is obliged to do something.

In this theory, errors occur when situations cannot be mapped easily into pragmatic schemata, or errors may arise directly from the inferences generated by schemata (because the rules in the schema may not conform to those sanctioned by propositional logic). Cheng and Holyoak have also suggested that in situations where the schemata do not apply (e.g., abstract content) abstract rules and other strategies may come into play.

Social contract theory: Domain-specific rules encoding contracts

Cosmides (1989) has an alternative domainspecific-rule theory based on an evolutionary approach to cognition. She suggests that people have rules—called Darwinian algorithms—that maximise their ability to achieve their goals in social situations. She concentrates on situations involving social exchange, where two people must co-operate for mutual benefit. The social contract theory proposes that people have schemata pertaining to these sorts of social contracts, such as schemata that encode the following rules:

Standard social contract: "If you take a benefit then you must pay the cost" Switched social contract: "If you pay the cost then you take the benefit"

Along with these schemata, for evolutionary reasons, people must have a mechanism for detecting those who might break a contract: a "look for cheaters" algorithm. If the standard contract schema is applied to the selection task, along with the cheater-detection algorithm, then people should make the correct selections (of P and not-Q), but when the switched-contract schema is applied they will make less optimal selections (taking the logical view of rationality). Social contract situations are a subset of permission situations, so those permission situations that are social contracts will show facilitation, but permission situations that are not social contracts will not. The postal rules used in some realistic contents (see earlier) can be viewed as reified social contracts, as can the rules about ages at which people can legally drink alcohol, which have been used in some realistic-content studies. Furthermore, instructions to find violators of the rule can be viewed as a means to facilitate the application of the cheater-detection algorithm (see also Gigerenzer & Hug, 1992, for a further elaboration of this theory; and Cummins, 1996).

Following on Cheng and Holyoak's (1985) results, Cosmides examined cases of permission situations that were not social contracts and found less facilitation than in permission situations that were social contracts (see also Gigerenzer & Hug, 1992). However, the levels of correct responses in permission cases that were not social contracts were still higher than the usual levels on the abstract task. So, while social contract situations appear to be highly facilitating, there is also some facilitation for pure permission situations. Although many of Cosmides' results have been replicated and further clarified (see Platt & Griggs, 1993), some researchers contested her claims. For example, some have disputed her interpretation that some realistic-content versions of the task are, in fact, social contracts; for example, the Sears' receipts version of the task (see earlier) seems to have no obvious contractual component to it (see Cheng & Holyoak, 1989; Gigerenzer & Hug, 1992). As we saw earlier, there is a growing consensus that these effects are due to the deontic aspects of the tasks, rather than whether they are specifically permissions or contracts, although this does not completely invalidate these theories as accounts of such deontic effects.

Evaluation of domain-specific-rule theories

In general, domain-specific-rule or schema theories are an interesting addition to the theoretical corpus of research on the selection task. They have been used to generate new predictions about the underlying basis of subjects' performance on the task and many of these predictions have been confirmed.

The major criticism of this view is that they are not complete theories of reasoning; they can characterise some versions of the selection task, but do not go far beyond this point. For example, they make no predictions for other logical connectives (e.g., and and or) and have no account of the pattern of selections on the abstract task. In other words, these theories are silent on what people are doing when they are not using domain-specific rules. Cheng and Holyoak suggest that some responses may be due to non-logical biases but they do not elaborate this further. O'Brien (1993) has suggested a marriage between the abstract-rule account and Cheng and Holyoak's theory.

Finally, the theories themselves have many underspecified components. Both pragmatic-rule and social-contract theory do not specify how natural language is parsed into the rules used for reasoning. So many aspects of their predictions could be open to ad-hoc interpretation.

Stop Anxiety Attacks

Stop Anxiety Attacks

Here's How You Could End Anxiety and Panic Attacks For Good Prevent Anxiety in Your Golden Years Without Harmful Prescription Drugs. If You Give Me 15 minutes, I Will Show You a Breakthrough That Will Change The Way You Think About Anxiety and Panic Attacks Forever! If you are still suffering because your doctor can't help you, here's some great news...!

Get My Free Ebook


Post a comment