A Critique of Artificial Reason Hubert L. Dreyfus . HUBERT L. DREYFUS What Computers Still Can't Do Thi s One XZKQ-GSY-8KDG What. WHAT COMPUTERS STILL CAN'T DO Front Cover.

This unique book provides a gentle introduction to an increasingly important type of formal logic called CLAUSAL FORM LOGIC (CFL). Having evolved out of human reasoning, CFL represents the ideal to which all computer-based systems must approximate. Most present day computational logic systems are based on it, including the well-know artificial intelligence language PROLOG.

Computer simulations are widely used for surrogative reasoning in scientific research. They also play a crucial role in engineering, more specifically in the design of new robotic systems, yet the nature of this role has been little discussed so far in the philosophy of technology literature. The main claim made in this article is that the notion of surrogative reasoning is central to understanding how computer simulations can serve the purpose of designing new robots. More specifically, it is (...) argued that computer simulations can support two forms of surrogative reasoning, which are called model-oriented and prediction-oriented, whose inferential structure is reconstructed to some extent. And it is argued that, when computer simulations are used to design new robots, they are distinctively used in the model-oriented way. By unravelling the structure of the computer simulation-supported methods adopted in robotic design, this article may contribute to a finer-grained understanding of the epistemic processes involved in technological research. (shrink)

This article begins with an introduction to defeasible (nonmonotonic) reasoning and a brief description of a computer program, EVID, which can perform such reasoning. I then explain, and illustrate with examples, how this program can be applied in computational representations of ordinary dialogic argumentation. The program represents the beliefs and doubts of the dialoguers, and uses these propositional attitudes, which can include commonsense defeasible inference rules, to infer various changing conclusions as a dialogue progresses. It is proposed (...) that computational representations of this kind are a useful tool in the analysis of dialogic argumentation, and, in particular, demonstrate the important role of defeasible reasoning in everyday arguments using commonsense reasoning. (shrink)

Without inductive reasoning, we couldn't generalize from one instance to another, derive scientific hypotheses, or predict that the sun will rise again tomorrow morning. Despite the widespread nature of inductive reasoning, books on this topic are rare. Indeed, this is the first book on the psychology of inductive reasoning in twenty years. The chapters survey recent advances in the study of inductive reasoning and address questions about how it develops, the role of knowledge in induction, how (...) best to model people's reasoning, and how induction relates to other forms of thinking. Written by experts in philosophy, developmental science, cognitive psychology, and computational modeling, the contributions here will be of interest to a general cognitive science audience as well as to those with a more specialized interest in the study of thinking. (shrink)

Argument mapping is a way of diagramming the logical structure of an argument to explicitly and concisely represent reasoning. The use of argument mapping in critical thinking instruction has increased dramatically in recent decades. This paper overviews the innovation and provides a procedural approach for new teaches wanting to use argument mapping in the classroom. A brief history of argument mapping is provided at the end of this paper.

Theories of moral, and more generally, practical reasoning sometimes draw on the notion of coherence. Admirably, Paul Thagard has attempted to give a computationally detailed account of the kind of coherence involved in practical reasoning, claiming that it will help overcome problems in foundationalist approaches to ethics. The arguments herein rebut the alleged role of coherence in practical reasoning endorsed by Thagard. While there are some general lessons to be learned from the preceding, no attempt is made (...) to argue against all forms of coherence in all contexts. Nor is the usefulness of computational modelling called into question. The point will be that coherence cannot be as useful in understanding moral reasoning as coherentists may think. This result has clear implications for the future of Machine Ethics, a newly emerging subfield of AI. (shrink)

The idea that logic and reasoning are somehow related goes back to antiquity. It clearly underlies much of the work in logic, as witnessed by the development of computability, and formal and mechanical deductive systems, for example. On the other hand, a platitude is that logic is the study of correct reasoning; and reasoning is cognitive if anything Is. Thus, the relationship between logic, computation, and correct reasoning makes an interesting and historically central case study for (...) mechanism. The purpose of this article is to begin the articulation of this relationship, pointing out its sources and its limitations. (shrink)

Defeasible reasoning is concerned with the logics of non-deductive argument. As is described in the literature, the study of this type of reasoning is considerably more involved than the study of deductive argument, even so that, in realistic applications, there is often a lack of resources to perform an exhaustive analysis. It follows that, in a theory of defeasible reasoning, the order and direction in which arguments are developed, i.e. theprocedure, is important. The aim of this article (...) is to show that debate is the most efficient procedure to argue in the presence of limited resources. To do so, there is first some general theory on defeasible argumentation, which is followed by an introduction to the problem of dialectical search. The problem of dialectical search is (or at least, should be) the essential issue in every theory on argumentation, and emerges at every occasion that involves adjudication on competing arguments. Starting with an example, it is explained that dialectical search can be best scheduled according to classical debating techniques, that work along well-tried methods. These methods (which include various forms of curtailment, interruption, and interpretation) have proven their value in keeping debating efforts within reasonable bounds. How they apply in a theory of formal argument, will be shown in this article. (shrink)

Leibniz's overall view of the relationship between reasoning and computation is discussed on the basis of two broad claims that one finds in his writings, concerning respectively the nature of human reasoning and the possibility of replacing human thinking by a mechanical procedure. A joint examination of these claims enables one to appreciate the wide scope of Leibniz's interests for mechanical procedures, concerning a variety of philosophical themes further developed both in later logical investigations and in methodological contributions (...) to cognitive psychology. (shrink)

Partial functions are ubiquitous in both mathematics and computer science. Therefore, it is imperative that the underlying logical formalism for a general-purpose mechanized mathematics system provide strong support for reasoning about partial functions. Unfortunately, the common logical formalisms — first-order logic, type theory, and set theory — are usually only adequate for reasoning about partial functionsin theory. However, the approach to partial functions traditionally employed by mathematicians is quite adequatein practice. This paper shows how the traditional approach to (...) partial functions can be formalized in a range of formalisms that includes first-order logic, simple type theory, and Von-Neumann—Bernays—Gödel set theory. It argues that these new formalisms allow one to directly reason about partial functions; are based on natural, well-understood, familiar principles; and can be effectively implemented in mechanized mathematics systems. (shrink)

Sieg has proposed axioms for computability whose models can be reduced to Turing machines. This lecture will investigate to what extent these axioms hold for reasoning. In particular we focus on the requirement that the configurations that a computing agent (whether human or machine) operates on must be ’immediately recognisable’. If one thinks of reasoning as derivation in a calculus, this requirement is satisfied; but even in contexts which are only slightly less formal, the requirement cannot be met. (...) Our main example will be the Wason selection task, a propositional reasoning task in which in a typical (undergraduate) subject group only around 5% arrive at the answer dictated by classical logic. The instructions for this task (as well as other standard tasks in the psychology of reasoning, such as syllogisms) do not contain any ’immediately recognisable’ configurations. The subject must try to find an interpretation of the task by making the various elements in the instructions cohere, in effect solving a difficult constraint satisfaction problem, which has no unique solution. The subject has given a complete interpretation of the task if she can formulate the problem posed in the task as a theorem to be proved. The complexity of such theorems can be quite high; e.g. for the propositional Wason selection task the theorem can be in Σ1 3 . This sounds implausible, but we’ll present experimental data confirming this point. (shrink)

Stanovich & West interpret errors in syllogistic reasoning in terms of computational limitations. I argue that the variety of strategies used by reasoners in solving syllogisms requires us to consider also performance errors. Although reasoners' performance from one trial to another is quite consistent, it can be different, in line with the definition of performance errors. My argument has methodological implications for reasoning theories.

Theory of Mind (ToM), the cognitive capacity to attribute internal mental states to oneself and others, is a crucial component of social skills. Its formal study has become important, witness recent research on reasoning and information update by intelligent agents, and some proposals for its formal modelling have put forward settings based on Epistemic Logic (EL). Still, due to intrinsic idealisations, it is questionable whether EL can be used to model the high-order cognition of ‘real’ agents. This manuscript proposes (...) a framework that takes visibility, memory, perspective and communication as the main factors underpinning mental attributions. This is more in-line with findings in cognitive science and allows us to model well-known False-Belief Tasks that are typically used in testing people's ToM. We discuss some of the framework's technical features, arguing why it does justice to empirical observations. (shrink)

This article aims to achieve two goals: to show that probability is not the only way of dealing with uncertainty ; and to provide evidence that logic-based methods can well support reasoning with uncertainty. For the latter claim, two paradigmatic examples are presented: logic programming with Kleene semantics for modelling reasoning from information in a discourse, to an interpretation of the state of affairs of the intended model, and a neural-symbolic implementation of input/output logic for dealing with uncertainty (...) in dynamic normative contexts. (shrink)

The notion of an ideal reasoner has several uses in epistemology. Often, ideal reasoners are used as a parameter of (maximum) rationality for finite reasoners (e.g. humans). However, the notion of an ideal reasoner is normally construed in such a high degree of idealization (e.g. infinite/unbounded memory) that this use is unadvised. In this dissertation, I investigate the conditions under which an ideal reasoner may be used as a parameter of rationality for finite reasoners. In addition, I present and justify (...) the research program of computational epistemology, which investigates the parameter of maximum rationality for finite reasoners using computer simulations. (shrink)

This article aims to achieve two goals: to show that probability is not the only way of dealing with uncertainty ; and to provide evidence that logic-based methods can well support reasoning with uncertainty. For the latter claim, two paradigmatic examples are presented: logic programming with Kleene semantics for modelling reasoning from information in a discourse, to an interpretation of the state of affairs of the intended model, and a neural-symbolic implementation of input/output logic for dealing with uncertainty (...) in dynamic normative contexts. (shrink)

Following Marr (1982), any computational account of cognition must satisfy constraints at three explanatory levels: computational, algorithmic, and implementational. This paper focuses on the first two levels and argues that current theories of reasoning cannot provide explanations of everyday defeasible reasoning, at either level. At the algorithmic level, current theories are not computationally tractable: they do not “scale-up” to everyday defeasible inference. In addition, at the computational level, they cannot specify why people behave as they (...) do both on laboratory reasoning tasks and in everyday life (Anderson, 1990). In current theories, logic provides the computational-level theory, where such a theory is evident at all. But logic is not a descriptively adequate computational-level theory for many reasoning tasks. It is argued that better computational-level theories can be developed using a probabilistic framework. This approach is illustrated using Oaksford and Chater's (1994) probabilistic account of Wason's selection task. (shrink)

In the paper we examine the use of non-classical truth values for dealing with computation errors in program specification and validation. In that context, 3-valued McCarthy logic is suitable for handling lazy sequential computation, while 3-valued Kleene logic can be used for reasoning about parallel computation. If we want to be able to deal with both strategies without distinguishing between them, we combine Kleene and McCarthy logics into a logic based on a non-deterministic, 3-valued matrix, incorporating both options as (...) a non-deterministic choice. If the two strategies are to be distinguished, Kleene and McCarthy logics are combined into a logic based on a 4-valued deterministic matrix featuring two kinds of computation errors which correspond to the two computation strategies described above. For the resulting logics, we provide sound and complete calculi of ordinary, two-valued sequents. (shrink)

Inventing novel knowledge to solve problems is a crucial, creative, mechanism employed by humans, to extend their range of action. In this talk, I will show how commonsense reasoning plays a crucial role in this respect. In particular, I will present a cognitively inspired reasoning framework for knowledge invention and creative problem solving exploiting TCL: a non-monotonic extension of a Description Logic (DL) of typicality able to combine prototypical (commonsense) descriptions of concepts in a human-like fashion. The proposed (...) approach has been tested both in the task of goal-driven concept invention and has additionally applied within the context of serendipity-based recommendation systems. I will present the obtained results, the lessons learned, and the road ahead of this research path. (shrink)

In this paper I attempt to cast the current program verification debate within a more general perspective on the methodologies and goals of computer science. I show, first, how any method involved in demonstrating the correctness of a physically executing computer program, whether by testing or formal verification, involves reasoning that is defeasible in nature. Then, through a delineation of the senses in which programs can be run as tests, I show that the activities of testing and formal verification (...) do not necessarily share the same goals and thus do not always constitute alternatives. The testing of a program is not always intended to demonstrate a program's correctness. Testing may seek to accept or reject nonprograms including algorithms, specifications, and hypotheses regarding phenomena. The relationship between these kinds of testing and formal verification is couched in a more fundamental relationship between two views of computer science, one properly containing the other. (shrink)

The emergence of new computational media is radically changing the practices of science, particularly in the way computational models are built and used to understand and engineer complex biological systems. These new practices present a novel variation of model-based reasoning, based on dynamic and opaque models. A new cognitive account of MBR is needed to understand the nature of this practice and its implications. To develop such an account, I first outline two cases where the building and (...) use of computational models led to discoveries. A theoretical model of the possible cognitive and neural mechanisms underlying such discoveries is then presented, based on the way the body schema is extended during tool use. This account suggests that the process of building the computational model gradually ‘incorporates’ the external model as a part of the internal imagination system, similar to the way tools are incorporated into the body schema through their active use. A central feature of this incorporation account is the critical role played by tacit and implicit reasoning. Based on this account, I examine how computational modeling would change model-based reasoning in science and science education. (shrink)

This study examines the relationship between cognitive moral development (CMD), productivity features of information technology (IT) and unethical behavior or misconduct. Using an experimental design that randomly assigns subjects to one of four unique technology conditions, we assess the relationship between a subjects' predominant level of CMD and ethical misconduct on IT-oriented work tasks. Our results show that both higher levels of CMD and increased levels of IT productivity features at one's disposal have a significant role to play in explaining (...) observed behavior in our sample. We find that CMD as measured by the Defining Issues Test's P-score is negatively related to task misconduct. Conversely, IT productivity features such as copy-and-paste are positively related to task misconduct. In addition, the CMD—misconduct relationship is significantly diminished by the introduction of IT productivity features. Lastly, a series of hazard analyses are conducted to explore the boundaries of our principal findings. These results demonstrate the significant role of technology in enabling negative behavior and the relative inability of subjects' use of principled moral reasoning to overcome it. Implications of these findings for academics and business managers are offered, as well as recommendations for mitigating misconduct in both academic and workplace environments. (shrink)

Empirical insights into language processing have a philosophical relevance that extends well beyond philosophical questions about language. This chapter will discuss this wider relevance: We will consider how experimental philosophers can examine language processing in order to address questions in several different areas of philosophy. To do so, we will present the emerging research program of experimental argument analysis (EAA) that examines how automatic language processing shapes verbal reasoning – including philosophical arguments. The evidential strand of experimental philosophy uses (...) mainly questionnaire-based methods to assess the evidentiary value of intuitive judgments that are adduced as evidence for philosophical theories and as premises for philosophical arguments. Extending this prominent strand of experimental philosophy, EAA underpins such assessments, extends the scope of the assessments, and expands the range of the empirical methods employed: EAA examines how automatic inferences that are continually made in language comprehension and production shape verbal reasoning, and draws on findings about comprehension biases that affect the contextualisation of such default inferences, in order to explain and expose fallacies. It deploys findings to assess premises and inferences from premises to conclusions, in philosophical arguments. To do so, it adapts methods from psycholinguistics and recruits methods from computational linguistics. (shrink)