Humans are often extraordinary at performing practical reasoning. There are cases where the human computer, slow as it is, is faster than any artificial intelligence system. Are we faster because of the way we perceive knowledge as opposed to the way we represent it? -/- The authors address this question by presenting neural network models that integrate the two most fundamental phenomena of cognition: our ability to learn from experience, and our ability to reason from what has been (...) learned. This book is the first to offer a self-contained presentation of neural network models for a number of computer science logics, including modal, temporal, and epistemic logics. By using a graphical presentation, it explains neural networks through a sound neural-symbolic integration methodology, and it focuses on the benefits of integrating effective robust learning with expressive reasoning capabilities. -/- The book will be invaluable reading for academic researchers, graduate students, and senior undergraduates in computer science, artificial intelligence, machine learning, cognitive science and engineering. It will also be of interest to computational logicians, and professional specialists on applications of cognitive, hybrid and artificial intelligence systems. (shrink)

Inductive reasoning is a fundamental and complex aspect of human intelligence. In particular, how do subjects, given a set of particular examples, generate general descriptions of the rules governing that set? We present a biologically plausible method for accomplishing this task and implement it in a spiking neuron model. We demonstrate the success of this model by applying it to the problem domain of Raven's Progressive Matrices, a widely used tool in the field of intelligence testing. The model is (...) able to generate the rules necessary to correctly solve Raven's items, as well as recreate many of the experimental effects observed in human subjects. (shrink)

The ability to conduct logical reasoning is a fundamental aspect of intelligent behavior, and thus an important problem along the way to human-level artificial intelligence. Traditionally, symbolic methods from the field of knowledge representation and reasoning have been used to equip agents with capabilities that resemble human reasoning qualities. More recently, however, there has been an increasing interest in applying alternative approaches based on machine learning rather than logic-based formalisms to tackle this kind of tasks. Here, we (...) make use of state-of-the-art methods for training deep neural networks to devise a novel model that is closely coupled to symbolic reasoning methods, and thus able to learn how to effectively perform basic ontology reasoning. This term describes an important and at the same time very natural kind of problem settings where the rules for conducting reasoning are specified alongside with the actual information. Many problems in practice may be viewed as such reasoning tasks, which is why the presented approach is applicable to a plethora of important real-world problems. To demonstrate the effectiveness of the suggested method, we present the outcomes of several experiments that have been conducted on both toy datasets as well as real-world data, which show that our model learned to perform precise reasoning on a number of diverse inference tasks that require comprehensive deductive proficiencies. Furthermore, it turned out that the suggested model suffers much less from different obstacles that prohibit symbolic reasoning. (shrink)

Abduction is or subsumes a process of inference. It entertains possible hypotheses and it chooses hypotheses for further scrutiny. There is a large literature on various aspects of non-symbolic, subconscious abduction. There is also a very active research community working on the symbolic (logical) characterisation of abduction, which typically treats it as a form of hypothetico-deductive reasoning. In this paper we start to bridge the gap between the symbolic and sub-symbolic approaches to abduction. We are interested in benefiting from (...) developments made by each community. In particular, we are interested in the ability of non-symbolic systems (neural networks) to learn from experience using efficient algorithms and to perform massively parallel computations of alternative abductive explanations. At the same time, we would like to benefit from the rigour and semantic clarity of symbolic logic. We present two approaches to dealing with abduction in neural networks. One of them uses Connectionist Modal Logic and a translation of Horn clauses into modal clauses to come up with a neural network ensemble that computes abductive explanations in a top-down fashion. The other combines neural-symbolic systems and abductive logic programming and proposes a neural architecture which performs a more systematic, bottom-up computation of alternative abductive explanations. Both approaches employ standard neural network architectures which are already known to be highly effective in practical learning applications. Differently from previous work in the area, our aim is to promote the integration of reasoning and learning in a way that the neural network provides the machinery for cognitive computation, inductive learning and hypothetical reasoning, while logic provides the rigour and explanation capability to the systems, facilitating the interaction with the outside world. Although it is left as future work to determine whether the structure of one of the proposed approaches is more amenable to learning than the other, we hope to have contributed to the development of the area by approaching it from the perspective of symbolic and sub-symbolic integration. (shrink)

The ability to conduct logical reasoning is a fundamental aspect of intelligent behavior, and thus an important problem along the way to human-level artificial intelligence. Traditionally, symbolic methods from the field of knowledge representation and reasoning have been used to equip agents with capabilities that resemble human reasoning qualities. More recently, however, there has been an increasing interest in applying alternative approaches based on machine learning rather than logic-based formalisms to tackle this kind of tasks. Here, we (...) make use of state-of-the-art methods for training deep neural networks to devise a novel model that is closely coupled to symbolic reasoning methods, and thus able to learn how to effectively perform basic ontology reasoning. This term describes an important and at the same time very natural kind of problem settings where the rules for conducting reasoning are specified alongside with the actual information. Many problems in practice may be viewed as such reasoning tasks, which is why the presented approach is applicable to a plethora of important real-world problems. To demonstrate the effectiveness of the suggested method, we present the outcomes of several experiments that have been conducted on both toy datasets as well as real-world data, which show that our model learned to perform precise reasoning on a number of diverse inference tasks that require comprehensive deductive proficiencies. Furthermore, it turned out that the suggested model suffers much less from different obstacles that prohibit symbolic reasoning. (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)

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 recent advance of large language models (LLMs) demonstrates that these large-scale foundation models achieve remarkable capabilities across a wide range of language tasks and domains. The success of the statistical learning approach challenges our understanding of traditional symbolic and logical reasoning. The first part of this paper summarizes several works concerning the progress of monotonicity reasoning through neural networks and deep learning. We demonstrate different methods for solving the monotonicity reasoning task using neural and (...) symbolic approaches and also discuss their advantages and limitations. The second part of this paper focuses on analyzing the capability of large-scale general-purpose language models to reason with monotonicity. (shrink)

This book explores why, regarding practical reasoning, humans are sometimes still faster than artificial intelligence systems. It is the first to offer a self-contained presentation of neural network models for many computer science logics.

Abduction is or subsumes a process of inference. It entertains possible hypotheses and it chooses hypotheses for further scrutiny. There is a large literature on various aspects of non-symbolic, subconscious abduction. There is also a very active research community working on the symbolic (logical) characterisation of abduction, which typically treats it as a form of hypothetico-deductive reasoning. In this paper we start to bridge the gap between the symbolic and sub-symbolic approaches to abduction. We are interested in benefiting from (...) developments made by each community. In particular, we are interested in the ability of non-symbolic systems (neural networks) to learn from experience using efficient algorithms and to perform massively parallel computations of alternative abductive explanations. At the same time, we would like to benefit from the rigour and semantic clarity of symbolic logic. We present two approaches to dealing with abduction in neural networks. One of them uses Connectionist Modal Logic and a translation of Horn clauses into modal clauses to come up with a neural network ensemble that computes abductive explanations in a top-down fashion. The other combines neural-symbolic systems and abductive logic programming and proposes a neural architecture which performs a more systematic, bottom-up computation of alternative abductive explanations. Both approaches employ standard neural network architectures which are already known to be highly effective in practical learning applications. Differently from previous work in the area, our aim is to promote the integration of reasoning and learning in a way that the neural network provides the machinery for cognitive computation, inductive learning and hypothetical reasoning, while logic provides the rigour and explanation capability to the systems, facilitating the interaction with the outside world. Although it is left as future work to determine whether the structure of one of the proposed approaches is more amenable to learning than the other, we hope to have contributed to the development of the area by approaching it from the perspective of symbolic and sub-symbolic integration. (shrink)

Here we argue how Hoerl & McCormack's dual system proposal may change the current view about the neural correlates underlying temporal information processing. We also consider that the concept of the dual system may help characterize various timing disabilities in neuropsychiatric disorders from the new perspective.

Few automated legal reasoning systems have been developed in domains of law in which a judicial decision maker has extensive discretion in the exercise of his or her powers. Discretionary domains challenge existing artificial intelligence paradigms because models of judicial reasoning are difficult, if not impossible to specify. We argue that judicial discretion adds to the characterisation of law as open textured in a way which has not been addressed by artificial intelligence and law researchers in depth. We (...) demonstrate that systems for reasoning with this form of open texture can be built by integrating rule sets with neural networks trained with data collected from standard past cases. The obstacles to this approach include difficulties in generating explanations once conclusions have been inferred, difficulties associated with the collection of sufficient data from past cases and difficulties associated with integrating two vastly different paradigms. A knowledge representation scheme based on the structure of arguments proposed by Toulmin has been used to overcome these obstacles. The system, known as Split Up, predicts judicial decisions in property proceedings within family law in Australia. Predictions from the system have been compared to those from a group of lawyers with favourable results. (shrink)

Analogy making from examples is a central task in intelligent system behavior. A lot of real world problems involve analogy making and generalization. Research investigates these questions by building computer models of human thinking concepts. These concepts can be divided into high level approaches as used in cognitive science and low level models as used in neural networks. Applications range over the spectrum of recognition, categorization and analogy reasoning. A major part of legal reasoning could be formally (...) interpreted as an analogy making process. Because it is not the same as reasoning in mathematics or the physical sciences, it is necessary to use a method, which incorporates first the ability to specify likelihood and second the opportunity of including known court decisions. We use for modelling the analogy making process in legal reasoning neural networks and fuzzy systems. In the first part of the paper a neural network is described to identify precedents of immaterial damages. The second application presents a fuzzy system for determining the required waiting period after traffic accidents. Both examples demonstrate how to model reasoning in legal applications analogous to recent decisions: first, by learning a system with court decisions, and second, by analyzing, modelling and testing the decision making with a fuzzy system. (shrink)

The centerpiece of the scientific study of consciousness is the search for the neural correlates of consciousness. Yet science is typically interested not only in discovering correlations, but also – and more deeply – in explaining them. When faced with a correlation between two phenomena in nature, we typically want to know why they correlate. The purpose of this chapter is twofold. The first half attempts to lay out the various possible explanations of the correlation between consciousness and its (...) neural correlate – to provide a sort of “menu” of options from which we probably would ultimately have to choose. The second half raises considerations suggesting that, under certain reasonable assumptions, the choice among these various options may be in principle underdetermined by the relevant scientific evidence. (shrink)

As the demand for explainable deep learning grows in the evaluation of language technologies, the value of a principled grounding for those explanations grows as well. Here we study the state-of-the-art in explanation for neural models for natural-language processing (NLP) tasks from the viewpoint of philosophy of science. We focus on recent evaluation work that finds brittleness in explanations obtained through attention mechanisms.We harness philosophical accounts of explanation to suggest broader conclusions from these studies. From this analysis, we assert (...) the impossibility of causal explanations from attention layers over text data. We then introduce NLP researchers to contemporary philosophy of science theories that allow robust yet non-causal reasoning in explanation, giving computer scientists a vocabulary for future research. (shrink)

Colombo (Phenomenology and the Cognitive Sciences, 2012) argues that we have compelling reasons to posit neural representations because doing so yields unique explanatory purchase in central cases of social norm compliance. We aim to show that there is no positive substance to Colombo’s plea—nothing that ought to move us to endorse representationalism in this domain, on any level. We point out that exposing the vices of the phenomenological arguments against representationalism does not, on its own, advance the case for (...) representationalism one inch—beyond establishing its mere possibility. We criticize the continual confounding of constitutive and explanatory claims and the lack of recognition of a Hard Problem of having to provide a naturalistic account of content, coupled with an inability to face up to it. We point at the inadequacy of various deflationary moves that end up driving representationalists towards the idea of neural representations with non-standard contents or without content altogether, both of which either render neural representationalism unfit for purpose or vacuous. Referring to possibilities for neural manipulation and control, or to established scientific practice does not help representationalism either. (shrink)

The historical debate on representation in cognitive science and neuroscience construes representations as theoretical posits and discusses the degree to which we have reason to posit them. We reject the premise of that debate. We argue that experimental neuroscientists routinely observe and manipulate neural representations in their laboratory. Therefore, neural representations are as real as neurons, action potentials, or any other well-established entities in our ontology.

Can findings from psychology and cognitive neuroscience about the neural mechanisms involved in decision-making can tell us anything useful about the commonly-understood mental phenomenon of making voluntary choices? Two philosophical objections are considered. First, that the neural data is subpersonal, and so cannot enter into illuminating explanations of personal level phenomena like voluntary action. Secondly, that mental properties are multiply realized in the brain in such a way as to make them insusceptible to neuroscientific study. The paper argues (...) that both objections would be weakened by the discovery of empirical generalisations connecting subpersonal properties with the personal level. It gives three case studies that furnish evidence to that effect. It argues that the existence of such interrelations are consistent with a plausible construal of the personal-subpersonal distinction. Furthermore, there is no reason to suppose that the notion subpersonal representation relied on in cognitive neuroscience illicitly imports personal-level phenomena like consciousness or normativity, or is otherwise explanatorily problematic. (shrink)

The target article criticises neural-symbolic systems as inadequate for analogical reasoning and proposes a model of analogy as transformation (i.e., learning). We accept the importance of learning, but we argue that, instead of conflicting, integrated reasoning and learning would model analogy much more adequately. In this new perspective, modern neural-symbolic systems become the natural candidates for modelling analogy.

Formal models of argumentation have been investigated in several areas, from multi-agent systems and artificial intelligence (AI) to decision making, philosophy and law. In artificial intelligence, logic-based models have been the standard for the representation of argumentative reasoning. More recently, the standard logic-based models have been shown equivalent to standard connectionist models. This has created a new line of research where (i) neural networks can be used as a parallel computational model for argumentation and (ii) neural networks (...) can be used to combine argumentation, quantitative reasoning and statistical learning. At the same time, non-standard logic models of argumentation started to emerge. In this paper, we propose a connectionist cognitive model of argumentation that accounts for both standard and non-standard forms of argumentation. The model is shown to be an adequate framework for dealing with standard and non-standard argumentation, including joint-attacks, argument support, ordered attacks, disjunctive attacks, meta-level attacks, self-defeating attacks, argument accrual and uncertainty. We show that the neural cognitive approach offers an adequate way of modelling all of these different aspects of argumentation. We have applied the framework to the modelling of a public prosecution charging decision as part of a real legal decision making case study containing many of the above aspects of argumentation. The results show that the model can be a useful tool in the analysis of legal decision making, including the analysis of what-if questions and the analysis of alternative conclusions. The approach opens up two new perspectives in the short-term: the use of neural networks for computing prevailing arguments efficiently through the propagation in parallel of neuronal activations, and the use of the same networks to evolve the structure of the argumentation network through learning (e.g. to learn the strength of arguments from data). (shrink)

The conclusions drawn by Benjamin Libet from his work with collegues on the timing of somatosensorial conscious experiences has met with a lot of praise and criticism. In this issue we find three examples of the latter. Here I attempt to place the divide between the two opponent camps in a broader perspective by analyzing the question of the relation between physical timing, neural timing, and experiential timing. The nervous system does a sophisticated job of recombining and recoding messages (...) from the sensorial surfaces and if these processes are slighted in a theory, it might become necessary to postulate weird operations, including subjective back-referral. Neuroscientifically inspired theories are of necessity still based on guesses, extrapolations, and philosophically dubious manners of speech. They often assume some neural correlate of consciousness as a part of the nervous system that transforms neural activity in reportable experiences. The majority of neuroscientists appear to assume that the NCC can compare and bind activity patterns only if they arrive simultaneously at the NCC. This leads to a search for synchrony or to theories in terms of the compensation of differences in neural delays . This is the main dimension of the Libet discussion. Examples from vision research, such as “temporal-binding-by-synchrony” and the “flash-lag” effect, are then used to illustrate these reasoning patterns in more detail. Alternatively one could assume symbolic representations of time and space that are not coded in their own dimension . Unless such tags are multiplexed with the quality message , one gets a binding problem for tags. One of the hidden aspects of the discussion between Libet and opponents appears to be the following. Is the NCC smarter than the rest of the nervous system, so that it can solve the problems of local sign and timing on its own, or are these pieces of information coded symbolically early on in the system? A supersmart NCC appears to be the assumption of Libet's camp . The wish to distribute the smartness evenly across all stages of processing in the nervous system appears to motivate the opponents. I argue that there are reasons to side with the latter group. (shrink)

Chemosensation and mechanosensation cover an enormous spectrum of processes by which animals use information from the environment to adapt their behavior. For pragmatic reasons, these sensory modalities are commonly investigated independently. Recent advances, however, have revealed numerous situations in which they function together to control animals’ actions. Highlighting examples from diverse vertebrates and invertebrates, we first discuss sensory receptors and neurons that have dual roles in the detection of chemical and mechanical stimuli. Next we present cases where peripheral chemosensory and (...) mechanosensory pathways are discrete but intimately packaged to permit coordinated reception of external cues. Finally, we consider how chemical and mechanical signals converge in central neural circuitry to enable multisensory integration. These insights demonstrate how investigation of the interplay between different sensory modalities is key to a more holistic and realistic understanding of sensory‐guided behaviors. (shrink)

Introduction of deep neural networks to the machine translation research ameliorated conventional machine translation systems in multiple ways, specifically in terms of translation quality. The ability of deep neural networks to learn a sensible representation of words is one of the major reasons for this improvement. Despite machine translation using deep neural architecture is showing state-of-the-art results in translating European languages, we cannot directly apply these algorithms in Indian languages mainly because of two reasons: unavailability of the (...) good corpus and Indian languages are morphologically rich. In this paper, we propose a neural machine translation (NMT) system for four language pairs: English–Malayalam, English–Hindi, English–Tamil, and English–Punjabi. We also collected sentences from different sources and cleaned them to make four parallel corpora for each of the language pairs, and then used them to model the translation system. The encoder network in the NMT architecture was designed with long short-term memory (LSTM) networks and bi-directional recurrent neural networks (Bi-RNN). Evaluation of the obtained models was performed both automatically and manually. For automatic evaluation, the bilingual evaluation understudy (BLEU) score was used, and for manual evaluation, three metrics such as adequacy, fluency, and overall ranking were used. Analysis of the results showed the presence of lengthy sentences in English–Malayalam, and the English–Hindi corpus affected the translation. Attention mechanism was employed with a view to addressing the problem of translating lengthy sentences (sentences contain more than 50 words), and the system was able to perceive long-term contexts in the sentences. (shrink)

If artificial neural networks are ever to form the foundation for higher level cognitive behaviors in machines or to realize their full potential as explanatory devices for human cognition, they must show signs of autonomy, multifunction operation, and intersystem integration that are absent in most existing models. This model begins to address these issues by integrating predictive learning, sequence interleaving, and sequence creation components to simulate a spectrum of higher-order cognitive behaviors which have eluded the grasp of simpler systems. (...) Its capabilities are described based on simulations calling for increasing levels of functionality and are used to show how the model can progress from fundamental sequence learning and recall tasks to sophisticated behaviors such as an ability to solve simple mathematical expressions and a creative capacity for the formation and application of inductive rules. (shrink)

Problems are raised with the global workspace hypothesis of consciousness, for example about exactly how global the workspace needs to be for consciousness to suddenly be present. Problems are also raised with Carruthers's version that excludes conceptual representations, and in which phenomenal consciousness can be reduced to physical processes, with instead a different levels of explanation approach to the relation between the brain and the mind advocated. A different theory of phenomenal consciousness is described, in which there is a particular (...) computational system involved in which Higher Order Syntactic Thoughts are used to perform credit assignment on first order thoughts of multiple step plans to correct them by manipulating symbols in a syntactic type of working memory. This provides a good evolutionary reason for the evolution of this kind of computational module, with which, it is proposed, phenomenal consciousness is associated. Some advantages of this HOST approach to phenomenal consciousness are then described with reference not only to the global workspace approach, but also to Higher Order Thought theories. It is hypothesized that the HOST system which requires the ability to manipulate first order symbols in working memory might utilize parts of the prefrontal cortex implicated in working memory, and especially the left inferior frontal gyrus, which is involved in language and probably syntactical processing. Overall, the approach advocated is to identify the computations that are linked to consciousness, and to analyze the neural bases of those computations. [copyright: Copyright © 2020 Rolls.]. (shrink)

Proponents of neuroeconomics often argue that better knowledge of the human neural architecture enables economists to improve standard models of choice. In their view, these improvements provide compelling reasons to use neural findings in constructing and evaluating economic models. In a recent article, I criticized this view by pointing to the trade-offs between the modeling desiderata valued by neuroeconomists and other economists, respectively. The present article complements my earlier critique by focusing on three modeling desiderata that figure prominently (...) in economic and neuroeconomic modeling. For each desideratum, I examine findings that neuroeconomists deem to be especially relevant for economists and argue that neuroeconomists have failed to substantiate their calls to use these findings in constructing and evaluating economic models. In doing so, I identify methodological and evidential constraints that will continue to hinder neuroeconomists’ attempts to improve such models. Moreover, I draw on the literature on scientific modeling to advance the ongoing philosophical discussion regarding the prospects of interdisciplinary models of choice. (shrink)

Computational approaches to the law have frequently been characterized as being formalistic implementations of the syllogistic model of legal cognition: using insufficient or contradictory data, making analogies, learning through examples and experiences, applying vague and imprecise standards. We argue that, on the contrary, studies on neural networks and fuzzy reasoning show how AI & law research can go beyond syllogism, and, in doing that, can provide substantial contributions to the law.

Moral judgment depends critically on theory of mind, reasoning about mental states such as beliefs and intentions. People assign blame for failed attempts to harm and offer forgiveness in the case of accidents. Here we use fMRI to investigate the role of ToM in moral judgment of harmful vs. helpful actions. Is ToM deployed differently for judgments of blame vs. praise? Participants evaluated agents who produced a harmful, helpful, or neutral outcome, based on a harmful, helpful, or neutral intention; (...) participants made blame and praise judgments. In the right temporo-parietal junction, and, to a lesser extent, the left TPJ and medial prefrontal cortex, the neural response reflected an interaction between belief and outcome factors, for both blame and praise judgments: The response in these regions was highest when participants delivered a negative moral judgment, i.e., assigned blame or withheld praise, based solely on the agent's intent. These results show enhanced attention to mental states for negative moral verdicts based exclusively on mental state information. (shrink)

Cognitive Science, Volume 46, Issue 3, March 2022.

Decision-making is an intricate subject in neuroscience. It is often argued that laboratorial research is not capable of dealing with the necessary complexity to study the issue. Whereas philosophers in general neglect the physiological features that constitute the main aspects of thought and behaviour, I advocate that cutting-edge neuroscientific experiments can offer us a framework to explain human behaviour in its relationship with will, self-control, inhibition, emotion and reasoning. It is my contention that self-control mechanisms can modulate more basic (...) stimuli. Assuming the aforementioned standpoints, I show the physiological mechanisms underlying social assessment and decision-making. I also establish a difference between veridical and adaptive decision-making, useful to create experimental designs that can better mimic the complexity of our day-by-day decisions in more ecologically relevant laboratorial research. Moreover, I analyse some experiments in order to develop an epistemological reflection about the necessary neural mechanisms to social assessment and decision-making. (shrink)

Decision-making is an intricate subject in neuroscience. It is often argued that laboratorial research is not capable of dealing with the necessary complexity to study the issue. Whereas philosophers in general neglect the physiological features that constitute the main aspects of thought and behaviour, I advocate that cutting-edge neuroscientific experiments can offer us a framework to explain human behaviour in its relationship with will, self-control, inhibition, emotion and reasoning. It is my contention that self-control mechanisms can modulate more basic (...) stimuli. Assuming the aforementioned standpoints, I show the physiological mechanisms underlying social assessment and decision-making. I also establish a difference between veridical and adaptive decision-making, useful to create experimental designs that can better mimic the complexity of our day-by-day decisions in more ecologically relevant laboratorial research. Moreover, I analyse some experiments in order to develop an epistemological reflection about the necessary neural mechanisms to social assessment and decision-making. (shrink)

We critically review themushrooming literature addressing the neuralmechanisms of moral cognition (NMMC), reachingthe following broad conclusions: (1) researchmainly focuses on three inter-relatedcategories: the moral emotions, moral socialcognition, and abstract moral reasoning. (2)Research varies in terms of whether it deploysecologically valid or experimentallysimplified conceptions of moral cognition. Themore ecologically valid the experimentalregime, the broader the brain areas involved.(3) Much of the research depends on simplifyingassumptions about the domain of moral reasoningthat are motivated by the need to makeexperimental progress. This is (...) a valuablebeginning, but as more is understood about theneural mechanisms of decision-making, morerealistic conceptions will need to replace thesimplified conceptions. (4) The neuralcorrelates of real-life moral cognition areunlikely to consist in anything remotely like a``moral module'' or a ``morality center.'' Moralrepresentations, deliberations and decisionsare probably highly distributed and notconfined to any particular brainsub-system. Discovering the basic neuralprinciples governing planning, judgment anddecision-making will require vastly more basicresearch in neuroscience, but correlatingactivity in certain brain regions withwell-defined psychological conditions helpsguide neural level research. Progress on socialphenomena will also require theoreticalinnovation in understanding the brain'sdistinctly biological form of computationthat is anchored by emotions, needs, drives,and the instinct for survival. (shrink)

Most consciousness researchers, almost no matter what their views of the metaphysics of consciousness, can agree that the first step in a science of consciousness is the search for the neural correlate of consciousness (the NCC). The reason for this agreement is that the notion of ‘correlation’ doesn’t by itself commit one to any particular metaphysical view about the relation between (neural) matter and consciousness. For example, some might treat the correlates as causally related, while others might view (...) the correlation as evidence for identity between conscious states and brain states. The common ground therefore seems to be that the scientific search for the NCC is largely independent of the metaphysics of consciousness. (shrink)

This chapter explains how artificial neural networks may be used as models for reasoning, conditionals, and conditional logic. It starts with the historical overlap between neural network research and logic, it discusses connectionism as a paradigm in cognitive science that opposes the traditional paradigm of symbolic computationalism, it mentions some recent accounts of how logic and neural networks may be combined, and it ends with a couple of open questions concerning the future of this area of (...) research. (shrink)

In daily life, perceivers often need to predict and interpret the behavior of group agents, such as corporations and governments. Although research has investigated how perceivers reason about individual members of particular groups, less is known about how perceivers reason about group agents themselves. The present studies investigate how perceivers understand group agents by investigating the extent to which understanding the ‘mind’ of the group as a whole shares important properties and processes with understanding the minds of individuals. Experiment 1 (...) demonstrates that perceivers are sometimes willing to attribute a mental state to a group as a whole even when they are not willing to attribute that mental state to any of the individual members of the group, suggesting that perceivers can reason about the beliefs and desires of group agents over and above those of their individual members. Experiment 2 demonstrates that the degree of activation in brain regions associated with attributing mental states to individuals—i.e., brain regions associated with mentalizing or theory-of-mind, including the medial prefrontal cortex (MPFC), temporo-parietal junction (TPJ), and precuneus—does not distinguish individual from group targets, either when reading statements about those targets' mental states (directed) or when attributing mental states implicitly in order to predict their behavior (spontaneous). Together, these results help to illuminate the processes that support understanding group agents themselves. (shrink)

Although some cognitive studies provided reasons that children with low socioeconomic status (SES) showed poor mathematical achievements, there was no explicit evidence to directly explain the root of lagged performance in children with low SES. Therefore, the present study explored the differences in neural correlates in the process of symbolic magnitude comparison between children with different SES by the event-related potentials (ERP). A total of 16 second graders from low SES families and 16 from middle/high SES families participated in (...) this study. According to the results of anterior N1 (early attention) and P2 (extraction of numerical meaning) over frontal region, the differences among children with different SES were manifested as differences in general neural activities in terms of attention and top-down cognitive control. In the late stage of cognitive processing, there was no significant difference in the average amplitude of the LPC between the children with different SES, indicating that low SES did not influence the information encoding and memory updating of numerical representation responsible by the parietal lobe. The educational implications of this study were mentioned in the discussion. (shrink)

Only problems that cannot be solved without representational changes can be regarded as pure insight problems; others are classified as pseudo-insight problems. Existing studies using neuroimaging...

Brains, unlike artificial neural nets, use symbols to summarise and reason about perceptual input. But unlike symbolic AI, they “ground” the symbols in the data: the symbols have meaning in terms of data, not just meaning imposed by the outside user. If neural nets could be made to grow their own symbols in the way that brains do, there would be a good prospect of combining neural networks and symbolic AI, in such a way as to combine (...) the good features of each. The article argues the cluster analysis provides algorithms to perform the task, and that any solution to the task must be a form of cluster analysis. (shrink)

There are two concepts of consciousness that are easy to confuse with one another, access-consciousness and phenomenal consciousness. However, just as the concepts of water and H 2 O are different concepts of the same thing, so the two concepts of consciousness may come to the same thing in the brain. The focus of this paper is on the problems that arise when these two concepts of consciousness are conflated. I will argue that John Searle's reasoning about the function (...) of consciousness goes wrong because he conflates the two senses. And Francis Crick and Christof Koch fall afoul of the ambiguity in arguing that visual area V1 is not part of the neural correlate of consciousness. Crick and Koch's work raises issues that suggest that these two concepts of consciousness may have different neural correlates – despite Crick and Koch's implicit rejection of this idea. (shrink)

We use the phrase "moral intuition" to describe the appearance in consciousness of moral judgments or assessments without any awareness of having gone through a conscious reasoning process that produces this assessment. This paper investigates the neural substrates of moral intuition. We propose that moral intuitions are part of a larger set of social intuitions that guide us through complex, highly uncertain and rapidly changing social interactions. Such intuitions are shaped by learning. The neural substrates for moral (...) intuition include fronto-insular, cingulate, and orbito-frontal cortices and associated subcortical structure such as the septum, basil ganglia and amygdala. Understanding the role of these structures undercuts many philosophical doctrines concerning the status of moral intuitions, but vindicates the claim that they can sometimes play a legitimate role in moral decision-making. (shrink)

In the late summer of 1998, the authors, a cognitive scientist and a logician, started talking about the relevance of modern mathematical logic to the study of human reasoning, and we have been talking ever since. This book is an interim report of that conversation. It argues that results such as those on the Wason selection task, purportedly showing the irrelevance of formal logic to actual human reasoning, have been widely misinterpreted, mainly because the picture of logic current (...) in psychology and cognitive science is completely mistaken. We aim to give the reader a more accurate picture of mathematical logic and, in doing so, hope to show that logic, properly conceived, is still a very helpful tool in cognitive science. The main thrust of the book is therefore constructive. We give a number of examples in which logical theorizing helps in understanding and modeling observed behavior in reasoning tasks, deviations of that behavior in a psychiatric disorder (autism), and even the roots of that behavior in the evolution of the brain. (shrink)