Augustine tells us in the Confessions that his reading of Cicero's Hortensius at the age of nineteen aroused in him a burning ‘passion for the wisdom of eternal truth’. He was inspired ‘to love wisdom itself, whatever it might be, and to search for it, pursue it, hold it, and embrace it firmly’. And thus he embarked on his arduous journey to the truth, which was at the same time a conversion to Catholic Christianity, and which culminated twelve years later (...) in his experience in the garden in Milan. (shrink)

While there is virtually a consensus among contemporary philosophers of perception that some form of direct realism is true, there is less than complete agreement about whether normal, direct perceptions involve mental inferences in any sense. In taking another look at this recurrent question, my aim is twofold: first, to examine some of the arguments and evidences that have been offered in favor of inferences and to see if they can be accommodated within the direct realist framework, and second, to (...) attempt to clarify and defend the insight of direct realism that normal perceptions are noninferential. (shrink)

In this study I am going to present and discuss some of the central themes of Gustav Bergmann's theory of perception. I shall be concerned, however, only with "later Bergmann," that is, with the perceptual theory worked out in a series of essays in which Bergmann shifts from phenomenalism to a form of intentional realism. This label ("intentional realism") indicates the two dominant themes in Bergmann's later thought about perception: perceivings are analyzed as mental acts (thoughts) which are intentionally related (...) to real and mind-independent objects and states of affairs. In a timely essay, " Intentionality" (1955) Bergmann presented an impressive defense of mental acts, and although the framework of that essay was still phenomenalistic, the structural analysis of mental acts and their intentionality presented there has not been significantly altered by Bergmann in later writings. In two subsequent essays, "Acts" (1960) and "Realistic Postscript" (1963) his realistic turn was worked out in detail in the context of giving a satisfactory account of ordinary perceptual experience. These three essays will comprise the core texts of the study that follows. (shrink)

I interpret and defend Sellars’ intemalist view of perceptual justification which argues that perceivers have evidence for their perceptual beliefs that includes a higher-order belief about the circumstances in which those beliefs arise, and an epistemic belief about the reliability of beliefs that are formed in those circumstances. The pattem of inference that occurs in ordinary cases of perception is elicited.I then defend this account of perceptual evidence against 1) AIston’s objection that ordinary perceivers are not as critical and reflective (...) as this view requires them to be; and 2) the charge that intemalism leads to various forms of infinite regress and circular reasoning. It is granted that subjects must have further grounds for their justifying reasons, and an attempt is made to identify these second-order reasons. In particular, I argue that epistemic beliefs are grounded in the perceiver’s awareness that his present experience-cum-conditions fits into a larger pattem of similar past experiences that were reliably connected with their objects. (shrink)

I interpret and defend Sellars’ intemalist view of perceptual justification which argues that perceivers have evidence for their perceptual beliefs that includes a higher-order belief about the circumstances in which those beliefs arise, and an epistemic belief about the reliability of beliefs that are formed in those circumstances. The pattem of inference that occurs in ordinary cases of perception is elicited.I then defend this account of perceptual evidence against 1) AIston’s objection that ordinary perceivers are not as critical and reflective (...) as this view requires them to be; and 2) the charge that intemalism leads to various forms of infinite regress and circular reasoning. It is granted that subjects must have further grounds for their justifying reasons, and an attempt is made to identify these second-order reasons. In particular, I argue that epistemic beliefs are grounded in the perceiver’s awareness that his present experience-cum-conditions fits into a larger pattem of similar past experiences that were reliably connected with their objects. (shrink)

The general theory of perception proposed by Roderick Chisholm in his book Perceiving: A Philosophical Study1 has gained considerable acceptance among contemporary philosophers of perception. In this paper, I will review and evaluate one part of this theory and show where I believe an important modification is necessary. Chisholm distinguishes what he thinks are two importantly different senses of “perceive,” a propositional and a nonpropositional sense, and then proposes a definition of each. The propositional sense of “perceive” is expressed in (...) contexts in which what is perceived is referred to by a propositional clause, as in 1. George perceives that this is a door. The nonpropositional sense is expressed in contexts in which the perceptual object is referred to by a noun or noun phrase, as in 2. George perceives the door. The problem areas I will deal with are, first and foremost, the adequacy of the definitions Chisholm gives of these two senses, and second, the ways in which these senses are related to one another. In the first section, dealing with the propositional sense, I will not be concerned with propositional perceivings in their entirety. Specifically, I will not be concerned with them as instances of knowing that something is such and such, which is a central feature of Chisholm’s definition. Rather, I will be concerned mainly with the mental state, qua mental, which is involved in these perceptions, that is, the mental state considered apart from any relation it may have to the object it is about. I will assume that the mental state involved in a propositional perceiving is a conceptual state, i.e., one which involves the exercise of concepts or thoughts, together with an accompanying propositional attitude. The key locution Chisholm uses to describe this conceptual mental state is “S takes something to have some property.” I will argue that Chisholm’s definition of taking will not do the job it is intended to do in these perceptual contexts. In the second section, I will explore what Chisholm calls the “simplest” nonpropositional sense of “perceive,” which he defines wholly in terms of a causal relation existing between the object perceived and the subject’s sensory state. Significantly, this definition does not make any reference to a taking or to any propositional state on the part of the subject, but only to the subject’s sensory state. Consequently, if one accepts the idea, as I believe Chisholm would, that sensing is a nonconceptual state, then Chisholm’s nonpropositional sense of “perceive” would not refer to any conceptual fact at all. I will consider Chisholm’s definition, and raise the question of whether nonpropositional perceptual contexts should ever be construed as having a propositional meaning, that is, as including the idea that the subject exhibits some thought. As we shall see, Chisholm makes allowance for such a possibility. (shrink)

This paper develops and defends an internalist account of having authority for one’s claim. It begins with Robert Brandom’s pragmatist account of thinking which locates the root notion of reasoning in a primitive language game of asking for and giving reasons. The idea is that the authority of a claim can be spelled out pragmatically in terms of the social practice of undertaking commitments and attributing entitlements. It is argued that this account fails to acknowledge the role of the subject’s (...) grasp of the higher-order concept of the evidence on which I base my claim. (shrink)

To sum up the main results of this study: I have disentangled two distinct patterns of argument that Taylor runs together in his attempt to show that there is a reason or explanation for the world as a whole. The first is based on the causal dependency of things in the world, the second is based on their logical contingency. It seems to make the most sense of Taylor's discussion if we interpret him not as invoking the principle of sufficient (...) reason at the crucial juncture, but as using these arguments to give backing to that principle by showing that it applies to the world in its totality. However, these arguments do not succeed in doing that. The first fails because it depends on a remote analogy between the world as a whole and the physical objects in the world. Concerning the second, an analysis of the logic of why-questions about the existence of things has revealed that the logical contingency of something is not a ground for thinking it has an explanation. The only promising interpretation of the principle of sufficient reason that we have found is as a causal principle pertaining to things in nature. (shrink)

The question that is front and center at the outset of William James's great work The Varieties of Religious Experience is What is religion?1 He opens his discussion of this question in the second lecture, "Circumscription of the Topic," with the observation that the field of religion is so "wide" and the definitions of it so varied that it would be "foolish" to try to find a single definition that formulates its essence. Even if, instead, we look for a defining (...) characteristic of religion within the self or mind, we find the same disagreement among authors "in the psychologies and in the philosophies of religion": "One man allies it to the feeling of dependence; one makes it a derivative from fear; others connect... (shrink)

When someone believes something that is justified for her, what part does the subject play in her state of being justified? I will answer this question by developing a strong internalist account of justification according to which the justification of a believing for a subject consists in her having grounds for her belief, and holding the belief in recognition of those grounds. But the internalist theory I defend incorporates key elements of reliabilism into its account. Using perception as a model (...) for justification, I show how ordinary perceivers would appeal to external factors to support their perceptual beliefs, and normally suppose that their beliefs are reliably connected to the objects their beliefs are about. I find in this feature of our common justificatory practice a sufficient basis for positing an externalist condition on justification—namely that subjects are only justified if their beliefs are reliably connected to their objects. (shrink)

This paper contests social psychology’s emphasis on the biased, erroneous, and constructed nature of social cognition by: showing how the extent of bias and error in classic research is overstated; summarizing research regarding the accuracy of social beliefs; and describing how social stereotypes sometimes improve person perception accuracy. A Goodness of Judgment Index is also presented to extract evidence regarding accuracy from research focusing on bias. We conclude that accuracy is necessary for understanding social cognition.

We demonstrate that an unlexicalized PCFG can parse much more accurately than previously shown, by making use of simple, linguistically motivated state splits, which break down false independence assumptions latent in a vanilla treebank grammar. Indeed, its performance of 86.36% (LP/LR F1) is better than that of early lexicalized PCFG models, and surprisingly close to the current state-of-theart. This result has potential uses beyond establishing a strong lower bound on the maximum possible accuracy of unlexicalized models: an unlexicalized PCFG is (...) much more compact, easier to replicate, and easier to interpret than more complex lexical models, and the parsing algorithms are simpler, more widely understood, of lower asymptotic complexity, and easier to optimize. (shrink)

This chapter explores the debate between contemporary projectivists or expressivists, and the advocates of sensibility theory. Both positions are best viewed as forms of sentimentalism — the theory that evaluative concepts must be explicated by appeal to the sentiments. It argues that the sophisticated interpretation of such notions as “true” and “objective” that are offered by defenders of these competing views ultimately undermines the significance of their meta-ethical disputes over “cognitivism” and “realism” about value. Their fundamental disagreement lies in moral (...) psychology; it concerns how best to understand the emotions to which sentimentalist theories must appeal. (shrink)

Supporting Hamilton’s inclusive fitness theory, archival analyses of inheritance patterns in wills have revealed that people invest more of their estates in kin of closer genetic relatedness. Recent classroom experiments have shown that this genetic relatedness effect is stronger for relatives of direct lineage (children, grandchildren) than for relatives of collateral lineage (siblings, nieces, nephews). In the present research, multilevel modeling of more than 1,000 British Columbian wills revealed a positive effect of genetic relatedness on proportions of estates allocated to (...) relatives. This effect was qualified by an interaction with lineage, such that it was stronger for direct than for collateral relatives. Exploratory analyses of the moderating role of benefactors’ sex and estate values showed the genetic relatedness effect was stronger among female and wealthier benefactors. The importance of these moderators to understanding kin investment in modern humans is discussed. (shrink)

We present a novel generative model for natural language tree structures in which semantic (lexical dependency) and syntactic (PCFG) structures are scored with separate models. This factorization provides conceptual simplicity, straightforward opportunities for separately improving the component models, and a level of performance comparable to similar, non-factored models. Most importantly, unlike other modern parsing models, the factored model admits an extremely effective A* parsing algorithm, which enables efficient, exact inference.

This paper presents a novel approach to the unsupervised learning of syntactic analyses of natural language text. Most previous work has focused on maximizing likelihood according to generative PCFG models. In contrast, we employ a simpler probabilistic model over trees based directly on constituent identity and linear context, and use an EM-like iterative procedure to induce structure. This method produces much higher quality analyses, giving the best published results on the ATIS dataset.

We present a generative distributional model for the unsupervised induction of natural language syntax which explicitly models constituent yields and contexts. Parameter search with EM produces higher quality analyses than previously exhibited by unsupervised systems, giving the best published unsupervised parsing results on the ATIS corpus. Experiments on Penn treebank sentences of comparable length show an even higher F1 of 71% on nontrivial brackets. We compare distributionally induced and actual part-of-speech tags as input data, and examine extensions to the basic (...) model. We discuss errors made by the system, compare the system to previous models, and discuss upper bounds, lower bounds, and stability for this task. (shrink)

This paper presents empirical studies and closely corresponding theoretical models of the performance of a chart parser exhaustively parsing the Penn Treebank with the Treebank’s own CFG grammar. We show how performance is dramatically affected by rule representation and tree transformations, but little by top-down vs. bottom-up strategies. We discuss grammatical saturation, including analysis of the strongly connected components of the phrasal nonterminals in the Treebank, and model how, as sentence length increases, the effective grammar rule size increases as regions (...) of the grammar are unlocked, yielding super-cubic observed time behavior in some configurations. (shrink)

This paper separates conditional parameter estima- tion, which consistently raises test set accuracy on statistical NLP tasks, from conditional model struc- tures, such as the conditional Markov model used for maximum-entropy tagging, which tend to lower accuracy. Error analysis on part-of-speech tagging shows that the actual tagging errors made by the conditionally structured model derive not only from label bias, but also from other ways in which the independence assumptions of the conditional model structure are unsuited to linguistic sequences. The (...) paper presents new word-sense disambiguation and POS tagging experiments, and integrates apparently conflicting reports from other recent work. (shrink)

In a range of contexts, individuals arrive at collective decisions by sharing confidence in their judgements. This tendency to evaluate the reliability of information by the confidence with which it is expressed has been termed the ‘confidence heuristic’. We tested two ways of implementing the confidence heuristic in the context of a collective perceptual decision-making task: either directly, by opting for the judgement made with higher confidence, or indirectly, by opting for the faster judgement, exploiting an inverse correlation between confidence (...) and reaction time. We found that the success of these heuristics depends on how similar individuals are in terms of the reliability of their judgements and, more importantly, that for dissimilar individuals such heuristics are dramatically inferior to interaction. Interaction allows individuals to alleviate, but not fully resolve, differences in the reliability of their judgements. We discuss the implications of these findings for models of confidence and collective decision-making. (shrink)

While Ç ´Ò¿ µ methods for parsing probabilistic context-free grammars (PCFGs) are well known, a tabular parsing framework for arbitrary PCFGs which allows for botton-up, topdown, and other parsing strategies, has not yet been provided. This paper presents such an algorithm, and shows its correctness and advantages over prior work. The paper finishes by bringing out the connections between the algorithm and work on hypergraphs, which permits us to extend the presented Viterbi (best parse) algorithm to an inside (total probability) (...) algorithm. (shrink)

A* PCFG parsing can dramatically reduce the time required to find the exact Viterbi parse by conservatively estimating outside Viterbi probabilities. We discuss various estimates and give efficient algorithms for computing them. On Penn treebank sentences, our most detailed estimate reduces the total number of edges processed to less than 3% of that required by exhaustive parsing, and even a simpler estimate which can be pre-computed in under a minute still reduces the work by a factor of 5. The algorithm (...) extends the classic A* graph search procedure to a certain hypergraph associated with parsing. Unlike bestfirst and finite-beam methods for achieving this kind of speed-up, the A* parser is guaranteed to return the most likely parse, not just an approximation. The algorithm is also correct for a wide range of parser control strategies and maintains a worst-case cubic time bound. (shrink)

This paper discusses ensembles of simple but heterogeneous classifiers for word-sense disambiguation, examining the Stanford-CS224N system entered in the SENSEVAL-2 English lexical sample task. First-order classifiers are combined by a second-order classifier, which variously uses majority voting, weighted voting, or a maximum entropy model. While individual first-order classifiers perform comparably to middle-scoring teams’ systems, the combination achieves high performance. We discuss trade-offs and empirical performance. Finally, we present an analysis of the combination, examining how ensemble performance depends on error independence (...) and task difficulty. (shrink)

Unsupervised grammar induction systems commonly judge potential constituents on the basis of their effects on the likelihood of the data. Linguistic justifications of constituency, on the other hand, rely on notions such as substitutability and varying external contexts. We describe two systems for distributional grammar induction which operate on such principles, using part-of-speech tags as the contextual features. The advantages and disadvantages of these systems are examined, including precision/recall trade-offs, error analysis, and extensibility.

We present an improved method for clustering in the presence of very limited supervisory information, given as pairwise instance constraints. By allowing instance-level constraints to have spacelevel inductive implications, we are able to successfully incorporate constraints for a wide range of data set types. Our method greatly improves on the previously studied constrained -means algorithm, generally requiring less than half as many constraints to achieve a given accuracy on a range of real-world data, while also being more robust when over-constrained. (...) We additionally discuss an active learning algorithm which increases the value of constraints even further. (shrink)

erative clustering. First, we show formally that the common heuristic agglomerative clustering algorithms – Ward’s method, single-link, complete-link, and a variant of group-average – are each equivalent to a hierarchical model-based method. This interpretation gives a theoretical explanation of the empirical behavior of these algorithms, as well as a principled approach to resolving practical issues, such as number of clusters or the choice of method. Second, we show how a model-based viewpoint can suggest variations on these basic agglomerative algorithms. We (...) introduce adjusted complete-link, Mahalanobis-link, and line-link as variants, and demonstrate their utility. (shrink)

While symbolic parsers can be viewed as deduction systems, this view is less natural for probabilistic parsers. We present a view of parsing as directed hypergraph analysis which naturally covers both symbolic and probabilistic parsing. We illustrate the approach by showing how a dynamic extension of Dijkstra’s algorithm can be used to construct a probabilistic chart parser with an Ç´Ò¿µ time bound for arbitrary PCFGs, while preserving as much of the flexibility of symbolic chart parsers as allowed by the inherent (...) ordering of probabilistic dependencies. (shrink)

Multilevel research strategies characterize contemporary molecular inquiry into biological systems. We outline conceptual, methodological, and explanatory dimensions of these multilevel strategies in microbial ecology, systems biology, protein research, and developmental biology. This review of emerging lines of inquiry in these fields suggests that multilevel research in molecular life sciences has significant implications for philosophical understandings of explanation, modeling, and representation.

Integrating large language models (LLMs) like GPT-4 into medical ethics is a novel concept, and understanding the effectiveness of these models in aiding ethicists with decision-making can have significant implications for the healthcare sector. Thus, the objective of this study was to evaluate the performance of GPT-4 in responding to complex medical ethical vignettes and to gauge its utility and limitations for aiding medical ethicists. Using a mixed-methods, cross-sectional survey approach, a panel of six ethicists assessed LLM-generated responses to eight (...) ethical vignettes. The main outcomes measured were relevance, reasoning, depth, technical and non-technical clarity, as well as acceptability of GPT-4’s responses. The readability of the responses was also assessed. Of the six metrics evaluating the effectiveness of GPT-4’s responses, the overall mean score was 4.1/5. GPT-4 was rated highest in providing technical (4.7/5) and non-technical clarity (4.4/5), whereas the lowest rated metrics were depth (3.8/5) and acceptability (3.8/5). There was poor-to-moderate inter-rater reliability characterised by an intraclass coefficient of 0.54 (95% CI: 0.30 to 0.71). Based on panellist feedback, GPT-4 was able to identify and articulate key ethical issues but struggled to appreciate the nuanced aspects of ethical dilemmas and misapplied certain moral principles. This study reveals limitations in the ability of GPT-4 to appreciate the depth and nuanced acceptability of real-world ethical dilemmas, particularly those that require a thorough understanding of relational complexities and context-specific values. Ongoing evaluation of LLM capabilities within medical ethics remains paramount, and further refinement is needed before it can be used effectively in clinical settings. (shrink)

Metacognition – the ability to represent, monitor and control ongoing cognitive processes – helps us perform many tasks, both when acting alone and when working with others. While metacognition is adaptive, and found in other animals, we should not assume that all human forms of metacognition are gene-based adaptations. Instead, some forms may have a social origin, including the discrimination, interpretation, and broadcasting of metacognitive representations. There is evidence that each of these abilities depends on cultural learning and therefore that (...) cultural selectionmight shape humanmetacognition. The cultural origins hypothesis is a plausible and testable alternative that directs us towards a substantial new programme of research. (shrink)