Epistemic Inference Research Articles

Imprecise Probabilistic Model Checking for Stochastic Multi-agent Systems

Standard techniques for model checking stochastic multi-agent systems usually assume the transition probabilities describing the system dynamics to be stationary and completely specified. As a consequence, neither non-stationary systems nor systems whose stochastic behaviour is partially unknown can be treated. So far, most of the approaches proposed to overcome this limitation suffer from complexity issues making them poorly efficient in the case of large state spaces. A fruitful but poorly explored way out is offered by the formalism of imprecise probabilities and the related imprecise Markov models. The aim of this paper is to show how imprecise probabilities can be fruitfully involved to model-check multi-agent systems characterised by non-stationary behaviours. Specifically, the paper introduces a new class of multi-agent models called Imprecise Probabilistic Interpreted Systems and their relative extensions with rewards. It also introduces a proper logical language to specify properties of such models and corresponding model checking algorithms based on iterative procedures to compute probabilistic and epistemic inferences over imprecise Markov models.

Probabilistic modeling of rational communication with conditionals

While a large body of work has scrutinized the meaning of conditional sentences, considerably less attention has been paid to formal models of their pragmatic use and interpretation. Here, we take a probabilistic approach to pragmatic reasoning about indicative conditionals which flexibly integrates gradient beliefs about richly structured world states. We model listeners' update of their prior beliefs about the causal structure of the world and the joint probabilities of the consequent and antecedent based on assumptions about the speaker's utterance production protocol. We show that, when supplied with natural contextual assumptions, our model uniformly explains a number of inferences attested in the literature, including epistemic inferences, conditional perfection and the dependency between antecedent and consequent of a conditional. We argue that this approach also helps explain three puzzles introduced by Douven (2012) about updating with conditionals: depending on the utterance context, the listener's belief in the antecedent may increase, decrease or remain unchanged.

Children's epistemic inferences through modal verbs and prosody.

This study explores how young children infer nuances in epistemic modality through prosody. A forced-choice task was used, testing children's (ages three to seven) comprehension of the might/will distinction (modal condition) as well their ability to modulate the strength of might through two prosodic tunes (prosody condition). Positive and negative valence conditions were included. Younger children were shown to start off performing above chance for the modal condition, and at around chance for the prosody condition, but after age four performance on the prosody condition quickly improved. For both modal verbs and prosody, children performed significantly better when valence was positive. By age seven, children performed at ceiling for all conditions. Qualitative analysis of children's justifications for prosody responses showed metalinguistic awareness of prosodic meaning as early as age four, with the ability to relate prosody to epistemic modal meaning becoming quite common by age seven.

Attitudes in discourse: Italian polar questions and the particle mica

This paper explores ways in which discourse participants convey an attitude about another discourse participant’s conversational move. We examine the semantics/pragmatics of Italian positive and negative polar questions (building on the literature on biased questions) and propose the first fully compositional analysis of the Italian particle ‘mica’, appearing in negative polar questions and negative assertions. The core is that ‘mica’ is member of a family of presuppositional, epistemic ‘common ground management’ operators, leading to a new account of epistemic inferences in biased polar questions that relies on the presuppositional nature of these operators. We argue that ‘mica’ is a high-left-periphery particle that indicates a presupposed bias against a proposition being added to the common ground, anchored uniformly to the speaker and therefore not showing ‘interrogative flip’. The paper develops connections between common-ground management operators and evidentials, arguing that interrogative flip (and lack thereof) is a phenomenon that should be studied for a wide variety of discourse particles. EARLY ACCESS

Simulation-Based Research in Information Systems

Simulations provide a useful methodological approach for studying the behavior of complex socio-technical information systems (IS), in which humans and IT artifacts interact to process information. However, the use of simulations is relatively new in IS research and the current presence and impact of simulation-based studies is still limited. Furthermore, simulation-based research is quite different from other approaches, making it difficult to position and evaluate it adequately. Therefore, this paper first analyses the epistemic particularities of simulation-based IS research. Based on this analysis, a structured literature review of the status quo of simulation-based IS research was conducted, to understand how IS scholars currently employ simulation. A comparison of the epistemic particularities of simulation-based research with its status quo in IS literature allows to critically examine epistemic inferences in the respective research process. The results provide guidance for prospective simulation-based IS research through discussing the theory-based derivation of simulation models, as well as different simulation techniques, validation techniques, and simulation uses.

A new perspective on the shielding property of positive polarity

In certain languages, disjunctions exhibit positive polarity behavior, which Szabolcsi (2002) argues can be diagnosed via the following four properties: (i) anti-licensing: no narrow scope interpretation under a clausemate negation, (ii) rescuing: acceptable in the scope of an even number of negative operators, (iii) shielding: acceptable under a clausemate negation if a universal quantifier intervenes, and (iv) locality of anti-licensing: acceptable in the scope of an extra-clausal negation. In recent work, Nicolae (2016, 2017), building on Spector 2014, argues that what distinguishes PPI disjunctions from polarity insensitive disjunctions is the fact that PPI-disjunctions obligatorily trigger epistemic inferences. That analysis, however, only accounts for the first two PPI properties. This paper extends that analysis to account for the second two properties, concluding that they should be seen as instantiations of the same phenomena, namely shielding by a universal quantifier.

Scalar Implicature in Absence of Epistemic Reasoning? The Case of Autism Spectrum Disorder

ABSTRACTWe investigated “scalar implicature” in adolescents and children with autism spectrum disorder (ASD) to test whether theory of mind deficits associated with autism affect pragmatic inferences in language. We tested scalar implicature computation in adolescents with ASD (12–18 years) and asked whether they reason about mental states when computing inferences. Like previous studies, we found the adolescents with ASD computed implicatures to the same degree as neurotypical adults. However, we also found that this ability may not rely on epistemic reasoning. In a test of epistemic reasoning (which probed so-called “ignorance implicature”) we found that adolescents with ASD were able to make epistemic inferences required by Gricean models of scalar implicature when they were explicitly required by the task to do so. However, in a second task, which asked whether subjects spontaneously reason about mental states in the service of scalar implicature when not explicitly asked to, we found that adolescents with ASD did not engage in epistemic reasoning, leading them to compute scalar implicatures in contexts in which they were not justified. Based on these data, we argue that epistemic reasoning may not be a core, constitutive component of scalar implicature.

Inducing Knowledgeability From Niceness: Children Use Social Features for Making Epistemic Inferences

In many ways, evaluating informants based on their features is a problem of induction: Children rely on the assumption that observable informant characteristics (e.g., traits, behaviors, social categories) will predict unobservable characteristics (e.g., future behavior, knowledge states, intentions). Yet to make sensible inferences, children must recognize what informant features are relevant for what types of inferences. The current research investigated whether preschoolers use social features (e.g., niceness) for making epistemic inferences and, conversely, whether they use intellectual features (e.g., expertise) for making social inferences. In the study, 96 preschoolers (Mage = 4.96 years) were asked to attribute knowledge and behaviors to a mean informant, a nice informant, and a neutral informant. Between subjects, we varied which informant had expertise. We found that when attributing knowledge, children used both features: attributing more knowledge to nicer informants, but also attributing more knowledge to an informant when he had expertise. In contrast, when making social inferences, children relied primarily on social features.

Informants' Traits Weigh Heavily in Young Children's Trust in Testimony and in Their Epistemic Inferences

Informants' Traits Weigh Heavily in Young Children's Trust in Testimony and in Their Epistemic Inferences

Informants' Traits Weigh Heavily in Young Children's Trust in Testimony and in Their Epistemic Inferences

This study examined how informants' traits affect how children seek information, trust testimony, and make inferences about informants' knowledge. Eighty-one 3- to 6-year-olds and 26 adults completed tasks where they requested and endorsed information provided by one of two informants with conflicting traits (e.g., honesty vs. dishonesty). Participants also completed tasks where they simultaneously considered informants' traits and visual access to information when inferring their knowledge and trusting their testimony. Children and adults preferred to ask and endorse information provided by people who are nice, smart, and honest. Moreover, these traits influenced the knowledge that young children attributed to informants. Children younger than 5 years of age reported that people with positive traits were knowledgeable even when they lacked access to relevant information.

Making claims and counterclaims through factuality: The uses of Mandarin Chinese qishi (‘actually’) and shishishang (‘in fact’) in institutional settings

The study reported here, building on the research methods of Conversation Analysis (Sacks et al., 1974), Politeness Theory (Brown and Levinson, 1987), and Relevance Theory (Sperber and Wilson, 1986/1995), attempts to examine the distribution of Mandarin qishi (‘actually’) and shishishang (‘in fact’) across two different discourse modes in formal speech settings: formal lectures and TV panel news discussions. The results indicate that qishi is prevalent in TV panel news discussion data, which fall into the interactional mode, whereas shishishang is more prevalent in formal speech data, which fall into the transactional mode. The study shows that in interaction, qishi is addressee-oriented and signals alignment (agreement) or divergence (disagreement), whereas shishishang is message-oriented and asserts a proposition with a tone of certainty. In addition, the study suggests that although the literal meanings of qishi (‘it’s fact’) and shishishang (‘in the aspect of fact’), which are concerned with factuality, are seemingly unrelated to emotive expressivity, they offer a rhetorical strategy for expressing the speaker’s attitudinal position, and can both serve to indicate the speaker’s epistemic inference.

INDUCTION OF CLASSIFIERS THROUGH NON-PARAMETRIC METHODS FOR APPROXIMATE CLASSIFICATION AND RETRIEVAL WITH ONTOLOGIES

This work concerns non-parametric approaches for statistical learning applied to the standard knowledge representation languages adopted in the Semantic Web context. We present methods based on epistemic inference that are able to elicit and exploit the semantic similarity of individuals in OWL knowledge bases. Specifically, a totally semantic and language-independent semi-distance function is introduced, whence also an epistemic kernel function for Semantic Web representations is derived. Both the measure and the kernel function are embedded in non-parametric statistical learning algorithms customized for coping with Semantic Web representations. Particularly, the measure is embedded in a k-Nearest Neighbor algorithm and the kernel function is embedded in a Support Vector Machine. The implemented algorithms are used to perform inductive concept retrieval and query answering. An experimentation on real ontologies proves that the methods can be effectively employed for performing the target tasks, and moreover that it is possible to induce new assertions that are not logically derivable.

What do uncertainty-averse decision-makers believe?

This paper introduces the concept of firm belief, which is proposed as a new epistemic model for a wide class of preferences. In particular, firm beliefs are shown to have the following desirable properties: (i) they are derived from preferences according to a plausible rule of epistemic inference; (ii) they satisfy standard logical properties; and (iii) tractable representations of firm belief are available for all (suitably continuous) biseparable preferences [13, 14], including the Choquet expected utility [30] and maxmin expected utility [16] classes. We also use firm belief to construct a generalization of Nash equilibrium for (two-player) normal form games.

Epistemic Inference and Illative Judgment

Summary In considering the main weaknesses of epistemologies like Chisholm's, this paper introduces a concept which is foreign to such systems, Cardinal Newman's notion of illative judgment. The method of concrete inference involves the cumulation of probabilities, and principles of epistemic logic can never accommodate all relevant circumstances. Epistemic principles like Chisholm's, perhaps best rephrased as principles of inductive logic, will not, when applied, elicit assent by themselves. Systems like Chisholm's do not ring true phenomenologically because they overemphasize inference at the expense of judgment. Resume Pour remedier aux faiblesses majeures des epistemologies telles que celle de Chisholm, cet article introduit un conceptetranger a de tels systemes: la notion de‘ jugment illatif’ du Cardinal Newman. La methode ?on;inference concrete suppose la cumulation de probabilityes et les principes ?on;une logique epistemique ne peuvent jamais tenir compte de toutes les circonstances qui inter-viennent. Des principes epistemiques tels que ceux de Chisholm, qu'on pourrait peut-etre qualifier de principes de logique inductive, ne vont pas, lorsqu'ils sont appliques, entrainer par eux-memes ľassentiment. De tels systemes ne sonnent pas phenomenologiquement juste parce qu'ils surestiment ľinference aux depens du jugement. Zusammenfassung Indem er die Hauptschwachen von Erkenntnistheorien wie die von Chisholm hervorhebt, fuhrt der Verfasser einen Begriff ein, der solchen Systemen fremd ist, namlich der Begriff des illativen Urteils von Kardinal Newman. Die Methode des konkreten Folgerns involviert die Kumulation von Warscheinlichkeiten, und die Prinzipien der epistemischen Logik konnen nie alle relevanten Umstande erfassen. Epistemische Prinzipien wie diejenigen von Chisholm, die man vielleicht am besten als Prinzipien der induktiven Logik reformuliert, werden, wenn sie angewen-det werden, nicht von selbst Zustimmung erwecken. Systeme wie das von Chisholm scheinen phanomenologisch inadaquat, weil sie Folgerungen auf Kosten von Urteilen uberbewerten.