From Bayesian epistemology to inductive logic

Inductive logic admits a variety of semantics (Haenni et al. (2011) [7, Part 1]). This paper develops semantics based on the norms of Bayesian epistemology (Williamson, 2010 [16, Chapter 7]). Section 1 introduces the semantics and then, in Section 2, the paper explores methods for drawing inferences in the resulting logic and compares the methods of this paper with the methods of Barnett and Paris (2008) [2]. Section 3 then evaluates this Bayesian inductive logic in the light of four traditional critiques of inductive logic, arguing (i) that it is language independent in a key sense, (ii) that it admits connections with the Principle of Indifference but these connections do not lead to paradox, (iii) that it can capture the phenomenon of learning from experience, and (iv) that while the logic advocates scepticism with regard to some universal hypotheses, such scepticism is not problematic from the point of view of scientific theorising.     

THE epistemologies of parsimony: A review of Ockham's razors: A user's manual by Elliott Sober

Sober analyzes two paradigms of parsimony that have been used successfully in science. These are associated with two interpretations of probability: Bayesian and frequentist. Sober applies these paradigms to problems in biology, psychology, and philosophy. In the chapter on psychology, he argues that objective data consisting of environmental input and two or more concurrent responses could be used to refute empirically the radical behaviorist thesis that probability of learned responses can be accounted for solely on the basis of environmental variables. Sober believes that such data are readily available and offers a thought experiment to illustrate his point. Behavior analysts, however, would want actual experimental data, undoubtedly with animals, before accepting any such refutation. Nonetheless, Sober's philosophical point about the type of experiment that would be capable of refuting this thesis is valid. The behavior analytic program, however, does not depend upon the truth of this thesis.     

How to Be a Bayesian Dogmatist

Rational agents have (more or less) consistent beliefs. Bayesianism is a theory of consistency for partial belief states. Rational agents also respond appropriately to experience. Dogmatism is a theory of how to respond appropriately to experience. Hence, Dogmatism and Bayesianism are theories of two very different aspects of rationality. It's surprising, then, that in recent years it has become common to claim that Dogmatism and Bayesianism are jointly inconsistent: how can two independently consistent theories with distinct subject matter be jointly inconsistent? In this essay I argue that Bayesianism and Dogmatism are inconsistent only with the addition of a specific hypothesis about how the appropriate responses to perceptual experience are to be incorporated into the formal models of the Bayesian. That hypothesis isn't essential either to Bayesianism or to Dogmatism, and so Bayesianism and Dogmatism are jointly consistent. That leaves the matter of how experiences and credences are related, and so in the remainder of the essay I offer an alternative account of how perceptual justification, as the Dogmatist understands it, can be incorporated into the Bayesian formalism.     

Updating a progic

This paper presents a progic, or probabilistic logic, in the sense of Haenni et al. [8]. The progic presented here is based on Bayesianism, as the progic discussed by Williamson [15]. However, the underlying generalised Bayesianism differs from the objective Bayesianism used by Williamson, in the calibration norm, and the liberalisation and interpretation of the reference probability in the norm of equivocation. As a consequence, the updating dynamics of both Bayesianisms differ essentially. Whereas objective Bayesianism is based on a probabilistic re-evaluation, orthodox Bayesianism is based on a probabilistic revision. I formulate a generalised and iterable orthodox Bayesian revision dynamics. This allows to define an updating procedure for the generalised Bayesian progic. The paper compares the generalised Bayesian progic and Williamson's objective Bayesian progic in strength, update dynamics and with respect to language (in)sensitivity.     

Non-monotonicity and Informal Reasoning: Comment on Ferguson (2003)

In this paper, it is argued that Ferguson’s (2003, Argumentation 17, 335–346) recent proposal to reconcile monotonic logic with defeasibility has three counterintuitive consequences. First, the conclusions that can be derived from his new rule of inference are vacuous, a point that as already made against default logics when there are conflicting defaults. Second, his proposal requires a procedural “hack” to the break the symmetry between the disjuncts of the tautological conclusions to which his proposal leads. Third, Ferguson’s proposal amounts to arguing that all everyday inferences are sound by definition. It is concluded that the informal logic response to defeasibility, that an account of the context in which inferences are sound or unsound is required, still stands. It is also observed that another possible response is given by Bayesian probability theory (Oaksford and Chater, in press, Bayesian Rationality: The Probabilistic Approach to Human Reasoning, Oxford University Press, Oxford, UK; Hahn and Oaksford, in press, Synthese).     

Gibbard on meaning and normativity

The paper is a critique of Allan Gibbard’s impressively crafted monograph Meaning and Normativity. The book relies on a subtle form of logical empiricism, developing a normative verificationist semantics within a subjective Bayesian framework. I argue that the resulting account of synonymy is too fine-grained, since it counts clearly synonymous words in different languages as non-synonymous. For similar reasons, Gibbard’s account of analytic implication relies on postulating untenable connections between semantics and epistemology. I conclude that one of the main obstacles to robust realism about normativity and morality, the supposed conceptual connections between normative language and action, is a myth.     

Bayesian probability

Bayesian decision theory is here construed as explicating a particular concept of rational choice and Bayesian probability is taken to be the concept of probability used in that theory. Bayesian probability is usually identified with the agent’s degrees of belief but that interpretation makes Bayesian decision theory a poor explication of the relevant concept of rational choice. A satisfactory conception of Bayesian decision theory is obtained by taking Bayesian probability to be an explicatum for inductive probability given the agent’s evidence.     

Did Tom Kuhn actually Meet Tom Bayes?

Wesley Salmon and John Earman have presented influential Bayesian reconstructions of Thomas Kuhn’s account of theory-change. In this paper I argue that all attempts to give a Bayesian reading of Kuhn’s philosophy of science are fundamentally misguided due to the fact that Bayesian confirmation theory is in fact inconsistent with Kuhn’s account. The reasons for this inconsistency are traced to the role the concept of incommensurability plays with reference to the ‘observational vocabulary’ within Kuhn’s picture of scientific theories. The upshot of the discussion is that it is impossible to integrate both Kuhn’s claims and Bayesianism within a coherent account of theory-change.

Reasoning about evidence

Bayesians understand the notion of evidential support in terms of probability raising. Little is known about the logic of the evidential support relation, thus understood. We investigate a number of prima facie plausible candidate logical principles for the evidential support relation and show which of these principles the Bayesian evidential support relation does and which it does not obey. We also consider the question which of these principles hold for a stronger notion of evidential support.     

Assertion,Moore, and Bayes

It is widely believed that the so-called knowledge account of assertion best explains why sentences such as “It’s raining in Paris but I don’t believe it” and “It’s raining in Paris but I don’t know it” appear odd to us. I argue that the rival rational credibility account of assertion explains that fact just as well. I do so by providing a broadly Bayesian analysis of the said type of sentences which shows that such sentences cannot express rationally held beliefs. As an interesting aside, it will be seen that these sentences also harbor a lesson for Bayesian epistemology itself.

A Note on Binary Inductive Logic

We consider the problem of induction over languages containing binary relations and outline a way of interpreting and constructing a class of probability functions on the sentences of such a language. Some principles of inductive reasoning satisfied by these probability functions are discussed, leading in turn to a representation theorem for a more general class of probability functions satisfying these principles.     

Languages and designs for probability judgment

Theories of subjective probability are viewed as formal languages for analyzing evidence and expressing degrees of belief. This article focuses on two probability langauges, the Bayesian language and the language of belief functions (Shafer, 1976). We describe and compare the semantics (i.e., the meaning of the scale) and the syntax (i.e., the formal calculus) of these languages. We also investigate some of the designs for probability judgment afforded by the two languages.     

A Model of Jury Decisions where all Jurors have the same Evidence

Under the independence and competence assumptions of Condorcet’s classical jury model, the probability of a correct majority decision converges to certainty as the jury size increases, a seemingly unrealistic result. Using Bayesian networks, we argue that the model’s independence assumption requires that the state of the world (guilty or not guilty) is the latest common cause of all jurors’ votes. But often – arguably in all courtroom cases and in many expert panels – the latest such common cause is a shared ‘body of evidence’ observed by the jurors. In the corresponding Bayesian network, the votes are direct descendants not of the state of the world, but of the body of evidence, which in turn is a direct descendant of the state of the world. We develop a model of jury decisions based on this Bayesian network. Our model permits the possibility of misleading evidence, even for a maximally competent observer, which cannot easily be accommodated in the classical model. We prove that (i) the probability of a correct majority verdict converges to the probability that the body of evidence is not misleading, a value typically below 1; (ii) depending on the required threshold of ‘no reasonable doubt’, it may be impossible, even in an arbitrarily large jury, to establish guilt of a defendant ‘beyond any reasonable doubt’.     

The Logic and Meaning of Plurals. Part II

In this sequel to “The logic and meaning of plurals. Part I”, I continue to present an account of logic and language that acknowledges limitations of singular constructions of natural languages and recognizes plural constructions as their peers. To this end, I present a non-reductive account of plural constructions that results from the conception of plurals as devices for talking about the many. In this paper, I give an informal semantics of plurals, formulate a formal characterization of truth for the regimented languages that results from augmenting elementary languages with refinements of basic plural constructions of natural languages, and account for the logic of plural constructions by characterizing the logic of those regimented languages.

A Grim Semantics For Logics of Belief

Patrick Grim has presented arguments supporting the intuition that any notion of a totality of truths is incoherent. We suggest a natural semantics for various logics of belief which reflect Grim’s intuition. The semantics is a topological semantics, and we suggest that the condition can be interpreted to reflect Grim’s intuition. Beyond this, we present a natural canonical topological model for K4 and KD4.     

Probabilities on Sentences in an Expressive Logic

Automated reasoning about uncertain knowledge has many applications. One difficulty when developing such systems is the lack of a completely satisfactory integration of logic and probability. We address this problem directly. Expressive languages like higher-order logic are ideally suited for representing and reasoning about structured knowledge. Uncertain knowledge can be modeled by using graded probabilities rather than binary truth values. The main technical problem studied in this paper is the following: Given a set of sentences, each having some probability of being true, what probability should be ascribed to other (query) sentences? A natural wish-list, among others, is that the probability distribution (i) is consistent with the knowledge base, (ii) allows for a consistent inference procedure and in particular (iii) reduces to deductive logic in the limit of probabilities being 0 and 1, (iv) allows (Bayesian) inductive reasoning and (v) learning in the limit and in particular (vi) allows confirmation of universally quantified hypotheses/sentences. We translate this wish-list into technical requirements for a prior probability and show that probabilities satisfying all our criteria exist. We also give explicit constructions and several general characterizations of probabilities that satisfy some or all of the criteria and various (counter)examples. We also derive necessary and sufficient conditions for extending beliefs about finitely many sentences to suitable probabilities over all sentences, and in particular least dogmatic or least biased ones. We conclude with a brief outlook on how the developed theory might be used and approximated in autonomous reasoning agents. Our theory is a step towards a globally consistent and empirically satisfactory unification of probability and logic.     

Weak Bayesian coherentism

Recent results in probability theory have cast doubt on coherentism, purportedly showing (a) that coherence among a set of beliefs cannot raise their probability unless individual beliefs have some independent credibility, and (b) that no possible measure of coherence makes coherence generally probability-enhancing. I argue that coherentists can reject assumptions on which these theorems depend, and I derive a general condition under which the concurrence of two information sources lacking individual credibility can raise the probability of what they report.