Justifying the evidential use of linguistic intuitions

Linguistic intuitive judgements are the de facto data source of choice within generative linguistics. But why we are justified in relying on intuitive judgements as evidence for grammars? In the philosophy of linguistics, this question has been hotly debated. I argue that the three most prominent views of that debate all have their problems. Devitt’s Modest Explanation accounts for the wrong kind of intuitive judgements. The Voice of Competence view and Rey’s account both lack independent evidence. I introduce and defend a novel proposal that accounts for the evidential role of linguistic intuitive judgements and avoids these shortcomings. On this account, linguistic intuitive judgements are reports of the speaker’s immediate experience of trying to comprehend the sentence. This experience is due to the speaker’s linguistic competence, at least in part, and so the justification for the evidential use of linguistic intuitions ultimately comes from the speaker’s competence. However, the account does not rely on any special input from the speaker’s competence being available as the basis for linguistic intuitive judgements.

     

A competence framework for artificial intelligence research

ABSTRACT

While over the last few decades AI research has largely focused on building tools and applications, recent technological developments have prompted a resurgence of interest in building a genuinely intelligent artificial agent – one that has a mind in the same sense that humans and animals do. In this paper, I offer a theoretical and methodological framework for this project of investigating “artificial minded intelligence” (AMI) that can help to unify existing approaches and provide new avenues for research. I first outline three desiderata that a framework for AMI research should satisfy. In Section 1, I further motivate the desiderata as well as the need for a new framework. Section 2 develops a general methodological approach, the “generative methodology,” and Section 3 develops a version of this methodology, the “Competence Framework for AI research” (CFAI).     

Using models to correct data: paleodiversity and the fossil record

Despite an enormous philosophical literature on models in science, surprisingly little has been written about data models and how they are constructed. In this paper, I examine the case of how paleodiversity data models are constructed from the fossil data. In particular, I show how paleontologists are using various model-based techniques to correct the data. Drawing on this research, I argue for the following related theses: first, the ‘purity’ of a data model is not a measure of its epistemic reliability. Instead it is the fidelity of the data that matters. Second, the fidelity of a data model in capturing the signal of interest is a matter of degree. Third, the fidelity of a data model can be improved ‘vicariously’, such as through the use of post hoc model-based correction techniques. And, fourth, data models, like theoretical models, should be assessed as adequate (or inadequate) for particular purposes.

     

Proxytypes and Linguistic Nativism

Prinz (Perceptual the Mind: Concepts and Their Perceptual Basis, MIT Press, 2002) presents a new species of concept empiricism, under which concepts are off-line long-term memory networks of representations that are ‘copies’ of perceptual representations – proxytypes. An apparent obstacle to any such empiricism is the prevailing nativism of generative linguistics. The paper critically assesses Prinz’s attempt to overcome this obstacle. The paper argues that, prima facie, proxytypes are as incapable of accounting for the structure of the linguistic mind as are the more traditional species of empiricism. This position is then confirmed by looking in detail at two suggestions (one derived from recent connectionist research) from Prinz of how certain aspects of syntactic structure might be accommodated by the proxytype theory. It is shown that the suggestions fail to come to terms with both the data and theory of contemporary linguistics.     

Anti-Skepticism Under a Linguistic Guise

In this paper I consider the plausibility of developing anti-skepticism by framing the question in linguistic terms: instead of asking whether we know, we ask what falls within the extension of the word “know”. I first trace two previous attempts to develop anti-skepticism in this way, from Austin (particularly as presented by Kaplan) and from epistemic contextualism, and I present reasons to think that both approaches are unsuccessful. I then focus on a recently popular attempt to develop anti-skepticism from the “function-first” approach associated with Craig, which I also show to be problematic. I then argue that the apparent prima facie plausibility of fighting skepticism on linguistic grounds is due to a methodological spill-over from linguistics. Once we recognize this, it becomes clear that the skepticism debate should not be conducted in linguistic terms.

     

Normative principles and the nature of mind-dependence

One of the most fundamental debates in metaethics is whether (subsets of) the normative facts are mind-dependent. Yet some philosophers are skeptical that mind-dependence is a category that's significant in the way metaethicists have assumed it is. In this paper, I consider a puzzle that showcases this skepticism, explaining how it undermines the most natural reading of the mind-dependence claim. I then go on to show that no modification of this reading within a certain class can hope to solve the problem. I conclude by suggesting a new way that mind-dependence should be understood: mind-dependence is ultimately a matter of how normative principles are grounded. I develop this view briefly before concluding.

     

A Difference of Some Consequence Between Conventions and Rules

Lewis’s view of the way conventions are passed on may have some especially interesting consequences for the study of language. I’ll start by briefly discussing agreements and disagreements that I have with Lewis’s general views on conventions and then turn to how linguistic conventions spread. I’ll compare views of main stream generative linguistics, in particular, Chomsky’s views on how syntactic forms are passed on, with the sort of view of language acquisition and language change advocated by usage-based or construction grammars, which seem to fit better with Lewis’s ideas. Then I will illustrate the interest of Lewis’s perspective on the dissemination of conventions with a variety of linguistic examples.     

Statistical models for the induction and use of selectional preferences

Selectional preferences have a long history in both generative and computational linguistics. However, since the publication of Resnik's dissertation in 1993, a new approach has surfaced in the computational linguistics community. This new line of research combines knowledge represented in a pre‐defined semantic class hierarchy with statistical tools including information theory, statistical modeling, and Bayesian inference. These tools are used to learn selectional preferences from examples in a corpus. Instead of simple sets of semantic classes, selectional preferences are viewed as probability distributions over various entities. We survey research that extends Resnik's initial work, discuss the strengths and weaknesses of each approach, and show how they together form a cohesive line of research.     

Realism,Reliabilism, and the ‘Strong Programme’ in the Sociology of Scientific Knowledge

In this essay, I respond to Tim Lewens’s proposal that realists and Strong Programme theorists can find common ground in reliabilism. I agree with Lewens, but point to difficulties in his argument. Chief among these is his assumption that reliabilism is incompatible with the Strong Programme’s principle of symmetry. I argue that the two are, in fact, compatible, and that Lewens misses this fact because he wrongly supposes that reliabilism entails naturalism. The Strong Programme can fully accommodate a reliabilism which has been freed from its inessential ties to naturalism. Unlike naturalistic epistemologists, the Strong Programme’s sociologistic reliabilist insists that all scientific facts are the product of both natural and social causal phenomena. Anticipating objections, I draw on Wittgenstein’s rule‐following considerations to explain how the sociologistic reliabilist can account for standard intuitions about the objective elements of knowledge. I also explain how the Strong Programme theorist can distinguish between a belief’s seeming reliable and its being reliable.

?Ich setzte

?den Fuß in die Luft,

?und sie trug.

?(Hilde Domin)     

Deflationism about the necessary a posteriori and Twin Earth

Some necessary truths are a posteriori. That’s widely agreed and is presumed here. Their existence might appear to show that discoveries about how things are in fact—about how things actually are—can lead to discoveries about all the ways things might be, about the nature of logical space. I detail one way of resisting this conclusion for a number of examples, and the implications of Twin Earth for the issue. Central is the notion of a Cambridge discovery.

     

Why We Did It: An Anscombian Account of Collective Action

Abstract

In this paper, I am concerned with persons’ capacity for joint action. I start by suggesting that approaches which seek to account for that capacity in terms of collective intentionality face a problem: there are actions that clearly seem to qualify as collective even though the involved persons cannot be said to entertain an overarching ‘We’‐intention (however one characterizes this notion). I then go on to develop an alternative account of action that loosely draws on Elizabeth Anscombe’s action theory and show how this alternative account can be applied to joint action. In so doing, I stress the importance of the phenomenal dimension of agency.     

Licking and liking: The assessment of hedonic responses in rodents

Affective processes are a key determinant of behaviour: At its simplest, liked stimuli are approached while disliked stimuli are avoided. Although assessing hedonic responses in nonverbal animals can be difficult, one relatively tractable approach relies on detailed analyses of rodents' consummatory behaviour. Rodents typically produce rhythmic sets of licks that can be grouped into clusters on the basis of the intervals between licks. The mean number of licks in a cluster (cluster size) is directly related to the concentration of palatable and unpalatable solutions. These relationships suggest that lick cluster size might be a useful index of an animal's hedonic reaction to the solution being consumed. I begin by reviewing studies of conditioned flavour preference and aversion that support the idea that lick cluster size can provide useful information about rats' hedonic reactions. I then describe how this methodology has been used to address previously intractable issues in the investigation of contrast effects as well as revealing an analogue of effort justification effects that, in humans, are commonly explained in terms of cognitive dissonance reduction. Finally, I consider how lick analysis might provide information about hedonic responses in animal models of human psychiatric disorders. In all these cases, how an animal did something was particularly informative about why it was doing it.

     

Evolution, selection and cognition: from "learning" to parameter setting in biology and in the study of language

Most biologists and some cognitive scientists have independently reached the conclusion that there is no such thing as learning in the traditional "instructive" sense. This is, admittedly, a somewhat extreme thesis, but I defend it here in the light of data and theories jointly extracted from biology, especially from evolutionary theory and immunology, and from modern generative grammar. I also point out that the general demise of learning is uncontroversial in the biological sciences, while a similar consensus has not yet been reached in psychology and in linguistics at large. Since many arguments presently offered in defense of learning and in defense of "general intelligence" are often based on a distorted picture of human biological evolution, I devote some sections of this paper to a critique of "adaptationism," providing also a sketch of a better evolutionary theory (one based on "exaptation"). Moreover, since certain standard arguments presented today as "knock-down" in psychology, in linguistics and in artificial intelligence are a perfect replica of those once voiced by biologists in favor of instruction and against selection, I capitalize on these errors of the past to draw some lessons for the present and for the future.     

Does rationality presuppose irrationality

Abstract

Although irrationality always presupposes rationality, I think there are good arguments to claim that sometimes rationality presupposes irrationality.This paper tries to show how irrational action can support rationality in two ways: it can develop and preserve rationality. I also argue that sometimes the development and the conservation of rationality can only be realized by irrational action.     

Suspending is believing

A good account of the agnostic attitude of Suspending Judgement should explain how it can be rendered more or less rational/justified according to the state of one’s evidence—and one’s relation to that evidence. I argue that the attitude of suspending judgement whether p constitutively involves having a belief; roughly, a belief that one cannot yet tell whether or not p. I show that a theory of suspending that treats it as a sui generis attitude, wholly distinct from belief, struggles to account for how suspension of judgement can be rendered more or less rational (or irrational) by one’s evidence. I also criticise the related idea that suspension essentially requires an ‘Inquiring Attitude’. I show how a belief-based theory, in contrast, neatly accounts for the rational and epistemic features of suspending and so neatly accounts for why an agnostic has a genuine neutral opinion concerning the question whether p, as opposed to simply having no opinion.

     

Compositional Semantics and Normative ‘Ought’

According to the paradigm view in linguistics and philosophical semantics, it is lexical semantics (LS) plus the principle of compositionality (PC) that allows us to compute the meaning of an arbitrary sentence. The job of LS is to assign meaning to individual expressions, whereas PC says how to combine these individual meanings into larger ones. In this paper I argue that the pair LS?+?PC fails to account for the discourse-relevant meaning of normative ‘ought’. If my hypothesis is tenable, then the failure of LS?+?CS extends to normative language in general. The reason I offer that this is so is that semantics for normative language is, in an important respect, a substantive semantics (SS). The ‘substantive’ in question means that the meaning of normative vocabulary in use is driven by metanormative views associated with a particular normative concept. SS rejects the model LS?+?CS and replaces it with a discourse-relevant semantics built around an interactional principle that ascribes to a particular surface syntactical form of ‘ought’ sentences a logical form that represents its discourse-salient normative content. In the paper I shall sketch how SS works and why it is worth serious consideration.

     

The role of prediction in mental processing: A process approach

Although prediction plays a prominent role in mental processing, we have only limited understanding of how the brain generates and employs predictions. This paper develops a theoretical framework in three steps. First I propose a process model that describes how predictions are produced and are linked to behavior. Subsequently I describe a generative mechanism, consisting of the selective amplification of neural dynamics in the context of boundary conditions. I hypothesize that this mechanism is active as a process engine in every mental process, and that therefore each mental process proceeds in two stages: (i) the formation of process boundary conditions; (ii) the bringing about of the process function by the operation – within the boundary conditions – of a relatively ‘blind’ generative process. Thirdly, from this hypothesis I derive a strategy for describing processes formally. The result is a multilevel framework that may also be useful for studying mental processes in general.     

Cowie on the Poverty of Stimulus

My paper defends the use of the poverty of stimulus argument (POSA) for linguistic nativism against Cowie's (1999) counter-claim that it leaves empiricism untouched. I first present the linguistic POSA as arising from a reflection on the generality of the child's initial state in comparison with the specific complexity of its final state. I then show that Cowie misconstrues the POSA as a direct argument about the character of the pld. In this light, I first argue that the data Cowie marshals about the pld does not begin to suggest that the POSA is unsound. Second, through a discussion of the so-called `auxiliary inversion rule', I show, by way of diagnosis, that Cowie misunderstands both the methodology of current linguistics and the complexity of the data it is obliged to explain.     

Freedom,Firearms, and Civil Resistance

The claim that guns can safeguard freedom is common in US political discourse. In light of a broadly republican understanding of freedom, I evaluate this claim and its implications. The idea is usually that firearms would enable citizens to engage in revolutionary violence against a tyrannical government. I argue that some of the most common objections to this argument fail, but that the argument is fairly weak in light of other objections. I then defend a different argument for the claim that guns can safeguard freedom. I claim that firearm ownership among members of oppressed groups can hinder the use of systematic violence aimed at preventing them from exercising their basic liberties. I show how a commitment to armed self-defense is compatible with non-violent civil resistance as a tool of political change, and show how the former facilitated the latter during the Civil Rights Movement. Finally, I consider the policy implications of my argument. I don’t think it vindicates lax gun control policies. However, it may vindicate some individuals acquiring guns and learning how to use them, and some organizations aiding them in doing so.