The aim of this paper is to evaluate etiological accounts of functions for the domain of technical artefacts. Etiological theories ascribe functions to items on the basis of the causal histories of those items; they apply relatively straightforwardly to the biological domain, in which neo‐Darwinian evolutionary theory provides a well‐developed and generally accepted background for describing the causal histories of biological items. Yet there is no well‐developed and generally accepted theory for describing the causal history of artefacts, so the application (...) of etiological theories to the technical domain is hardly straightforward. In this paper we consider the transposition of etiological theories in general from the biological to the technical domain. We argue that a number of etiological theories that appear defensible for biology become untenable for technology. We illustrate our argument by showing that the standard etiological accounts of Neander and Millikan, and some recent attempts to improve on them, provide examples of such untenable theories.1Introduction2Desiderata for theories of functions3Etiological theories in general3.1Common core and divergent aims3.2Reproduction versus non‐reproduction etiological theories3.3Intentionalist versus non‐intentionalist etiological theories4Problems for etiological theories in the technical domain5The failure of existing reproduction theories6The failure of existing non‐reproduction theories7Improving reproduction by hybridisation8Conclusions. (shrink)

Over the past several decades, the philosophical community has witnessed the emergence of an important new paradigm for understanding the mind.1 The paradigm is that of machine computation, and its influence has been felt not only in philosophy, but also in all of the empirical disciplines devoted to the study of cognition. Of the several strategies for applying the resources provided by computer and cognitive science to the philosophy of mind, the one that has gained the most attention from philosophers (...) has been the Computational Theory of Mind (CTM). CTM was first articulated by Hilary Putnam (1960, 1961), but finds perhaps its most consistent and enduring advocate in Jerry Fodor (1975, 1980, 1981, 1987, 1990, 1994). It is this theory, and not any broader interpretations of what it would be for the mind to be a computer, that I wish to address in this paper. What I shall argue here is that the notion of symbolic representation employed by CTM is fundamentally unsuited to providing an explanation of the intentionality of mental states (a major goal of CTM), and that this result undercuts a second major goal of CTM, sometimes refered to as the vindication of intentional psychology. This line of argument is related to the discussions of derived intentionality by Searle (1980, 1983, 1984) and Sayre (1986, 1987). But whereas those discussions seem to be concerned with the causal dependence of familiar sorts of symbolic representation upon meaning-bestowing acts, my claim is rather that there is not one but several notions of meaning to be had, and that the notions that are applicable to symbols are conceptually dependent upon the notion that is applicable to mental states in the fashion that Aristotle refered to as paronymy. That is, an analysis of the notions of meaning applicable to symbols reveals that they contain presuppositions about meaningful mental states, much as Aristotle's analysis of the sense of healthy that is applied to foods reveals that it means conducive to having a healthy body, and hence any attempt to explain mental semantics in terms of the semantics of symbols is doomed to circularity and regress. I shall argue, however, that this does not have the consequence that computationalism is bankrupt as a paradigm for cognitive science, as it is possible to reconstruct CTM in a fashion that avoids these difficulties and makes it a viable research framework for psychology, albeit at the cost of losing its claims to explain intentionality and to vindicate intentional psychology. I have argued elsewhere (Horst, 1996) that local special sciences such as psychology do not require vindication in the form of demonstrating their reducibility to more fundamental theories, and hence failure to make good on these philosophical promises need not compromise the broad range of work in empirical cognitive science motivated by the computer paradigm in ways that do not depend on these problematic treatments of symbols. (shrink)

According to an idea which is widespread among philosophers, linguistic entities derive their intentionality from the intentionality of mental entities by virtue of some relation between them. Typically, it is some kind of intention on the speaker’s part – e.g., an intention to produce in the hearer a belief with a certain content – that is supposed to endow words with content. This paper argues that the concept of the derivation of content from one entity to another, if understood realistically, (...) is flawed: derived intentionality, I will argue, is merely ascribed intentionality, not a real property of its possessor. Irrealistic-ascriptivist senses are suggested for the ideas of content derivation, of original intentionality, and of the mind as the source of linguistic intentionality. Thus, endorsing the idea that mental intentionality is the source of non-mental intentionality need not tempt one to intentional realism. In an intentional irrealistic framework, what forms of intentionality are original and what are derived is a deeply contingent matter, determined by our practice of content ascription. But while intentional irrealism accommodates all those ideas, this paper defends “content-derivation irrealism” but not thoroughgoing intentional irrealism – the idea that there is real original intentionality is not ruled out. Still, assuming that some entities possess real intentionality, what can make them endow intentionality upon other entities is also our practice of content ascription. (shrink)

Most contemporary philosophers of mind claim to be in search of a 'naturalistic' theory. However, when we look more closely, we find that there are a number of different and even conflicting ideas of what would count as a 'naturalization' of the mind. This article attempts to show what various naturalistic philosophies of mind have in common, and also how they differ from one another. Additionally, it explores the differences between naturalistic philosophies of mind and naturalisms found in ethics, epistemology, (...) and philosophy of science. Section 1 introduces a distinction between two types of project that have been styled 'naturalistic', which I call philosophical naturalism and empirical naturalism . Sections 2 to 6 canvass different strands of philosophical naturalism concerning the mind, followed by a much briefer discussion of attempts to provide empirical naturalizations of the mind in Section 7 . Section 8 concludes the essay with a consideration of the relations between philosophical and empirical naturalism in philosophy of mind, arguing that at least some types of philosophical naturalism are incompatible with empirical naturalism. (shrink)

Reflections on Meaning refines Paul Horwich’s use theory of meaning. Horwich holds that the meaning of a word is constituted by the nonsemantic property that best explains a certain law. For a given word, the law to be explained governs that word’s use by specifying the “acceptance conditions” of a privileged class of sentences containing the word (26). Horwich devotes considerable energy to details in Reflections on Meaning and focuses on especially pressing problems for his use theory of meaning. As (...) a result the book’s topics run the gamut, and the connections between its chapters are not always strong. Rather than try to provide a synoptic overview, I’ll discuss three areas where it seems further clarification and detail could be fruitful: the distinction between semantic and nonsemantic properties, context sensitivity, and compositionality. Horwich thinks ours is a “fundamentally non-semantic world” (27), making it crucial that meaning be explained in nonsemantic terms. In particular, he insists that we “exclude from the analyzing-properties [of word meaning] anything that would itself require analysis in terms of meaning”: we can’t appeal to reference, belief, or intention, for example (37). But Horwich does not object to “accounts of meaning in psychological terms,” and his own theory relies heavily on a psychological, nonsemantic relation that Horwich calls “acceptance” (37). It’s difficult to see a substantive difference between this technical notion and belief. Acceptance, for Horwich, is “the psychological (but nonsemantic) relation to a sentence that is manifested in our relying on it as a premise in theoretical and practical inference” (40–41). Belief, on the other hand, is a relation with these properties except that it is semantic. Horwich’s other characterizations of acceptance don’t sharpen the distinction very much: “S accepts a sentence just in case that sentence, or its mental correlate, is in S’s belief box” (41); “believing a given proposition is nothing more than accepting some sentence that expresses it” (61).. (shrink)

In this essay, I shall defend a transcendental argument for epistemological realism: the view that mind-independent yet cognitively accessible entities exist. The proposed argument reasons from the fact that we are conceptual creatures to the existence of a knowable outer world as a condition of the possibility of such creatures. I first lay down my general approach to concepts and conceptualization, according to which concepts are rules that agents follow in their cognitive activities. I go on to explicate the peculiar (...) normative nature of rules and rule-following, from which I extract, following Wittgensteinian considerations, an intractable problem for any idealist account of concept possession. I argue that the very possibility of conceptualizing requires the existence of external objects that enable the cognizer to regulate their use of concepts, and close with some remarks on the resultant nature and scope of metaphysical knowledge. (shrink)

First, we discuss issues raised with respect to the Theory of Event Coding (TEC)'s scope, that is, its limitations and possible extensions. Then, we address the issue of specificity, that is, the widespread concern that TEC is too unspecified and, therefore, too vague in a number of important respects. Finally, we elaborate on our views about TEC's relations to other important frameworks and approaches in the field like stages models, ecological approaches, and the two-visual-pathways model. Footnotes1 We acknowledge the precedence (...) of both Freud¹s Instincts and Their Vicissitudes (1915) and Neisser¹s Stimulus Information and Its Vicissitudes (a term Neisser borrowed from Freud for his monograph “Cognitive psychology,” 1967). (shrink)

In this paper I examine the connection between accounts of biological teleology and the biocentrist claim that all living beings have a good of their own. I first present the background for biocentrists’ appeal to biological teleology. Then I raise a problem of scope for teleology-based biocentrism and, drawing in part on recent work by Basl and Sandler, I discuss Taylor and Varner’s responses to this problem. I then challenge Basl and Sandler’s own response to the scope problem for its (...) reliance on a selectionist account of organismic teleology. Finally I examine the prospects for a biocentrist response to the problem of scope based on an alternative organisational account of internal teleology. I conclude by assessing the prospects for teleology-based biocentrism. (shrink)

A popular defense of physicalist theories of consciousness against anti-physicalist arguments invokes the existence of ‘phenomenal concepts’. These are concepts that designate conscious experiences from a first person perspective, and hence differ from physicalistic concepts; but not in a way that precludes co-referentiality with them. On one version of this strategy phenomenal concepts are seen as (1) type demonstratives that have (2) no mode of presentation. However, 2 is possible without 1-call this the ‘bare recognitional concept’ view-and I will argue (...) that this avoids certain recent criticisms while retaining the virtue of finessing the ‘mode of presentation’ problem for phenomenal concepts. But construing phenomenal concepts this way seems to not do justice to the phenomenology of conscious experience. In this paper I examine whether or not this impression can be borne out by a good argument. As it turns out, it is harder to do so than one might think. It can be done, but it involves somewhat more convoluted reasoning than one might have supposed necessary. Having shown that, I will end with a few brief remarks on what my argument means for attempts to preserve a physicalist account of consciousness. (shrink)

Frank Jackson formulated his knowledge argument as an argument for dualism. In this paper I show how the argument can be modified to also establish the irreducibility of the secondary qualities to the properties of physical theory, and ultimately "secondary quality eliminativism"- the view that the secondary qualities are physically uninstantiated.

According to the so-called metaphysical conception of analyticity, analytic truths are true in virtue of meaning (or content) alone and independently of (extralinguistic) facts. Quine and Boghossian have tried to present a conclusive argument against the metaphysical conception of analyticity. In effect, they tried to show that the metaphysical conception inevitably leads into a highly implausible view about the truthmakers of analytic truths. We would like to show that their argument fails, since it relies on an ambiguity of the notion (...) of 'independence of (extralinguistic) facts'. If one distinguishes between variation independence and existence independence, the unwelcome view about the truthmakers of analytic truths no longer follows. Thus, there is at best a challenge, but no conclusive argument. The door to the metaphysical conception of analyticity is still open. (shrink)

More intellectual modesty, but also conceptual clarity is urgently needed in AI, perhaps more than in many other disciplines. AI research has been coined by hypes and hubris since its early beginnings in the 1950s. For instance, the Nobel laureate Herbert Simon predicted after his participation in the Dartmouth workshop that “machines will be capable, within 20 years, of doing any work that a man can do”. And expectations are in some circles still high to overblown today. This paper addresses (...) the demand for conceptual clarity and introduces precise definitions of “strong AI”, “superintelligence”, the “technological singularity”, and “artificial general intelligence” which ground in the work by the computer scientist Judea Pearl and the psychologist Howard Gardner. These clarifications allow us to embed famous arguments from the philosophy of AI in a more analytic context. (shrink)

Linguistic meaning underdetermines what is said. This has consequences for philosophical accounts of meaning, communication, and propositional attitude reports. I argue that the consequence we should endorse is that utterances typically express many propositions, that these are what speakers mean, and that the correct semantics for attitude reports will handle this fact while being relational and propositional.

When someone masters a skill, their performance looks to us like second nature: it looks as if their actions are smoothly performed without explicit, knowledge-driven, online monitoring of their performance. Contemporary computational models in motor control theory, however, are instructionist: that is, they cast skillful performance as a knowledge-driven process. Optimal motor control theory, as representative par excellence of such approaches, casts skillful performance as an instruction, instantiated in the brain, that needs to be executed—a motor command. This paper aims (...) to show the limitations of such instructionist approaches to skillful performance. We specifically address the question of whether the assumption of control-theoretic models is warranted. The first section of this paper examines the instructionist assumption, according to which skillful performance consists of the execution of theoretical instructions harnessed in motor representations. The second and third sections characterize the implementation of motor representations as motor commands, with a special focus on formulations from OMCT. The final sections of this paper examine predictive coding and active inference—behavioral modeling frameworks that descend, but are distinct, from OMCT—and argue that the instructionist, control-theoretic assumptions are ill-motivated in light of new developments in active inference. (shrink)

Teleosemantics is a naturalistic research programme in the philosophy of mind and language. Its ambition is to achieve a reduction, first, of mental content to teleological function; second, of teleological function to non-teleological notions. This article explores the second step, particularly as envisaged by Millikan’s etiological theory of function.

This paper clarifies and answers the following question: is technology morally neutral? It is argued that the debate between proponents and opponents of the Neutrality Thesis depends on different underlying assumptions about the nature of technological artifacts. My central argument centres around the claim that a mere physicalistic vocabulary does not suffice in characterizing technological artifacts as artifacts, and that the concepts of function and intention are necessary to describe technological artifacts at the right level of description. Once this has (...) been established, I demystify talk about the possible value-ladenness of technological artifacts by showing how these values can be empirically identified. I draw from examples in biology and the social sciences to show that there is a non-mysterious sense in which functions and values can be empirically identified. I conclude from this that technology can be value-laden and that its value-ladenness can both derive from the intended functions as well as the harmful non-intended functions of technological artifacts. (shrink)

P.F. Strawson contributed to many philosophical domains, including the philosophy of language, the history of philosophy, metaphysics, moral philosophy and philosophical methodology. Most of his contributions in these areas have influenced contemporary debates, either because his views are still defended or because they are still considered worthy of detailed responses. His views on the philosophy of logic have been only rarely discussed, however. My aim in this paper is threefold. First, I provide a systematic account of Strawson’s philosophy of logic. (...) I argue that Strawson is an important predecessor of logical expressivism, a contemporary position in the philosophy of logic most notably defended by Robert Brandom. My main focus is on Strawson’s largely-neglected 1982 paper ‘Logical Form and Logical Constants’, which contains his most mature ideas on the topic. Second, while Strawson’s position is of historical and independent philosophical interest, I argue that he leaves many points unclear. Finally, I propose several clarifications of Strawson’s position. (shrink)

Dennett's intended rapprochement between physical realism and intentional relativism fails because it is premised upon conflicting arguments governing the status of design. Indeed, Dennett's remarks on design serve to highlight tensions buried deep within his theory. For inasmuch as Dennett succeeds in objectifying attributions of design, attributions of intentionality readily follow suit, leading to a form of intentional realism. But inasmuch as Dennett is successful in relativizing attributions of design, scientific realism at large is subject to renewed anti-realistic criticism. Dennettian-inspired (...) considerations of adaptationism substantiate the former move towards intentional realism, while considerations of the relativity of artifactual design encourage the latter move towards physical relativism. The ambivalence intrinsic to Dennett's ``mild realism'' can be viewed as a function of these two conflicting positions on design, for Dennett can no more avoid objectifying intentionality when he is realistic about design than he can avoid relativizing physical causality when relativistic about design. (shrink)

This article looks at some of the metaphysical properties of cognitive artefacts. It first identifies and demarcates the target domain by conceptualizing this class of artefacts as a functional kind. Building on the work of Beth Preston, a pluralist notion of functional kind is developed, one that includes artefacts with proper functions and system functions. Those with proper functions have a history of cultural selection, whereas those with system functions are improvised uses of initially non-cognitive artefacts. Having identified the target (...) domain, it then briefly looks at the multiple usability of physical structures and the multiple realizability of cognitive function. Further developing insights from the “dual nature of artefacts thesis”, the article ends with conceptualizing the structure–function relations of cognitive artefacts. More specifically, it unpacks the relation between physical structure, representational structure, information, and cognitive function. (shrink)

This article looks at some of the metaphysical properties of cognitive artefacts. It first identifies and demarcates the target domain by conceptualizing this class of artefacts as a functional kind. Building on the work of Beth Preston, a pluralist notion of functional kind is developed, one that includes artefacts with proper functions and system functions. Those with proper functions have a history of cultural selection, whereas those with system functions are improvised uses of initially non-cognitive artefacts. Having identified the target (...) domain, it then briefly looks at the multiple usability of physical structures and the multiple realizability of cognitive function. Further developing insights from the “dual nature of artefacts thesis”, the article ends with conceptualizing the structure–function relations of cognitive artefacts. More specifically, it unpacks the relation between physical structure, representational structure, information, and cognitive function. (shrink)

It is often thought that non-junk or coding DNA is more significant than other cellular elements, including so-called junk DNA. This is for two main reasons: because coding DNA is often targeted by historical or current selection, it is considered functionally special and because its mode of action is uniquely specific amongst the other actual difference makers in the cell, it is considered causally special. Here, we challenge both these presumptions. With respect to function, we argue that there is previously (...) unappreciated reason to think that junk DNA is significant, since it can alter the cellular environment, and those alterations can influence how organism-level selection operates. With respect to causality, we argue that there is again reason to think that junk DNA is significant, since it too is remarkably causally specific. As a result, something is missing from the received view of significance in molecular biology—a view which emphasizes specificity and neglects something we term ‘reach’. With the special case of junk DNA in mind, we explore how to model and understand the causal specificity, reach, and corresponding efficacy of difference makers in biology. The account contains implications for how evolution shapes the genome, as well as advances our understanding of multi-level selection. (shrink)

John Searle's Chinese room argument is perhaps the most influential andwidely cited argument against artificial intelligence (AI). Understood astargeting AI proper – claims that computers can think or do think– Searle's argument, despite its rhetorical flash, is logically andscientifically a dud. Advertised as effective against AI proper, theargument, in its main outlines, is an ignoratio elenchi. It musterspersuasive force fallaciously by indirection fostered by equivocaldeployment of the phrase "strong AI" and reinforced by equivocation on thephrase "causal powers" (at least) equal (...) to those of brains." On a morecarefully crafted understanding – understood just to targetmetaphysical identification of thought with computation ("Functionalism"or "Computationalism") and not AI proper the argument is still unsound,though more interestingly so. It's unsound in ways difficult for high church– "someday my prince of an AI program will come" – believersin AI to acknowledge without undermining their high church beliefs. The adhominem bite of Searle's argument against the high church persuasions of somany cognitive scientists, I suggest, largely explains the undeserved reputethis really quite disreputable argument enjoys among them. (shrink)

Harnad''s proposed robotic upgrade of Turing''s Test (TT), from a test of linguistic capacity alone to a Total Turing Test (TTT) of linguisticand sensorimotor capacity, conflicts with his claim that no behavioral test provides even probable warrant for attributions of thought because there is no evidence of consciousness besides private experience. Intuitive, scientific, and philosophical considerations Harnad offers in favor of his proposed upgrade are unconvincing. I agree with Harnad that distinguishing real from as if thought on the basis of (...) (presence or lack of) consciousness (thus rejecting Turing (behavioral) testing as sufficient warrant for mental attribution)has the skeptical consequence Harnad accepts — there is in factno evidence for me that anyone else but me has a mind. I disagree with hisacceptance of it! It would be better to give up the neo-Cartesian faith in private conscious experience underlying Harnad''s allegiance to Searle''s controversial Chinese Room Experiment than give up all claim to know others think. It would be better to allow that (passing) Turing''s Test evidences — evenstrongly evidences — thought. (shrink)

This note addresses two of Gibbard's central contentions in Meaning and Normativity: first, that the concept of meaning is normative, and second, that an expressivist account of semantic concepts and statements can shed light on the hard problem of intentionality, the problem of explaining intentionality in naturalistic terms.

Many people claim that semantic content is normative, and that therefore naturalistic theories of content face a potentially insuperable difficulty. The normativity of content allegedly undermines naturalism by introducing a gap between semantic 'ought's and the explanatory resources of naturalism. I argue here that this problem is not ultimately pressing for naturalists. The normativity thesis, I maintain, is ambiguous; it could mean either that the content of a term prescribes a pattern of use, or that it merely determines which pattern (...) of use can be described as 'correct'. For the antinaturalist argument to go forward, content must be prescriptive. I argue, however, that it is not. Moreover, the thesis that content supplies standards for correct use is insufficient to supply a similar, a priori objection to naturalism. (shrink)

A new theory that naturalizes biological function is explained and compared with earlier etiological and causal role theories. Etiological theories explain functions from how they are caused over their evolutionary history. Causal role theories analyze how functional mechanisms serve the current capacities of their containing system. The new proposal unifies the key notions of both kinds of theories, but goes beyond them by explaining how functions in an organism can exist as factors with autonomous causal efficacy. The goal-directedness and normativity (...) of functions exist in this strict sense as well. The theory depends on an internal physiological or neural process that mimics an organism’s fitness, and modulates the organism’s variability accordingly. The structure of the internal process can be subdivided into subprocesses that monitor specific functions in an organism. The theory matches well with each intuition on a previously published list of intuited ideas about biological functions, including intuitions that have posed difficulties for other theories. (shrink)

Journal of Social Philosophy, EarlyView.

What does a naturalistic epistemologist do? Content Type Journal Article Pages 1-4 DOI 10.1007/s11016-011-9531-7 Authors William F. Harms, Humanities and Social Sciences, Seattle Central Community College, 1701 Broadway, Seattle, WA 98122-9905, USA Journal Metascience Online ISSN 1467-9981 Print ISSN 0815-0796.

Information theory offers a measure of "mutual information" which provides an appropriate measure of tracking efficiency for the naturalistic epistemologist. The statistical entropy on which it is based is arguably the best way of characterizing the uncertainty associated with the behavior of a system, and it is ontologically neutral. Though not appropriate for the naturalization of meaning, mutual information can serve as a measure of epistemic success independent of semantic maps and payoff structures. While not containing payoffs as terms, mutual (...) information places both upper and lower bounds on payoffs. This constitutes a non-trivial relationship to utility. (shrink)

Much work in contemporary philosophy of mind and neurophilosophy hinges on the concept of ‘representation,’ but that concept inherits a problematic ambiguity from neuroscience, where scientists may distinguish between cognitive and physiological levels of representation only tacitly. First, I explicate two potentially distinct senses of representation corresponding to these levels. I then argue that ambiguity about the nature of representation in philosophy of mind is problematic for at least one prominent philosophical project that aims to use neuroscientific work on representation (...) to defend the existence or explanatory relevance of intentional mental states, namely Schroeder’s (2004) scientific explication of desire in terms of the neurobiology of reward. I argue that philosophical treatments of the relationship between cognitive and neural architecture must attend more carefully to the ambiguity in the concept of representation. I conclude by outlining a strategy for addressing the gap between levels of representation, one that privileges a local or narrow philosophical approach to interpreting scientific concepts and data over a global or general approach. (shrink)

Despite its centrality in the philosophy of cognitive science, there has been little prior philosophical work engaging with the notion of representation in contemporary NLP practice. This paper attempts to fill that lacuna: drawing on ideas from cognitive science, I introduce a framework for evaluating the representational claims made about components of neural NLP models, proposing three criteria with which to evaluate whether a component of a model represents a property and operationalising these criteria using probing classifiers, a popular analysis (...) technique in NLP (and deep learning more broadly). The project of operationalising a philosophically-informed notion of representation should be of interest to both philosophers of science and NLP practitioners. It affords philosophers a novel testing-ground for claims about the nature of representation, and helps NLPers organise the large literature on probing experiments, suggesting novel avenues for empirical research. (shrink)

In his Explaining Behavior, Fred Dretske uses a reliabilist theory of representation to try to vindicate the use of intentional explanation for behaviour against latter-day elitninativism. Although Dretske's indicator semantics turns on the notion of function, he himself never explicitly defines what function means. Dretske's reticence in discussing function may ultimately be an error, for, as I argue, his implicit understanding of what a function amounts to does not fit with data from op rant conditioning. Still, this need not be (...) a deep flaw in Dretske and I offer one way in which we may patch up the notion of function via the changes known to occur with learning in the brain. Ultimately, I conclude that the only facts needed to explain behaviour are (1) the conditions in the world that are paired with neuronal circuit activation (as picked out by goals in some circumstances); and (2) what motor output that condition triggers. (shrink)