Ontology and evaluation of theories in cognitive science concepts ** Guido Vallejos
Centro de Estudios Cognitivos
Departamento de Filosofía
Facultad de Filosofía y Humanidades
Universidad de Chile
Los conceptos desempeñan a central role in the cognitive processes that underlie intelligent behavior. The centrality of the concepts have in the mental life is well expressed in the phrase that Gregory Murphy begins his book The big book of concepts -an excellent review of findings on the subject has made experimental cognitive psychology the last twenty years:
"Concepts are the glue that holds our mental world. When we walk into a room, we tried a new restaurant, go to the supermarket to buy groceries, visit the doctor or read a story, we must rely on our concepts of the world to help us understand what is happening. "(2002:1) rescue Murphy's description of the intuition that most of us have about the importance of the concepts to explain our cognitive functioning. What seems obvious at this intuitive level, becomes a matter of lively debate within cognitive science in formulating theories that specify a clear and evidence based on the role of 'glue', to use the expression Murphy played the concepts in our mental life. Along with this, theories of concepts appeal explicitly or implicitly to questions concerning the nature of concepts, issues that go beyond what could be supported by experimental evidence. The many theories currently on offer in the field [1] can be organized by the different degrees of emphasis to give the one hand, the quality of the experimental evidence that supports them and, secondly, to plausibility of transcendental arguments tending to show that this or that abstract property is (or is) essential to the concepts. Thus, theories of cognitive science concepts could be classified, at least for the purposes of what I present here in three groups:
(i) those that rely heavily theorizing about the concepts in evidential support obtained through the use of reliable procedures (Murphy 2002) [2] and which comprise the largest group;
(ii) those who, despite being overtly philosophical approaches to the concepts of their programmatic nature, rely on an interpretation of some experimental findings as significant, according to the satisfaction of some not very clearly specified criteria, and are, by virtue of this, suggestions for joint empirical theories (Prinz 2003, Clark and Prinz 2004);
( iii) Finally, a small number of theoretical proposals that address purely philosophical about the nature of concepts (eg, a Fodor 1998 and Peacocke 1992) and, presumably, would suggest which are the foundations upon which empirical theories could stand forth in (i) or mixed or program mentioned in (ii) .
theories grouped under (i) compete for their place of hegemony in the field using the usual procedures for evaluating theories. The dependency relationship between theoretical formulations and experimental findings play a crucial role. It is the latter which ultimately determine changes in the theoretical formulations. The experimental data if that are reliable, given the minimum theoretical assumptions can be considered as sufficient evidence, albeit provisional, to maintain certain generalizations or reformulated. The reformulation of the generalizations are not radical, but rather add causal factors, or broken down as dictated by the evidence. Given its acute experimental these theories do not pay much attention to causal factors more abstract as it is difficult to support by experimental evidence. Additionally, it is considered that the issues concerning the nature of concepts can only be addressed on an experimental basis. If you are more abstract or contain assumptions which may not be supported by evidence arising from these bases are viewed more as an object of philosophical discussion. The sign "philosophical discussion" may well be interpreted here as leading to cofusiones by the absence of experimental support.
theories grouped under (ii) make statements depend on the nature of the concepts that could be called salient facts of cognition . Something is considered a salient fact when cognition is supported by one or more experimental findings that have a high impact on various areas of cognitive science research , Among others, cognitive neuroscience, artificial life, situated cognition, of embodied cognition, distributed cognition. So conceived, the salient facts of cognition determine what the nature of concepts, thus excluding any fact or reflection emanating from outgoing or obvious facts, such as, for example, explanations of folk psychology or architectures that Somehow these explanations are compatible with mentalists.
articulate a theory or outline of a research program on the basis of the salient facts of cognition has two dimensions. On the one hand, is an intellectual strategy for theorizing about cognition from a different perspective to the traditional cognitive and mentalist. On the other hand, appeals to find causal factors of intelligent behavior that would be more complex than postulated by the hegemonic conception.
Theories of group (iii) are closer to traditional thinking in philosophy of mind with a strong emphasis on ontological questions. While taking into account the theories of concepts supported experimental evidence, such evidence is not crucial when constructing arguments regarding the nature of concepts. What such theories are individuals looking for properties that are essential concepts. Individuation conditions of these properties can be justified by transcendental arguments. The statements are those arguments relate to an area that is beyond the experimental evidence. If any of the premises of these arguments are based on some kind of experimental evidence, is considered not enough to dislodge an ontological conclusion. Experimental evidence is directly related to procedures that allow a reliable epistemic access to a particular property. Found the individuation of a property in their mechanisms of epistemic access does not lead to clear individuation, as there is and can be a large number of different procedures for access to the same property epistemically. This requires that previously justified the establishment of which of the possible mechanisms turns out to be relevant to the individuation of that property. According to this theory, it appears that there is no criterion to determine the mechanism, nor to determine what would be the set of all mechanisms that are not relevant to the individuation of that property. Therefore, it makes sense to dispense with the appeal to such procedures indiividuar the property in question and based on transcendental arguments to avoid the truth of the premises depend the action of epistemic devices.
exclusion mechanisms ontological epistemic perspective is not universally shared by philosophers of mind who subscribe to theories of type (iii), especially if these philosophers are inclined to vindicate epistemology, understood as a theory attempts to justify how the beliefs, or concepts that are part of them have value or epistemic virtue. For example, Peacocke believes that concepts are entities that are residing in a third area metaphysical, which is neither the mind nor the physical world. However, the concepts are raised ( grasper ) by agents - rather cognizant - who have thoughts (propositional attitudes) of the concept that tries to single out is a part. Thus, the only way to access the essential properties of a concept for purposes of individuation is establishing the conditions of possession by an agent having propositional attitudes, for which the individual is part concept.
Peacocke's proposal might work if each concept corresponds to a specific access mechanism - in this case acquisition (grasping ) - so that this mechanism is relevant to the individuation of the constitutive properties the concept, in a still unspecified. In other words, be required to have only one relation of the subject or knowing who has the attitude, via an epistemic device, with the concept that individuals may be constitutive relationship of the concept. If there are two relations, R1 and R2 that an individual has a propositional content, and whether the concept C is part of that content, it is necessary to have at least one criterion to establish which of the two relationships is relevant to the individuation of C. If you can not find this approach, we can conclude that C is not one but two concepts C1 and C2. So exposed will not be problematic for a neofregeano since the concepts are individuated primarily by sense. Despite what I said, Peacocke, in so far as to individualize the constitutive properties of concepts without resorting to the evidence for empirical theories falls, though not as comfortably, within the theories of type (iii).
Both theories grouped into (i) and (ii) marginalize the ontological reflection of a more conventional field of discussion hardly considered relevant. The main reason is that much of the evidence that these theories put forward to support their claims have a character a priori and not obtained from experimental procedures reliable. Therefore, when the theories grouped under (iii) are subject to an assessment by scientists cognitvos holding some of the theories ascribed to (i) or (ii) the results of this assessment is bleak. Although these cognitive scientists would agree with much of the conclusions drawn inclined philosophers ontologically on the constitutive properties of concepts, saying that these visions are powerless to generate ontological explanations empirically contrasts. For its part, the theories are grouped under (iii) despite trying to justify or refute some of the current theories, the character strongly a priori argument prevents your reflections are a real contribution in the debate that rages between proponents of theories of type (i) and (ii).
The debate among theorists of concepts is to establish what is or what might be the best theory of concepts. For this there is no absolute parameters. Rather different criteria and conditions which yields satisfaction indicators that may lead to the conclusion that a theory has more merit than others to be considered, at least temporarily, as the best.
At this point I would consider a small digression important for a proper understanding of what I'm trying to expose. The tone that I used to express the above may seem a bit lax. By using this tone I want to avoid falling into the widespread temptation to make a monolithic group of conditions, whose satisfaction unequivocally by some theory allows to assign the property of being the best, meaning that property as a kind of epistemic virtue. As far as I understand the discussion among cognitive scientists, regardless of the vehemence and conviction stating that a determined experimentally supported theory is true, the epistemic virtue is rather an idea that regulates the debate about the assessment theories. This does not imply that there is a finite and clearly specified criteria and conditions whose satisfaction leads unequivocally to ascribe to a theory the property of being the best, understanding the subject property as an epistemic virtue. This situation would be analogous to that of a moral philosopher who thought that a finite set of requirements to ensure the achievement of moral virtue through proper implementation of the actions of the type that are contained in the prescriptions. While there are something like epistemic virtue, and although we incline to it without being able to define exactly what it is and, therefore, to establish what would their achievement, there is a finite and limited prescriptions whose satisfaction is sufficient or lead unequivocally to ascribe the property to a particular theory. However, considering the debate around the evaluation of theories and the way it articulates reveals that the debate is dominated by a search based on the best theory. This search is relentless and requires not only ongoing review of the criteria used to evaluate theories, but also encourages the review and reformulation of the theories that are subject to evaluation of their evidence and procedures used to collect the evidence.
Given the number and variety of theories of concepts currently being proposed in cognitive science, the debate between those who hold these theories is intended to show, on the basis of a complex of reasons, what theory is or may be the best . The reasons are articulated on different bases. Partly through the application of traditional criteria of parsimony, psychological reality and explanatory power. Criteria may also include correspondence between the evidence - as distinct from mere data - the characteristics of the process for obtaining evidence, and the nature of the properties and causal relationships between these properties that the hypothesis means. A more recent and increasingly popular exhibits in the field, to evaluate theories according to whether they would satisfy a number of desiderata . They try to express those problematic issues that explicitly or implicitly research in the field believe that a theory of concepts should explain, if this theory would aspire to be the best.
Procedures for evaluation of theories that were listed in the above paragraph does not include the ontological argument. In terms of concepts, these arguments are the prelude to the formulation individuation conditions of the constitutive properties of the concepts are well founded. Generally, theories of types (i) and (ii) tend to ignore this kind of argument. However, supporters of these types of theories make substantive claims about the constitutive properties of concepts, arguyemdo that these claims are either based on experimental evidence or can be inferred from the salient facts of cognition. In both cases the evidence seems compelling to support such claims substantive. Put another way, it is difficult for the constitutive properties of a concept can be guaranteed on these types of evidence, when even We could, on similar evidential bases ensure establishing what the mental domain. Within this context and bearing in mind that the revision of theories is an important consequence of debate assessments carried out by cognitive scientists, ontological arguments should have a place as a component for judging which theory could be considered as the better.
can be argued that the reasons should be excluded from such ontological debate about what is the best theory of concepts, as they have a radically different evidential basis of those, usually of epistemological and methodological that usually wield in such discussions. However, this does not seem a good reason to deny the role of ontology in the evaluation of theories. In fact, a significant group of contemporary alternative theoretical cognitive science, whose proposals currently enjoy huge popularity, appeal to phenomenological or experience of embodied (Clark 1997 and Varela et al. 1991) as a starting point for theorizing about cognition . Alternative theorists seek to naturalize the metaphysical notion of experience, with the aim of which becomes a factual starting point which should be based on the research of cognition. Although I believe that strategy and not the content of the alternative theoretical proposal is partly valid, for now I will not discuss the issue of whether the phenomenological experience is, or may be, a fact which can be based on a science of the mind. Suffice to say, for now, alternative theorists use the notion of metaphysics in question so that if a significant number of salient facts of cognition coincide with one or more of the many interpretations that can be made that concept, then this may taken as a pattern of inference to generate hypotheses on the various more specific areas of cognition.
Another field considerations, alternative theorists believe that the notion of phenomenological experience only has value to cognitive science if some features of it can be made intelligible by its compatibility with certain salient features of cognition. Apparently some of these theorists - such as Clark 1997, for example - believe that the metaphysical notion in question while it is experimentally validated by the salient facts of cognition is, by virtue of this validation, a concept that indicates the constitutive properties of cognition.
If it is considered as a general strategy of argument, what they propose the theoretical alternative is to provide a way to experimentally verify or, if you will, naturalize, a metaphysical notion, and on that basis, conclude that the notion is - or that there are good reasons to believe that such notion is - constitutes cognition. What I propose is to separate or distinguish the areas that alternative theoretical - and, for that matter, anyone using a strategy of naturalization as described - say they are inextricably linked. In other words, the indissoluble character of the nexus between these two areas is not the best, as obvious as it presents the theoretical alternative. On the one hand, the experimental validation they make the experience embodied, considered metaphysical notion does not seem to be sufficient to show that it is constitutive of cognition and, second, it is obvious that they describe experimental findings are empirical realizations of this notion.
Consequently, despite the enthusiasm with which alternative theoretical advertise the similarities between the experience and the salient facts of embodied cognition, common sense, at least for now, suggests to preserve the distinctions between what is possible to say about experimental basis and what it can be said about transcendental bases. However, the strategy of alternative theoretical has the virtue of emphasizing an aspect generally ignored by the theorists of the group (iii), namely that there is a link between the findings experimentally based and ontological claims. The fact that this relationship is not as strong as they seem to affirm the alternative theory suggests that it is better to preserve the distinctions between ontology and experimentally based theories, but at the same time, its contrast to estimate to what extent the claims are plausible on different evidentiary bases, are held in either area. The type of considerations that would authorize the weak link between between ontology and experimentally based theories seem to have little theoretical relevance. Most likely not have it. However, insofar as it promotes the review of theories - and in this case, the revision of ontological concepts - has, as far as I can see, a pragmatic relevance.
NOTES [1] For an overview from the most significant texts, see Margolis and Laurence compilation of 1999. An overview and assessment of the types of theory on offer can be found in Prinz 2003 Caps 1-4. For a discussion of the findings in experimental cognitive psychology, see Murphy 2002.
[2] References restricted only the texts I consider examples of each of the positions that I am characterizing.
REFERENCES Clark, A. 1997. Being there: Putting the brain, body and mind together again . Cambridge, MA: MIT Press.
Clark, A. and JJ Prinz. 2004. Putting concepts to work: Some Thoughts for the twentyfirst century. Mind and Language, 19:, 1: 57-69.
Fodor, JA 1998. Concepts: Where Cognitive Science Went Wrong . Oxford: Oxford University Press.
Margolis, E. and S. Laurence (eds.). 1999. Concepts: Core Readings . Cambridge, MA: MIT Press.
Murphy, G. 2002. The big book of concepts . Cambridge, MA: MIT Press.
Peacocke, C. 1992. A study of concepts . Cambridge, MA: MIT Press.
Prinz, JJ 2002. Furnishing the nd my : Concepts and Their Perceptual basis. Cambridge, MA: MIT Press.
Varela. F., E. Thompson and E. Rosch. 1991. The Embodied Mind: Cognitive science and human experience . Cambridge, MA: MIT Press.
