Malafouris, Lambros (2004)
***
A few years ago, the publication of Cognition and Material Culture (Renfrew & Scarre 1998) well exemplified that the science of mind and the science of material culture are two sides of the same coin. I consider the present volume to be an invitation to move a step further by placing our focus this time explicitly upon the realm where cognition and materiality intersect, mutually catalyzing and constituting each other. The process of material engagement
as recently introduced by Renfrew (2001a,b; this volume) offers a new analytic means for that
purpose, and my primary objective in this paper is to advance some proposals that will help the reception and better appropriation of this ‘hypostatic approach’ for the advancement of cognition-oriented archaeological research.
To this end, two main avenues are available. The first, as stated in the title of this volume, is to
rethink materiality; the second and correlated one to rethink cognition. Following the second avenue, and building from a cognitive basis, my aim is to propound a hypothesis of the constitutive intertwining of cognition and material culture. This I do on one hand as a method toward a theory of material engagement, and on the other as a means of reclaiming
cognition from the bonds of cognitivism. The important questions raised, both for archaeology and for the general domain of cognitive science call, of course, for a more extensive discussion than I can carry out here. My concern, however, is simply to clarify the ground and to stimulate a sort of direction, in the hope that the results so obtained will commend it to others.
I start with a brief note about the realm of material engagement which, strangely enough, can be
conceived as the most familiar and at the same time unknown existential territory. To exemplify, this territory is familiar, as when the hand grasps a stone and makes it a tool, yet it remains terra incognita, since—despite a long genealogy of analytic efforts—just what this grasping implies for the human condition remains elusive, and refuses to be read in the narrative fashion that hermeunetics have promised.
Various factors may underlie this blind spot, but there is one that I want to emphasize from the
very beginning. Despite the fact that contemporary archaeological theory appears in agreement about adopting a relational viewpoint, more often than not, it is either unwilling to follow the consequences of such a conviction or remains in a state of confusion about what this might imply in practice. The general call for non-dichotomous thinking in archaeology (e.g. Hodder 1999; Tilley 1994; Gosden 1994; Thomas 1996) seems captivated in the optical array of a Müller-Lyer illusion (Fig. 5.1). Knowing that the lines between the arrows are equal, we still perceive them as different; knowing that mind and matter are relational entities, we continue to approach them
through the Cartesian lenses of symbolic representation. It seems that the purification project of modernity(Latour 1993) that habituated our minds to think and talk in terms of clean divisions and fixed categories blocks our path as we seek to shift the focus away from the isolated internal mind and the demarcated external material world towards their mutual constitution as an inseparable analytic unit. Thus, material culture remains one ‘of the most resistant
forms of cultural expression in terms of our attempts to comprehend it’ (Miller 1987, 3), while cognition continues to look like a disembodied information-processing ghost captured in the laboratories of Artificial Intelligence.
I am afraid that, as long as cognition and material culture remain separated by this ontological gulf, our efforts to understand the nature of either is doomed to failure. Approaching the engagement of mind with the material world on such a basis will achieve nothing more than constantly reiterating a question-begging procedure which can be compared with an attempt to separate and analytically prioritize the process of ascending from descending in the famous drawing of Escher (1960; Bool et al. 1982). Finding an escape route from our Cartesian prison
demands more than a small displacement in our academic ‘language games’. Removing the arrows of modernity from the archaeological perceptual field is not an easy task; it will involve a great deal of cognitive dissonance (Malafouris 2003). Yet to tackle the complex intentionalities enacted through the materiality of the archaeological record, we need to move on and where necessary transgress the ontological tidiness of our modern taxonomies, just as conceptual art transgressed the aesthetic tidiness of the Renaissance.
With these remarks in mind let me now turn to defining the problem more closely.
Redefining the boundaries of mind: the problem with cognitivism
Ever since the famous Cartesian line between the ‘thinking thing’ and the ‘extended thing’ was drawn, the philosophy of mind has had to confront the crucial question of the so-called mind–body problem(Ryle 1949). In order to separate the mind from the body and by implication from the world, a mechanism was needed to account for how those independent
components interact. The notion of symbolic representation was gradually introduced to bridge this huge ontological gap, thus furnishing the principal mechanism by which we feed our cognitive apparatus with facts and information from the ‘external world’ as well as suggesting the way by which we materialize and externalize our mental contents.
Grounded on the premise of this representational thesis, cognitivism, or the so-called computational view of mind, emerged during the sixties as an attempt to re-define human conceptual architecture in the image of the digital computer (Gardner 1985; Dupuy 2000). That is to say, mind was viewed as a storehouse of passive internal representational structures and computational procedures, as a ‘filing cabinet’ capable of receiving and manipulating external sensory information (Clark 1997). Mind was then to the brain as a computer programme is to the hardware of the computer on which it runs. This to a large extent remains the dominant paradigm in contemporary cognitive science, as well as the implicit model behind most archaeological accounts of prehistoric cognition which conceptualize the human mind primarily through the idioms of representation and information-processing (for a concise discussion
of this trend see Mithen 1998, 8–10). Consequently, Zubrow’s statement that to understand
cognition means to ask ‘how do humans represent knowledge and what do they do with that
representation’ (1994, 109) echoes directly the major analytic imperative of cognitivism. This imperative can be summarized as follows: discover the representational and computational capacities of the mind and their structural and functional representation in the brain (Gardner 1985, 36).
The thing to note, however, is that behind the undeniable advances in the study of the human mind that this paradigm has brought about, one can easily trace some very important shortcomings. For example, in implementing computational theory in the laboratories of artificial intelligence (AI), it soon became manifest that although simulations based on computational logic proved extremely effective in complex analytic tasks, as for example running a program capable of winning a chess game, they were highly problematic in tasks as simple as instructing an automaton to find its way outside a room without running into the walls. In fact, when the first such autonomous devices (machina speculatrix) were constructed by Grey Walter (1953), they had nothing to do with complex algorithms and representational inputs. Their kinship was with W. Ross Ashby’s Homeostat (1952) and Norbert Wiener’s cybernetic feedbacks (1948) rather than with the complex representational structure of the by-that-time famous Turing machine (1950). On the basis of a very simple electromechanical circuitry, the so-called ‘turtles’ were capable of producing emergent properties and behaviour patterns that could not be determined by any of their system components, effecting in practice a cybernetic transgression of the mind–body divide, and materially exemplifying a model of human cognition the implications of which are yet to be realized and properly digested in contemporary cognitive science (see Brooks 1991).
What the above implies for the computational model in question cannot be pursued here in detail
(see Dupuy 2000; Boden 1990; Clark 2001). However, to make a long story short and easier to comprehend, it is safe to argue that the major problem with this paradigm was, and remains, that it provides a view of human cognition so purified and detached from the world that in the end it resembles a ‘brain in a vat’, a disembodied input–output device characterized by abstract, higher-level logical operations. This means that, using computational simulations as a method for gaining information about the human mind, you might learn a few things concerning the representational structures that support inferential logic and problem solving, but you will certainly also end up with a distorted picture as to how those structures relate to the environment, and probably with no picture at all as to how those structures are enacted in real-life situations and in different cultural settings. As Ingold (1998, 431) remarks:
. . . it makes no more sense to speak of cognition as the functioning of such a [computational] device than it does to speak of locomotion as the product of an internal motor mechanism analogous to the engine of a car. Like locomotion, cognition is the accomplishment of the whole animal, it is not accomplished by a mechanism interior to the animal and for which it serves as a vehicle.
In other words, computationalism in most cases failed the test of ecological validity. Turing’s algorithms and Chomskian grammars, however effective in mapping analytic procedures of disembodied intellects, scored very low when the cognitive task at issue involved embodied minds engaging with the material world in real-life settings.
.....page 55 (3)
댓글 없음:
댓글 쓰기