Part 1:Reflect on Alison Adam's discussion of Artificial Intelligence in the text,"Embodiment and Situatedness: the Artificial Life Alternative." How has framing problems of artificial intelligence as abstract problem solving influenced the direction of the field overall? How would approaching intelligence asrelating toembodiment change the conversation? How has your experience of technology influenced the way you view intelligence and/or the body?
Part 2: Heidegger discusses the relationship of human existence, which he calls "Dasein." with the equipment that makes up its environment. How is Heidegger's approach to technology different than the standard ones that suggest that technology is a collection of tools? What is something interesting that you took away from the Heidegger reading that you hadn't yet considered? What were you confused about?
Artificial Knowing Gender and the Thinking Machine by Alison Adam (z-lib.org) 5 EMBODIMENT AND SITUATEDNESS The artificial life alternative It is hard to overestimate the importance of the body in a consideration of knowledge, for not only does it have a potent cultural force, but also assigning both its role, and men’s and women’s roles in relation to the body, has formed a major element of mainstream Western philosophy, from Aristotle through Descartes, to the present day in artificial intelligence. Rationalist philosophy has sidelined the body in giving the mind the primary role in the making of knowledge and rationality. This reinforces the status of propositional knowledge as the product of mind alone, over and above skills knowledge, which is inconceivable separately from having a body. From this, it is not hard to make the links rationality - transcend- ence - mind - masculine, and to contrast this chain with irrationality - immanence - body - feminine. The question is, how far is the body or embodiment necessary for having knowledge and how does this relate to gender? If there are types of AI which take embodiment into account in a meaningful way, then are these better placed to enrol a concept of gender in their design? Situatedness is a related term which involves the question of the extent to which an agent is situated in an environment, participating in the world, rather than being given one problem after another, which is the typical traditional AI style of interaction. Clearly it is necessary to have a body to be situated. There are at least two aspects of situatedness which are of interest - being physiculZy situated in an environment (which relates to the embodiment problem) and being socially situated in a culture. The social constructivist position in science and technology studies, including actor-network theory, has much to say about the latter, but not much to say about the former, as it seems to shy away from dealing with messy bodies, maintaining a masculine, transcendental (albeit not necessarily rationalist) position. But on the other hand, research on AI which explicitly adopts a situated position, looks at the problem almost exclusively from the physical stance, and has little, if nothing, to say on what it might mean, in AI terms, to be socially situated. In disavowing the rationalist metaphor, there are several strands of philo- 129 EMBODIMENT AND SITUATEDNESS sophical enquiry leading in a broadly similar direction, which can be brought together in this discussion. In doing this I am not trying to impute a superficial feminism to some arguments where none exists. Yet as I hope earlier chapters have demonstrated, it is distinctly possible to use some non-feminist arguments, as in Ryle’s and Dreyfus’s commentaries on the propositional/skills distinction, to feminist ends. The arguments I examine in this chapter are united both in their critique of the rationalist or ‘objec- tivist’ metaphor, and also in arguing that the body plays a crucial role in the making of knowledge. Indeed the discussion unites analyses from categorization and prototype theory (the latter located in the realm of cognitive science itself) cultural studies of the body, and various aspects of feminist theory. Stated briefly, the question is how far our knowledge of the world belongs purely to the mental realm, separate from the body. Is having a body, a body which we have grown up with, which functions not only in the natural world but in the social world, necessary not only for some knowledge, but for all knowledge? If it is necessary, then not only fem- inism, but also parts of philosophy, cognitive science and AI itself, suggest that there are serious problems for large parts of the AI endeavour. Cer- tainly there is a problem with those parts of AI which are predicated on a symbolic view of mental processing, and which either ignore or miscon- ceive the role of the body in the making of knowledge. Once again, bringing in a proper concern with the body in knowledge necessarily brings in its wake a consideration of gender in the making of that knowledge. CATEGORIES AND METAPHORS Two of the most vocal advocates of embodiment as a crucial facet of reasoning are Lakoff (I~u), a cognitive scientist working on categorization and prototypes, and Johnson, a latter-day phenomenologist working on the ‘nonpropositional and figurative structures of embodied imagination’ (Johnson 1987: xxxv). It is no accident that their views have similarities as theirs is a long collaboration (see e.g. Lakoff and Johnson 1980). They have continued to be the chief, and often quoted, protagonists of the attack from embodiment from a view which they term as ‘objectivism’ or ‘logicism’.’ Category theory is the part of cognitive science which is concerned with the way the mind forms categories. If the dominant ‘information pro- cessing’ theory of mind in cognitive science could be termed rationalist then there is, by contrast, a strong realist thread to category theory. This is exemplified in the work of Lakoff, which seeks to establish a bodily basis for reason in a revolt against objectivist rationality. Although they are clearly not feminists themselves, in the increasingly common technique amongst feminist writers, I wish to press the work of these ‘mainstream’ 130 EMBODIMENT AND SITUATEDNESS theorists into service in a feminist argument without trying to rescue the original writers for feminism. * Their arguments provide a wealth of empirical evidence for a bodily view of rationality which is broadly in accord with the feminist view of embodiment and the phenomenology of Dreyfus. Their characterizations of objectivism are similar to each other?. For Lakoff, under the traditional view, reason is seen as abstract and disembodied. On the new view, reason has a bodily basis. The traditional view sees reason as literal, as primarily about propositions that can be objec- tively either true or false. The new view takes imaginative aspects of reason - metaphor, metonymy, and mental imagery - as central to reason, rather than as a peripheral and inconsequential adjunct to the literal. (Lakoff 1987: xi) Lakoff finds evidence for his view of experiential realism in the way people categorize the natural world, a study of which he develops in the domain of prototype theory. In a similar vein, the work of Eleanor Rosch (1973; Rosch and Lloyd 1978) is also influential for language and linguistic cat- egories and metonomy, where a subcategory has a socially recognized status as standing for the category as a whole. An instance of this is the housewife-mother category, where the category ‘working mother’ is defined in contrast to the stereotypical ‘housewife-mother’, even though a very substantial number of mothers with children of school age or younger may actually have paid employment. Lakoff acknowledges that such cat- egories are cultural conventions and that they fulfil a normative role. The ‘mother’ example clearly demonstrates a cultural imperative of homemaker- as-norm, no matter what the reality of the level of mothers’ participation in the workforce might be. The example also serves to reinforce the low esteem in which societies hold the work of women who stay at home, to the extent that their work is not even regarded as work at all. Lakoff looks at the relationship between the body and the formation of concepts. He suggests that the objectivist account of cognition, meaning and rationality makes no mention of who or what is doing the thinking. The function of the human organism is deemed irrelevant, and thought is characterized as symbol manipulation, where the symbols are taken to have a fixed correspondence with things and categories in the world. The notion of a fixed correspondence echoes Dennett’s (1993) concerns with the way that AI fails to ground the semantics of symbols in something in the world. And this also strikes a chord with the ‘view from nowhere’ of traditional epistemology. The rational mind is cut loose from the gen- dered body and floats free in a realm of pure thought - everywhere and yet nowhere. 131 EMBODIMENT AND SITUATEDNESS Meaning and rationality are transcendental - they transcend the limi- tations of any particular kind of being. Rational beings merely partake of transcendental rationality. Thus, in the characterization of what concepts and meaning and rationality are, there can be no mention of the human organism. (Lakoff 1987: 174) Perception is taken to be the way in which correspondences are made between external reality and the symbol system which we use to think. The objectivist account suggests that concepts are just there in the world for us to perceive, and the body has no role in adding to the meaning of concepts. Under Lakoff’s view, experientialist semantics must go beyond mere symbol manipulation as the embodiment of concepts is made directly, through perception, and indirectly through knowledge embodied in social groups. The independence of metaphysics and epistemology from each other is the cornerstone of objectivism which is epitomized in W. V. 0. Quine’s ‘To be is to be the value of a variable’. Yet such a view is completely contradicted by studies in cultural anthropology (ibid.: 208). It is as if mathematical logic is being asked to do too much work, especially in linguistics and in AI; it can be seen as doing a job for which it was not originally designed. So really Lakoff and Johnson’s ‘objectivism’ is a more general term which includes in its purview traditional epistemology and its manifestation in symbolic AI. Their attack is based, in particular, on the way that the role of the body has been ignored in objectivist or logicist accounts. In contrast, the experientialist approach attempts to characterize meaning in terms of the nature and experience of the organism doing the thinking. Meaning is seen in terms of embodiment; ‘meaning is understood via real experiences in a very real world with very real bodies. In objectivist accounts, such experiences are simply absent’ (ibid.: 206). We import struc- ture and order in the world through metaphor, where we use a number of important image schemas. For instance the ‘container schema’ gives us basic boundary distinctions between interiors and exteriors - what is ‘in’ and what is ‘out’. In mapping a source