”What Is It Like to Be a Robot?” Review of David Mcfarland's Guilty Robots, Happy Dogs
Total Page:16
File Type:pdf, Size:1020Kb
:KDW,V,W/LNHWR%HD5RERW" *XLOW\5RERWV+DSS\'RJV7KH4XHVWLRQRI$OLHQ0LQGVE\'DYLG0F)DUODQG 5HYLHZE\'DQLHO&'HQQHWW %LR6FLHQFH9RO1R 6HSWHPEHU SS 3XEOLVKHGE\University of California PressRQEHKDOIRIWKHAmerican Institute of Biological Sciences 6WDEOH85/http://www.jstor.org/stable/10.1525/bio.2009.59.8.14 . $FFHVVHG Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. University of California Press and American Institute of Biological Sciences are collaborating with JSTOR to digitize, preserve and extend access to BioScience. http://www.jstor.org Fall Focus on Books learn to appreciate some of the advances respect, that they go back to Harvard out the terrain using what you already and controversies in evolutionary de- Press and propose to use the current know to interpret what you find. For- velopmental biology while reading Greg version of their book as the seed for a tunately, David McFarland has chosen Wray. Then again, there is no essay on community-wide, online, open-ended the second option in Guilty Robots, the role of phenotypic plasticity in evo- effort. Of course, it would also be nice if Happy Dogs: The Question of Alien lution, a topic that has acquired central it were open access, but that’s another Minds, and there is much food for status during the past two decades; after story. thought here for both scientists and perusing Evolution a reader might be philosophers. excused for not appreciating the entire MASSIMO PIGLIUCCI It is written in the spirit of Valentino field of evolutionary genomics, or for Massimo Pigliucci (massimo Braitenberg’s brilliant little book Vehicles being ignorant of ongoing discussions @platofootnote.org) is a professor in the (1984), a series of thought experiments on crucial new concepts like evolvabil- Department of Philosophy at the City that led readers from robotic vehicles ity. Even attempts to move beyond strict University of New York, Lehman College. even simpler than bacteria to ever-more biology with entries on evolution and sophisticated and versatile agents capa- society, evolution and religion, and the References cited ble of tracking food, avoiding harm, above-mentioned essay on antievolu- Browne J, Ekman P, Kauffman S, May R, Pigliucci comparing situations, and remembering tionism barely scratch the surface—why M, and Raison C. 2009. Darwin’s Descendants. things. McFarland starts his project a is there no discussion of evolutionary New York Academy of Sciences. (7 August 2009; little higher on the ladder of sophisti- psychology, as controversial and some- www.nyas.org/Publications/Detail.aspx?cid=56e3 cation, with a robot designed to serve as what dubious as the field is? 5057-c0ad-4b83-8c02-bc38082419dc) a night watchman of sorts, identifying While some of these lacunae could Coyne JA. 2009. Why Evolution Is True. Viking. interlopers, calling for help when have been avoided during the planning Diderot D, d’Alembert JR. 1751–1777. Encyclopédie. needed, and, most important, preserv- stages of the volume, I think the under- (7 August 2009; http://quod.lib.umich.edu/d/did/) ing its energy supply for another day, lying problem is that encyclopedic efforts Müller GB. 2007. Evo-devo: Extending the evolu- budgeting its activities to stay alive at all are a thing of the past, certainly when it tionary synthesis. Nature Reviews Genetics 8: costs. This basic robot is then enhanced comes to the paper variety of encyclo- 943–949. in various ways, in a design process pedia. In this bold new era of ubiquitous Norenzayan A, Shariff AF. 2008. The origin and whose ultimate goal is a robot that can and increasingly cheap laptop comput- evolution of religious prosociality. Science 322: be held accountable and to whom things ers, 24/7 Internet access, e-readers, smart 58–62. matter—a robot with subjectivity and phones, and so on, I simply do not see Pigliucci M. 2007. Do we need an extended evolu- values. many people willing to lug around a tionary synthesis? Evolution 61: 2743–2749. How do nonhuman animals compare thousand pages of what is going to be a with such robots? Animal minds (in- necessarily incomplete and increasingly cluding our own) are the real quarry unrepresentative reference source like here, and McFarland uses the parallels Evolution. Publishers, editors, and au- and differences between clearly imagined thors would be much better off em- WHAT IS IT LIKE TO BE A ROBOT? robots and various well-studied animals bracing the anarchy and flexibility of to illuminate the issues in a host of re- the Web to develop decentralized and Guilty Robots, Happy Dogs: The search controversies currently raging in more focused projects, such as the Question of Alien Minds. David psychology and ethology. This has been excellent Complete Works of Charles Dar- McFarland. Oxford University Press, his larger strategy for many years, and win online (http://darwin-online.org.uk/). 2009. 256 pp., illus. $15.95 (ISBN this book gives us a summary of the Even encyclopedias are taking a de- 9780199219308 paper). lessons he has gleaned from this inter- cidedly different form these days, and if disciplinary exploration. one does not like the proletarian ny scientist who wants to investigate One message driven home most Wikipedia, excellent models of schol- A minds—our minds, animal minds, effectively, in my opinion, is that it is arly efforts are out there, such as the alien minds—will soon discover that entirely appropriate to consider natural Stanford Encyclopedia of Philosophy there is no way to proceed without ven- selection to be a (mindless, purpose- (http://plato.stanford.edu/). These take turing into the playgrounds and battle- less) designer, and to compare the de- seriously the idea of organic, grassroots fields of the philosophers. You can either signs churned up by eons of natural growth arising from the efforts of a ded- stumble into this investigation and selection on a par with designs generated icated community, based on what the thrash about with a big scientific stick, top-down by would-be intelligent community itself sees as worth writing thwacking yourself about as often as designers—engineers and roboticists. about, as opposed to the centralized your opponents, or you can enter cau- Sometimes the perspective is particu- planning typical of the standard model. tiously, methodically, trying to figure larly bracing, as when McFarland in- Indeed, let me suggest to Ruse and sists on situating his imagined robots Travis, both of whom I know and highly doi:10.1525/bio.2009.59.8.14 in a market economy so he can note www.biosciencemag.org September 2009 / Vol. 59 No. 8 • BioScience 707 Fall Focus on Books that nobody would buy such a robot— out explicit representation, and he pro- behavior seems to suffice: “Certainly we it wouldn’t pay for itself. Animals, sim- poses to define cognitive processes as can say that the teacher behaves as if it ilarly, are amazingly thrifty because they those that require “a certain kind of wants, hopes, or desires the pupil to be- have to be; they have superb layers of mechanism—one that requires manip- have in a certain way,” he says, but he also self-protection and repertoires of self- ulation of explicit representations” goes on to note that the teacher could advancing behaviors, but not a smidgen (p. 87). This sets the bar high and departs have a “strong theory of mind” about the more than can pay for itself in the long from standard usage, but perhaps it is pupil and be wrong (p. 105). The com- run. This often brings out the rationale best to follow his lead. Note that with parison with robots is always astringent for animal (or robot) features that would this definition, it isn’t clear that our here, and McFarland puts our built-in otherwise be lost in the shadows. It also hand-eye coordination or even our abil- skepticism about robot minds to good obliges McFarland to commit to a ity to find our way home counts as a use in reining in our romanticism about “behaviorist” approach—not the ideo- cognitive process (unless we use a map our furry friends. logical straitjacket of the Skinnerians or an explicit mental map). McFarland proposes a contrast but the behaviorism expressed by Tur- between two views of what is going on ing in 1937, when he noted about the inside: the hedonic model and the human computers of his day: “The be- automaton model. According to the he- havior of the computer at any moment donic, “the feelings of pleasure and dis- is determined by the symbols which he pleasure that arise from various parts is observing, and his ‘state of mind’ at of the body in situations of motivational that moment” (p. 241). Handsome is as compromise are combined in some way, handsome does, a motto enshrined in and behavioral adjustments are made the rationale for the Turing test, and the so as to maximize pleasure and minimize only way a science of mind can proceed. displeasure.” By contrast, in the au- But how much can one glean from tomaton, “the behavioral and physio- inner behavior (the machinery of the logical adjustments are automatic, mind, in effect) by observing the com- and...the system is attuned to produce petencies of outer behavior? Do ani- the best compromise among the com- mals, for instance, really have the beliefs peting demands” (p.