Scholarly book review of Understanding Computers and Cognition: A New Foundation for Design By: Terry Winograd and Fernando Flores November 23, 2002 Samuel A. Burns Information Technologies and the Information Professions LIS 386.13 Instructor: Dr. Ronald Wyllys Graduate School of Library and Information Science The University of Texas at Austin
Review of Understanding Computers and Cognition 2 Review of Understanding Computers and Cognition In Understanding Computers and Cognition (A New Foundation for Design) (1986), written by Terry Winograd a computer scientist and AI researcher and Fernando Flores a philosopher, former minister of finance in Chile and businessman the authors explore the theoretical and practical aspects of computer design and propose an alternative orientation to create a new understanding of how to design computer tools suited to human use and human purposes (p. 8). Using technology according to the authors, leads to fundamental changes in what we do, and ultimately in what it is to be human (p. xi). The impetus for this new understanding was born out of the many bold claims and predictions that were made about the capabilities of computers and artificial intelligent systems in the early 1980's. For Winograd and Flores (1986), the claims that computers could mimic human cognition and become electronic brains or autonomous intelligent agents revealed a dependence on a flawed intellectual tradition namely, the Rationalistic or Analytic tradition (p. 3). In their view, the Rationalistic Tradition is based on a misinterpretation of the nature of human cognition and language and their goal in the book is to develop a new orientation that centers on social activity as the foundation of cognition (p.78). Guided by this orientation, they assess the way we think about computers and ultimately conclude that computers would best serve us by supporting people in the complex conversational structures generated within organizations (p.12). Ultimately, the findings in this book are not what are most important. The authors divulge early on in the book what they are going to argue, how they will argue it and
Review of Understanding Computers and Cognition 3 what they will conclude from their arguments. They describe their analysis of language, cognition, and computers as a quest. Describing this quest, they explain that: Progress is made not by finding the right answers, but by asking meaningful questions ones that evoke an openness to new ways of being. We invite readers to create with us an openness that can alter our collective vision of how computer technology will develop in the coming decades. (1986, p. 13) Language and Rationalistic Tradition The first part of the book (chapters 1-6) introduces the rationalistic tradition and provides the theoretical framework that begins the shift towards a new orientation of understanding. They start by laying out the foundation of the rationalistic orientation. According to Winograd and Flores (1986), this orientation emphasizes the formulation of systematic rules that can be used to draw logical conclusions (p. 15). In it, there is an objective reality and language is used as a system of symbols that are composed into patterns that stand for things in the world (p. 17). Moreover, the authors say that the rationalistic orientation holds that decision making and problem solving is characterized as a process of information gathering and processing in which we choose among alternatives according to an evaluation of outcomes (p. 20). In the authors view, this tradition is not obvious to those immersed in it and as a result it generates blindness. The authors specifically cite the scientific method and cognitive science as two areas that fall prey to the blindness of the rationalistic tradition (p. 23). A central theme that develops early on in the book, which bears mentioning, is the importance of language. Critiquing the rationalistic theories of language and promoting alternative views of language and understanding are one of the primary aims of the book.
Review of Understanding Computers and Cognition 4 Winograd and Flores own feelings on this aspect of the book are expressed in the following passage: In this endeavor we are doubly concerned with language. First we are studying a technology that operated in a domain of language. Second, in looking at the impact of the computer, we find ourselves thrown back into questions of language how practice shapes our language and language in turn generates the space of possibilities for action. Much of our theory is a theory of language, and our understanding of the computer centers on the role it will play in mediating the facilitating linguistic action as the essential human activity. (p. 7) The significance of this point cannot be understated because Winograd and Flores intend with this book to do no less than propose a new foundation for computer design on the basis of a new discourse about language and thought. Towards a New Orientation The dismantling of the rationalistic tradition proceeds by offering up the phenomenological or hermeneutic school of philosophy (mostly Martin Heidegger but also Hans-Georg Gadamer, and Jürgen Habermas), the research of Chilean biologist Humberto Maturana, and the Speech Act Theory of J.L. Austin and John R. Searle as an alternative basis for cognition. In the rationalistic tradition, according to Winograd and Flores (1986), mental representations of the objective world and logical analysis of representations in the mind form the basis for cognition. This is how problems are solved, rational action occurs and understanding of the world is achieved. The authors new orientation proposes that cognition is based on relationships to our environment that
Review of Understanding Computers and Cognition 5 are unarticulated and cannot be expressed in the notions of commonsense knowledge or world knowledge or any other form of represented knowledge (p. 72-73). Winograd and Flores (1986) first call on Heidegger who they say emphasized that humans are in a state of thrownness as a condition of being in the world (p. 71). In such a state, humans are always engaged in acting within a situation, without the opportunity to fully disengage ourselves and function as detached observers (p. 71). As a consequence, our acts cannot be understood as the results of a process (conscious or non-conscious) of representing, planning, and reasoning (p. 71). From Maturana s studies of biological systems, Winograd and Flores (1986) derive the notion that humans beings ability to function as observers is generated from our functioning as structure-determined systems, shaped by structural coupling (p. 71). The authors conclude from both Heidegger and Maturana that the nature of cognition is not an activity in some mental realm, but a pattern of behavior that is relevant to the functioning of the person or organism in its world (p. 71). The ideas of structural coupling and being-in-the-world give way to the concernful activity that brings speech act theory into the Winograd and Flores (1986) argument. Their analysis of speech act theory emphasizes the importance of background and interpretation (p. 75). They examine many different utterances and possible communication breakdowns in order to show the patterns of interaction that occur within a shared background (p. 75). The meaning ascribed to language, in the authors view, is a result of listening to the commitment expressed in speech acts and in our articulation of content (p. 68). This analysis leads to the book s most provocative claim that nothing exists except through language (p. 68).
Review of Understanding Computers and Cognition 6 Implications for Computers and Design In the second part of the book, Winograd and Flores (1986) examine computers and design through the lens they create out of the hermeneutic, biological, and speech act theory of part 1. They feel that computers are valuable, but that they will never be intelligent. In their view, computers will always have to interpret the world within a set of concepts or symbols. The programmer acts within a context of language, culture, and previous understanding, both shared and personal. The program is forever limited to working within the world determined by the programmer s explicit articulation of possible objects, properties, and relations among them. (p. 97) In order to be considered intelligent, programmers would have to equip computers with the ability to participate in the structural coupling in which to negotiate the meaning of language and make commitments or with the ability to learn and evolve. In Winograd and Flores (1986) view, this is impossible and we should instead look in different directions for the design of powerful computer technology (p. 93). The following passage sums up this point: Our central claim in this book is that the current theoretical discourse about computers is based on a misinterpretation of the nature of human cognition and language. Computers designed on the basis of this misconception provide only impoverished possibilities for modeling and enlarging the scope of human understanding. They are restricted to representing knowledge as the acquisition and manipulation of facts, and communication as the transferring of information.
Review of Understanding Computers and Cognition 7 As a result, we are now witnessing a major breakdown in the design of computer technology. (78-79) A New Basis for Design Winograd and Flores (1986) conclude from this orientation that the goal of designing computers should not be for developing intelligent autonomous agents like expert systems, but rather to support the fundamental relationships between language and successful actions and to facilitate human work and interaction (p. 162-163). They acknowledge the success and importance of programs such as word processing software, spreadsheets, and other systems based on rule-governed manipulation of formal representations, but they emphasize that these systems are successful because they let people operate effectively in a systematic domain that is relevant to human work and not because they themselves are intelligent (p. 174-175). As for the next wave of computer development, Winograd and Flores (1986) propose a system called The Coordinator Workgroup Productivity System (p. 159). The product was marketed by Action Technologies, which was founded by Fernando Flores, and was presented as the first example of a new class of products which would be used at a workstation and provide a ready-to-hand tool that operates in the domain of conversations for action (p. 159). For Winograd and Flores, this system embodied their primary focus of designing computer-based tools as part of a larger perspective of ontological design (p. 177). Conclusions and Analysis Ultimately what Winograd and Flores succeed with in this book is in questioning our assumptions about computers and by extension our own understanding of what it
Review of Understanding Computers and Cognition 8 means to be intelligent. Sixteen years after the publication of this book, debates still rage over the possibilities of artificial intelligence. Proponents of strong AI support the view that programmed machines are capable of cognitive mental states while proponents of weak AI believe that programmed machines can simulate human cognition (Dennett, 6, 1988). Many of the linguistic problems mentioned in Understanding Computers and Cognition most notably abductive reasoning or the frames problem are still significant hurdles that have yet to be adequately dealt with by proponents of artificial intelligence (Sutherland, 4, 2001). As much as this book is notable for the insightful questions it raised early on about AI, its conclusions about human understanding, the limitations of computers and new foundations for design are largely unconvincing and somewhat problematic. For example, Winograd and Flores never adequately demonstrate that people immersed in the rationalistic tradition are unaware of the limitations of language and understanding that they are critical of. There is no evidence that rationalists are not aware of so-called breakdowns in communication and there is equally no proof that such breakdowns would be removed if one were to adopt the hermeneutic tradition of Heidegger. An example of an analytical approach that accounts for such a breakdown can be found in Warren Weaver s identification of the semantic problem and the effectiveness problem referred to by Winograd and Flores (1986) as communicative competence (p. 162) in Claude Shannon s Mathematical Theory of Communication. Weaver's contributions in acknowledging the potential for error in the meaning and effectiveness of language suggests that Winograd and Flores idea of blindness in the rationalistic tradition can be overcome. It also suggests that one need not be engaged in Winograd and
Review of Understanding Computers and Cognition 9 Flores' (1986) new orientation in order to be cognizant of the limitations of understanding (Wyllys, 2002). Another significant weakness with the book is that it does not successfully integrate the theories it develops in the first two parts of the book into the last one that deals with designing computer tools. The discussion in this part of the book only touches on significant socio-technical implications of computer tools and is almost exclusively related to business and work scenarios of social interaction (e.g. workers and managers communicating in an office environment). This point of reference seems especially limiting when contrasted with Nardi and O days (1999) vision of an information ecology. Despite Winograd and Flores reliance on biological theories, their new foundations for design do not seem nearly as all-encompassing as do Nardi and O days' (1999), which take into account people, practices, technologies, and values (p. 211). Nardi and O'days' perspective is a more holistic interpretation of collaborative technological communication that includes local differences as well as the strong interrelationships among the social, economic, and political contexts in which technology is invented and used (p. 47). Winograd and Flores ignore these concepts altogether. Impact on Library and Information Science The questions raised by Winograd and Flores in Understanding Computers and Cognition are important ones for the Library and Information Science domain. This profession concerns itself with how information flows through and enlivens and gives meaning to people's lives (Wyllys, 2002). As enablers of information search and retrieval, LIS professionals collect information; organize it; embed it in information systems; create interfaces for it; and provide search capabilities that exploit the
Review of Understanding Computers and Cognition 10 statistical characteristics of information and human information seeking behavior (2002 Bates). As Marcia Bates pointed out in her article After the Dot-Bomb: Getting Web Information Retrieval Right This Time (2002), good information retrieval design requires just as much expertise about information and systems of information organization as it does about the technical aspects of systems ( 4). Winograd and Flores (1986) view that we can create computer systems whose use leads to better domains of interpretation is a primary concern to information professionals because their work revolves around interpretations of information (p. 179). Examples of areas within the LIS domain that are concerned with communication and technology include but are not limited to usability, human computer interaction, information architecture, OPAC design, digital libraries and data mining. Also, LIS professionals work to develop information retrieval vocabularies, design metadata protocols, classification schemes and indexing systems that all seek to minimize errors and maximize efficiency on behalf of an information seeker. The potential for breakdowns in communication is a chief concern of library and information science practitioners because it strikes that the very heart of the information exchange that information professionals are seeking to facilitate. In this sense, Winograd and Flores exploration into computers and human cognition is central to the complicated and subtle business of making use of information resources (Bates, 5, 2002). Information professionals would be served well by exploring the meaningful questions about being and the development of computers that enrich this book.
Review of Understanding Computers and Cognition 11 References Bates, M.J., (2002, July 1). After the dot bomb: Getting web information retrieval right this time. First Monday, 7, 7. Retrieved November 10, 2002, from http://www.firstmonday.org/issues/issue7_7/bates/index.html Dennett, D. C. (1988). When philosophers encounter AI [Electronic version]. Daedalus, 117, 283-295. Nardi, B.A., & O'Day, V.L. (1999). Information ecologies. Cambridge ; Sebastopol, CA: O'Reilly. : MIT Press. Sutherland, K. (2001, March 10-11) A defining light in the darkness of the mind [Electronic version]. The Times Higher Education Supplement. Retrieved October 30, 2002, from http://www.thes.co.uk/search/story.asp?id=71426&state_value=archive Winograd, T., & Flores, F. (1986). Understanding computers and cognition: A new foundation for design. Norwood: Ablex. Wyllys, R.E. (2002). Information expertise: Comments by marcia bates. Retrieved September 18, 2002, from http://www.gslis.utexas.edu/~l38613dw/website_fall_02/readings/informationexpertise.html Wyllys, R.E. (2002). Information theory and digital representations. Retrieved September 5, 2002, from http://www.gslis.utexas.edu/~l38613dw/powerpoint/38613infothdigitalrep.ppt