《Definitions and dimensions of etiquette.doc》由会员分享，可在线阅读，更多相关《Definitions and dimensions of etiquette.doc（7页珍藏版）》请在三一办公上搜索。
1、Definitions and Dimensions of Etiquette Christopher A. MillerSmart Information Flow TechnologiesMinneapolis, MN.cmillerSIFTAbstractThis paper addresses the question “Why Etiquette?” What do we mean by etiquette in the realm of Human-Computer interaction? Does a focus on Human-Computer Etiquette prov
2、ide anything more than traditional usability or human-centered design? And finally, as we explore the topic of human-computer etiquette, do the characteristics of etiquette in human-human relationships give us any special insights for human-computer interactions? This paper provides a review of thes
3、e questions on the basis of the set of papers and interests expressed in the AAAI Fall Symposium on Etiquette for Human Computer Work. IntroductionCopyright 2002, American Association for Artificial Intelligence (www.aaai.org). All rights reserved. Approximately two and a half years ago, while co-ch
4、airing a AAAI Spring Symposium on Adaptive User Interfaces, I took advantage of the fact that, as a co-chair, it was unlikely that anyone would reject my paper to produce a soapbox polemic on the topic of Human-Computer Etiquette Miller, C. 2000. “Rules of Etiquette, or How a Mannerly AUI should Com
5、port Itself to Gain Social Acceptance and be Perceived as Gracious and Well-Behaved in Polite Society”. In Working Notes of the AAAI Spring Symposium Workshop on Adaptive User Interfaces. Stanford, CA; March 20-22. My purpose was primarily to draw attention to what I viewed as a flaw in much of the
6、exciting work in adaptive and intelligent user interfaces in the AI and AI-influenced world. Specifically, that they all too often behaved like little children: interrupting ongoing conversation or work to show off their capabilities at the least sign of interest, that they exhibited capabilities pr
7、imarily for the sake of showing off what they could do rather than for the sake of helping to advance the goals of their human users (their “betters”?), and that they persisted in exhibiting the same behavior long after it had ceased to be useful or interesting. While this behavior was tolerable in
8、young children and, perhaps, in young systems fresh out of the research lab, such systems needed to “grow up” and participate in the rules and conventions of the societies into which they hoped to be accepted if they were to ever gain acceptance.In fairness, I wasnt just pointing a finger at the wor
9、k of others, and I wasnt completely original. On the one hand, Eric Horvitz had written about a similar concern with regards to Microsofts Office AssistantsTM a year earlier Horvitz, E. 1999. Principles of Mixed-Initiative User Interfaces. Proceedings of CHI 99, ACM SIG-CHI Conference on Human Facto
10、rs in Computing Systems, Pittsburgh, PA, May. And for my part, I had noticed similar tendencies in my own projects: for example, pilots deemed initial versions of the Rotorcraft Pilots Associate Cockpit Information Manager Miller, C. and Hannen, M. 1999. “The Rotorcraft Pilots Associate: Design and
11、Evaluation of an Intelligent User Interface for a Cockpit Information Manager”. Knowledge Based Systems, 12. pp. 443-456. far too willing to provide aiding for behaviors it understood and recognized. On the other hand, we had also noted, that the human pilots in this domain spent as much as a third
12、of their time engaged in inter-crew coordination activitiesthat is, in “meta-communication” about their intents and plans. We designed and implemented a simple interface which allowed our smart Cockpit Information Manager to participate in that conversation, taking instruction and declaring its inte
13、nt and its understanding of the pilots intent. This modification seems to have resulted in some improvement in human + machine system performance, as well as larger gains in user acceptance Miller, C. and Funk, H. 2001. “Associates with Etiquette: Meta-Communication to Make Human-Automation Interact
14、ion more Natural, Productive and Polite”. In Proceedings of the 8th European Conference on Cognitive Science Approaches to Process Control. September 24-26, 2001; Munich, pp. 329-338. . In hindsight, I believe that part of the reason this form of interacting with a cockpit “associate” showed improve
15、ments was that it fit into the existing “etiquette” that pilots had evolved and expected from any new participant in the domain. The interface we implemented did not exhibit “etiquette” in the general sense of politeness, but it did behave according to the established rules and behavioral convention
16、s of the role within the domain for which it was intended. These factors have led me to think more deeply about the “etiquette” of human-computer relationships in work domains and, ultimately, to this symposium. What began as an extended metaphor embedded in a rant is now leading to the thought that
17、 we may be systematically missing, or at least failing to give proper consideration to, an important aspect of the way our systems will be used. In the remainder of this paper, I begin by providing some rationale and defense for the choice of the term “etiquette”. Next, I address the question of whe
18、ther a focus on etiquette provides any thing more or different than a traditional focus on usability or human-centered system design. Finally, I provide some characteristics of etiquette in human-human relationships and discuss their impact on the design of human-computer interfaces and human-comput
19、er interaction.General vs. Specialized EtiquettesWhy use the term “etiquette” as opposed to other, related terms that denote other, related fields of study (e.g., social computing Preece, J. (Ed.) 2002. Supporting Community and Building Social Capital. In _Communications of the ACM, 4(4), April. pp.
20、36-73., embodied or personified agents Cassell, J., Sullivan, J., Prevost, S. and Churchill, E. (Ed.) 2000. Embodied Conversational Agents. MIT Press, Cambridge, MA., human centered computing or design Billings, C. 1997. Aviation Automation: The search for a human-centered approach. Erlbaum. Mahwah,
21、 NJ. , etc.? How does a focus on etiquette differ from these other fields? As described above, it was initially suggested to me as a part of an extended metaphor indicating a direction that sophisticated, adaptive computer interfaces needed to move in if they were to perform acceptably for long term
22、 use by humans with jobs to do. It, of course, also has the benefit of being somewhat provocativeespecially among those who are used to thinking of computers as either unintelligent tools that must be forced to perform desired tasks, or as demons actively trying to thwart successful human endeavors.
23、The term is open to multiple interpretations. The American Heritage Dictionary defines “etiquette” as follows: “(1) the body of prescribed social usages. (2) Any special code of behavior or courtesy: “In the code of military etiquette, silence and fixity are forms of deference” (Ambrose Bierce). Syn
24、onyms: etiquette, propriety, decorum, protocol. These nouns refer to codes governing correct behavior.” Morris, W. (Ed.). 1978. The American Heritage Dictionary of the English Language. Houghton Mifflin, Boston. p. 451. Of the two definitions offered, the second is closer to the sense in which I mea
25、nt, and it led to the definition included in the Call for Participation in this workshop: By etiquette, we mean the defined roles and acceptable behaviors and interaction moves of each participant in a common social settingthat is, one that involves more than one intelligent agent. Etiquette rules c
26、reate an informal contract between participants in a social interaction, allowing expectations to be formed and used about the behavior of other parties, and defining what counts as good behavior. Etiquette, in this sense, need have little to do with politeness or the “social niceties,” and specific
27、 environments, specific work “cultures,” even individual teams of humans will all have their unique sets of expected behavioral norms or etiquettes. Essentially, an etiquette for a specific domain is formed by prescribing some subset of the range of possible human behaviors as appropriate or inappro
28、priate, expected or unexpected for those who participate in that domain (see Figure 1). These behaviors may pertain to speech, dress, movement , etc., or to more specific protocol behaviors (e.g., in this plant, we always empty a vessel when its not in use). As such, these are the behaviors that any
29、 human or automation agent should strive to adhere to if it wants to be accepted into that milieu.Figure 1. Etiquette for a domain defines acceptable behaviors for that domain.Nevertheless, the first definition applies equally well. Indeed, several of the papers to be presented at this symposium esp
30、ecially Davis, J. 2002. Understanding and decreasing adversive behavior in online social contexts. This volume., Whitworth, B. 2002. Polite Computing: Software that respects the user. This volume., Louwerse, M., Graesser, A., Olney, A. and the Tutoring Research Group. 2002. Good Computational manner
31、s: Mixed-initiative dialog in conversational agents. This volume. and Edwards, J., Scott, G., McFadden, S. and Hendy, K. 2002. Traps, pitfalls, swindles, lies, doubts and suspicions in human-computer interactions: A counter-case for the study of good etiquette. This volume. all take the perspective
32、of etiquette as good manners, though they investigate different aspects of what that might mean and how it might be applied. This is hardly problematic. Just as we encounter multiple etiquettes in our daily lives (what counts as appropriate, expected behavior is different at work, in a bar, in churc
33、h, etc.), so computers designed with etiquette in mind should be expected to exhibit different kinds of etiquette for the different kinds of contexts they are used in.The relationship between the “good manners” and the “established codes of behavior” meanings of the term “etiquette” seems to be a si
34、mple one of general vs. specialized application. The etiquette rules of good manners are general, at least in the sense that they are intended to apply in situations where no more specific set of rules is known to apply. For example, the use of “please” and “thank you” and the polite forms of addres
35、s can frequently be dispensed with, but they are a good way to start an interaction with someone who is unknown to us. In fact, Brown, P. and Levinson, S. 1987. Politeness: Some universals in language usage. Cambridge University Press; Cambridge, UK. claims that most of the polite forms in conversat
36、ion, regardless of their specific linguistic or cultural manifestation, are intended precisely to ensure that the least offensive interpretation possible is applied to our utterances and actions. By contrast, the etiquette rules of specialized domains may be less formal and may involve substantial d
37、eviation from the “polite” forms. Work domains are a common sphere in which specialized etiquettes prevailetiquettes that differ from formal and polite norms. This is true for forms of address and the use of “please” and “thank you,” but it is true in more interesting ways as well. For example, (as
38、Reeves and Nass demonstrate Reeves, B. and Nass, C. 1996. The Media Equation. Cambridge University Press, Cambridge, UK. ) we take it as inappropriate for a conversational partner in social interaction not to maintain a certain amount of small movementsindicative of a lack of intelligence or attenti
39、on, or of artificiality. Nevertheless, an assistant in a work setting who is always moving about the office may be seen as intrusive or at least fidgety. It is also true that many work domains have evolved highly specialized etiquettes designed to smooth operations and/or ensure safety. The conversa
40、tional conventions used between pilots and Air Traffic Controllers is one particularly formal example, about which Corker, K. 2002. Performance impacts of digital communication in aerospace operations: Roger, Over. This volume. will have more to say at the symposium. When such conventions exist, whe
41、ther formally documented and trained (as in the air traffic control case) or not, it will be important for computers which wish to perform actions similar to humans in the domain to perform in a fashion that preserves the expected etiquette which those humans would adhere to. So, in short, both sens
42、es of the term “etiquette” seem relevant for human-computer interactionthough perhaps in different settings. Politeness and adherence to social codes will likely be useful in exactly those situations where it is important in human-human interactions: in general situations, between strangers, when th
43、e specific context or relationship with the conversant is unknown, when it is important not to give offense, etc. On the other hand, especially given that computers are frequently (though certainly not exclusively) used as tools in the pursuit of work goals in specific work domains, the specialized
44、codes of etiquette which are familiar and expected in those domains will be important for the computer to use. Workers in an oil refinery are generally not accustomed to addressing each other with honorifics; therefore it is likely that a computer system that does so will be seen as out of place. Pi
45、lots acknowledge communications with a “Roger” and rarely use “thank you”; therefore it will be critical for an aviation system to acknowledge communications and it will be unlikely that the use of “Thank you” will be seen as appropriate.Is Etiquette just good interface design?The notion of etiquett
46、e does, I believe, force us to consider several aspects of human-computer relationships that traditional design concerns do not. By placing the system to be designed in the role of a well-behaved human collaborator we gain insights into how users might like or expect a system to act and how its acti
47、ons might be interpreted as a social being that rarely come from any other source of design (with the possible exception of usability reviews for an already-designed system). I find it instructive to ask, for example, “if this system were replaced by an ideal human assistant, how would that assistan
48、t behave?” and, alternatively, “if a human assistant were to provide this information/decision /recommendation/control input in this way, how would s/he be perceived by colleagues?” To pick, perhaps unfairly, on a well-known example: how would a human office assistant who, several times a day, interrupted my work to offer to help me write a letter be regarded? Would or should traditional approaches to human-computer interface design provide similar insights? Perhaps, but there is a growing set of criticisms claim otherwise. Traditional approaches to HCI design have
