Studying the organization of work has become an important element in the design of systems and artefacts for work settings. However, although studies of work have had an in#uence on design, there is still some concern about the nature of the relationship between the study of work and design. Concern has been variously expressed in terms of gaps between social science and systems engineering, formality and informality, prescription and negotiation and there have been a number of attempts to bridge those gaps. In contrast, the aim of this paper is to re-view the gap and to characterize it in such a way that bridging may not be an issue. This requires a reconceptualization of the relationship between designing artefacts and understanding work. Using Bateson's levels of analysis, an account of the paradox of framing the study of work by the norms and expectations of a rationalist approach to design, exempli"ed by software engineering, is developed. The study of work is characterized as inevitably involving self-referential observation and inscription, characteristics which create paradox when framed by the demands of rationalist design. However, Bateson's treatment of paradox allows us to see this relationship, not as a gap to be bridged, but as an opportunity to create new forms of punctuation for design, studying work, and relations between them. Many current attempts to reconceptualize design and its relationship with understanding work emphasize the dialogical aspects of practice and theory. Bakhtin's philosophy is used here to advance consideration of the dialogical aspects of &&understanding work for design'' with particular reference to the use of representations such as scenarios. A critique of exemplar representations is used to exemplify dimensions, such as addressivity and un"nalisability, that would characterize a dialogical punctuation of understanding work and design.